Tirzepatide SURMOUNT-1 Trial: NCT04184622

SURMOUNT-1 was a large phase 3 randomized, double-blind, placebo-controlled trial in 2,539 adults with obesity or overweight plus a qualifying weight-related condition and without diabetes. At week 72, all three tirzepatide groups had larger mean body-weight reductions and more participants reaching at least 5% weight reduction than placebo. The result applies to the sponsor's studied tirzepatide program, eligibility criteria, lifestyle co-intervention, estimand, and follow-up, not to every population or any independently sourced material.[1][2][3]

The primary hypotheses concerned the 10 mg and 15 mg groups. The same weight-change and at-least-5% comparisons for 5 mg were key secondary hypotheses, not additional co-primary hypotheses.[4][7]

FDA subsequently approved the finished Zepbound product under NDA 217806, and the April 2026 label identifies NCT04184622 as Study 1. FDA's history also lists a later labeling supplement, so that downloadable label is not presented here as a complete account of the latest safety wording. Approval belongs to the defined finished product, not every material called tirzepatide.[5][6][8]

Study at a glance

The protocol regimen is summarized to identify the study.[4] It is not individualized dosing or administration guidance.

Endpoint hierarchy and multiplicity

The registered protocol and analysis plan place percentage weight change and at-least-5% response for 10 mg and 15 mg in the primary family. The two corresponding 5 mg comparisons are key secondary objectives. Some higher response thresholds and cardiometabolic outcomes also belong to the controlled secondary family, with the exact dose or pooled-dose comparison specified in the plan.[4][7]

The paper describes graphical control of the false-positive error rate within each estimand independently. It does not claim control across the two estimands as one combined family. The paper's Table 2 also distinguishes the 5 mg group's 10%, 15%, and 20% response thresholds and waist change as additional secondary outcomes without confirmatory testing; the 25% threshold is exploratory and not controlled for this error rate.[1][7]

The posted SAP is version 2, approved September 28, 2021, and the protocol is amendment (b), approved October 4, 2021. These were amendments during the trial, not the original documents at enrollment. The SAP revision history records changes to the testing strategy, missing-data methods, and secondary objectives. Availability of a plan is not proof that every planned analysis was performed.[4][7]

Estimands and missing-data methods

The primary paper calls its headline analysis the treatment-regimen estimand, aiming to describe the randomized treatment comparison regardless of treatment discontinuation. It uses analysis of covariance with imputed missing values for the week-72 mean weight change. The paper separately reports an efficacy estimand addressing the hypothetical effect if treatment were administered as intended, using a repeated-measures model.[1]

The registered SAP needs an additional distinction. It specifies a hybrid estimand for the primary US analysis, using the full analysis set regardless of adherence. Missingness solely from exceptional circumstances such as a pandemic is handled under a missing-at-random assumption using the same arm's observed data. Other missingness uses measurements retrieved after treatment discontinuation in the same arm, with a placebo-reference fallback if those data cannot support the model. A separately named treatment-regimen sensitivity analysis uses retrieved-dropout imputation for missing week-72 values. Those labels and methods should not be silently treated as identical.[7]

The paper's figure notes distinguish COVID-only missingness from other missingness, and FDA's April 2026 Study 1 table reports the same headline values while explicitly calling its imputation hybrid. This page retains the paper's treatment-regimen label for the numbers below while making that methods distinction visible; it does not claim to have reproduced the analyses or verified every sensitivity result.[1][5][7]

The SAP's efficacy analysis set excludes data later than seven days after the last dose. The registry's first 72-week table describes exclusion of post-discontinuation observations and reports the larger efficacy-estimand weight changes, rather than the headline treatment-regimen figures. An outcome-analysis denominator is not automatically the number with an observed week-72 measurement.[3][7] See the guide to trial estimands and missing data for why these choices change the question being answered.

Week-72 body-weight results

At week 72, the paper reported the following mean percentage changes from baseline:

Each active group differed from placebo with p<0.001: primary comparisons for 10 mg and 15 mg and a key secondary comparison for 5 mg. The estimated differences from placebo were −11.9, −16.4, and −17.8 percentage points for 5 mg, 10 mg, and 15 mg, respectively.[1][7] The confidence intervals describe the precision of each modeled group mean. They do not by themselves measure bias, missing-data sensitivity, long-term durability, or transferability outside the protocol.

