Controlled study
SURMOUNT-4: continuation or withdrawal after a 36-week tirzepatide lead-in
After a 36-week tirzepatide lead-in, 670 participants were randomized. Over the next 52 weeks, weight change from randomization was −5.5% with continued tirzepatide and +14.0% after switching to placebo; adjusted difference −19.4 percentage points (95% CI −21.2 to −17.7).
During the SURMOUNT-4 randomized phase and safety follow-up, serious adverse events occurred in 10 of 335 participants in each group. Adverse-event discontinuation was 1.8% with continued tirzepatide and 0.9% with placebo.
Weight results during 68 or 72 treatment weeks and a two-week sleep intervention answer different questions from a withdrawal comparison after initial treatment. SURMOUNT-4 provides that specific comparison for 52 weeks after randomization.
- Evidence design
- Randomized withdrawal comparison
- Population
- Adults aged >=18 years with BMI >=30 kg/m2, or >=27 kg/m2 with a weight-related complication, without diabetes; randomized population had completed 36-week lead-in and achieved a tolerated tirzepatide dose of 10 or 15 mg. Lifestyle counseling continued in both randomized groups.
- Assessment period
- Randomization at week 36 to week 88; Weeks 36–88 and safety follow-up
- Reported units
- % change; percentage points; %
Keep in view
- The percentage reference is body weight at week 36, not initial week-zero weight.
- The randomized population was selected after reaching a tolerated dose: 670 of 783 lead-in enrollees.
- A 14% increase relative to week-36 weight does not mean regaining 14% of the weight previously lost.
- This does not evaluate gradual tapering, other maintenance strategies or individual treatment decisions.
See study details — SURMOUNT-4
Study design and methods
Randomized withdrawal comparison
Phase 3 trial: 36-week open-label tirzepatide lead-in, followed by 52-week randomized 1:1 double-blind placebo-controlled withdrawal phase.
Population
Adults aged >=18 years with BMI >=30 kg/m2, or >=27 kg/m2 with a weight-related complication, without diabetes; randomized population had completed 36-week lead-in and achieved a tolerated tirzepatide dose of 10 or 15 mg. Lifestyle counseling continued in both randomized groups.
- Outcome definition
- Percent change in body weight from randomization at week 36 to week 88
- Assessment period
- Value
- 52
- Unit
- weeks
- Starting study week