For the at-least-5% weight-reduction outcome, again primary for 10 mg and 15 mg and key secondary for 5 mg, the proportions were 85% (95% CI 82%–89%), 89% (86%–92%), and 91% (88%–94%) in the 5 mg, 10 mg, and 15 mg groups, versus 35% (30%–39%) with placebo.[1]

Secondary outcomes and what they add

At least 20% weight reduction was reported in 50% of the 10 mg group and 57% of the 15 mg group, compared with 3% of the placebo group. These 10 mg and 15 mg comparisons were key secondary outcomes in the amended plan. Selected prespecified cardiometabolic measures also changed, but their dose-specific or pooled comparisons must retain the hierarchy described above.[1][7]

These outcomes describe magnitude and associated measures over 72 weeks. They do not establish that every observed biomarker change caused a later health benefit, and they do not replace direct outcome trials for questions about cardiovascular events, mortality, or other diseases.

The ClinicalTrials.gov record now includes longer follow-up in the subgroup with prediabetes at randomization. That extension is a later analysis with a narrower population and should be covered separately rather than merged into the original 72-week paper's headline result.[3]

Adverse events

The most common adverse events with tirzepatide were gastrointestinal, mostly mild to moderate, and occurred mainly during escalation. Adverse events led to treatment discontinuation in 4.3%, 7.1%, and 6.2% of the 5 mg, 10 mg, and 15 mg groups, respectively, versus 2.6% with placebo.[1]

The paper's safety table uses 630, 636, 630, and 643 participants for 5 mg, 10 mg, 15 mg, and placebo. At least one treatment-period adverse event was reported by 510 (81.0%), 520 (81.8%), 497 (78.9%), and 463 (72.0%), respectively. Adverse-event treatment discontinuations were 27, 45, 39, and 17, corresponding to the percentages above.[1]

Serious adverse events affected 40 (6.3%), 44 (6.9%), 32 (5.1%), and 44 (6.8%) participants, respectively: 160 overall. Similar percentages do not establish equal safety. Eleven deaths were reported: four, two, one, and four in those groups. The external committee classified whether deaths were cardiovascular-related; that classification is not the same as deciding whether treatment caused a death. The raw distribution alone does not establish causality.[1]

This historical trial report cannot replace up-to-date approved-product labeling or provide patient-specific safety advice. Its original reporting window must also be kept separate from the registry's later adverse-event tables, which now extend through week 193.[3] The adverse-events and tolerability guide explains the distinction between event counts, seriousness, causality, and withdrawals.

Attrition and missing data

The paper reports that 86.0% of randomized participants completed the 72-week primary treatment period and 81.9% adhered to the assigned tirzepatide or placebo regimen. Completion ranged from 88.4% to 89.8% across the tirzepatide groups and was 77.0% with placebo; adherence ranged from 83.5% to 85.7% and was 73.6%, respectively.[1]

The current registry's primary-treatment-period flow lists 509 of 643 placebo participants, 569 of 630 in the 5 mg group, 566 of 636 in the 10 mg group, and 575 of 630 in the 15 mg group as completed: 2,219 of 2,539 overall. That total is not numerically identical to the paper's 86.0% summary, and this page has not established a participant-level reconciliation. Study completion, treatment adherence, and an outcome-analysis population are not interchangeable measures.[1][3]

The SAP's completion definition additionally requires specified weight measurements and, for participants without prediabetes, safety follow-up. Its modified intention-to-treat population includes randomized participants who received at least one dose; all 2,539 randomized participants did receive treatment. The first registry weight-outcome analysis includes 635, 623, 629, and 625 participants in placebo, 5 mg, 10 mg, and 15 mg groups, not simply those observed at week 72.[3][7]

FDA's April 2026 Study 1 table reports missing week-72 body weights for 21.6% of placebo participants and 10.2%, 10.5%, and 9.4% in the 5 mg, 10 mg, and 15 mg groups. The imputation model addresses these gaps under assumptions; it does not turn the missing values into observed measurements. Differential missingness and the hybrid-versus-efficacy analysis distinction remain important even with a large effect estimate.[5][7]

Statistical significance, effect size, and clinical importance

• Statistical significance: the 10 mg and 15 mg primary comparisons and the corresponding 5 mg key secondary comparisons differed from placebo under the reported testing framework.
• Effect size: absolute group means, placebo comparisons, and responder proportions show the magnitude; p-values do not.
• Uncertainty: group-mean intervals are not the same as intervals for the treatment difference, and neither captures every design or missing-data assumption.
• Clinical importance: the weight changes were substantial over 72 weeks, yet the trial was not designed to resolve every long-term clinical outcome or all populations.

Sponsor roles

Eli Lilly funded, designed, and oversaw the trial. Site investigators collected the data; the sponsor monitored sites, collated data, and performed analyses. The paper states that investigators had confidentiality agreements with the sponsor, all authors had full data access, and all participated in interpretation and manuscript review. It acknowledges Lilly personnel Wei Fan for data preparation/analysis and Farai Chigutsa for medical writing and editorial assistance.[1]

These roles are relevant transparency information. They do not themselves invalidate the randomized comparison or establish independent review by this website.

FDA scope and label-version limits

FDA announced Zepbound's original weight-management approval on November 8, 2023. The retrieved April 2026 label covers reduction of excess weight and long-term maintenance in adults with obesity, or overweight with at least one weight-related condition, alongside a reduced-calorie diet and increased physical activity. It also includes moderate-to-severe obstructive sleep apnea in adults with obesity, supported by separate studies, not by treating SURMOUNT-1 as an apnea trial.[5][6]

At the September 2 Edmonton / September 3 UTC check, Drugs@FDA listed an August 26, 2026 labeling supplement with no label file posted for that entry. The April 2026 PDF cited here therefore predates a later FDA action and must not be described as the complete latest safety label. Consult the live FDA application history for subsequent documents; this study summary does not provide a comprehensive current warning list.[5][8] Our guide to checking trial and FDA status explains this version distinction.

Important limitations

• The original publication reports 72 weeks, not lifetime durability or every long-term safety outcome.
• Adults with diabetes were excluded from this specific trial, as were several higher-risk or clinically unstable groups.
• All groups received a structured lifestyle intervention; the comparison is not tirzepatide versus no accompanying program.
• The comparator was placebo, not another active weight-management drug.
• Completion and adherence were lower in the placebo group, making estimand and missing-data handling important.
• The original study population was 67.5% female and about 70.6% White; generalizability should be assessed rather than assumed.
• Eli Lilly sponsored and funded the trial, and several authors were company employees.
• FDA approval of a finished product does not authenticate a separately sourced material or extend beyond current labeling.

What this study does not prove

SURMOUNT-1 does not prove the same result in people excluded by the protocol, under a different duration or co-intervention, with another outcome, or with a material that has not been shown to match the trial test article. It does not establish individual response or guarantee benefit without risk.

The trial also does not validate a third-party research material. A name, chromatogram, supplier statement, or nominal amount alone does not establish the application-specific identity, formulation, manufacturing controls, or clinical equivalence of FDA-approved Zepbound. See approved drug versus research material for this boundary.

Frequently asked questions

What were the SURMOUNT-1 primary endpoints?

They were percentage weight change and at least 5% weight reduction at week 72 for the 10 mg and 15 mg comparisons with placebo. The same two comparisons for 5 mg were key secondary objectives.[4][7]

Did SURMOUNT-1 include people with diabetes?

No. The 72-week trial excluded diabetes; findings should not be transferred to a different population without direct evidence.

Why do some sources show larger weight changes?

The paper's efficacy-estimand results and the first registry table exclude specified post-discontinuation data, unlike the headline analysis. The registered SAP also distinguishes hybrid imputation from a separate treatment-regimen sensitivity analysis. The analysis label, missing-data method, and denominator must travel with the number.

How many participants completed the primary period?

The paper reports 86.0% completion and 81.9% adherence. The registry's current primary-period flow reports 2,219 of 2,539 completed. These published summaries are not numerically identical and are not silently treated as the same measure here.

Were adverse events different from placebo?

Gastrointestinal events were the most common with tirzepatide, and adverse-event discontinuation was more frequent in each active group than with placebo. Serious-event percentages were similar across groups in the paper.

Is tirzepatide FDA approved for weight management?

Yes, FDA approved the finished Zepbound product for its labeled weight-management use. The cited label is revised April 2026, and FDA's application history lists a later August 2026 labeling action; this page is not a complete current safety-label summary.[5][6][8]

Does this trial validate any material sold as tirzepatide?

No. The evidence belongs to the defined sponsor test article and protocol. Shared naming does not establish product or lot equivalence.

Primary and official sources

1. Primary paper: SURMOUNT-1 in the New England Journal of Medicine
2. PubMed record: SURMOUNT-1, PMID 35658024
3. ClinicalTrials.gov: NCT04184622 registration and posted results
4. NCT04184622 protocol amendment (b), October 4, 2021
5. FDA Zepbound prescribing information, revised April 2026, supplement 050
6. FDA: Original approval announcement for Zepbound
7. NCT04184622 statistical analysis plan, version 2, September 28, 2021
8. Drugs@FDA: Zepbound application and supplement history

Review and corrections

Sources checked September 2, 2026 in Edmonton (September 3, 2026 UTC). Serum Research is responsible for this editorial summary and its corrections. No independent scientific, clinical, statistical, analytical, or regulatory review of this page is claimed. Registered protocol/SAP passages and the available official paper text were checked, but the journal's supplementary appendix was not retrieved and no participant-level analysis was reproduced. The completion-summary and analysis-label distinctions above are not silently resolved. This is educational, not medical advice or a preparation, administration, or dosing guide, and makes no material-quality or equivalence determination.

Read how we review evidence, browse the study records or Research Library, and contact Serum Research about corrections.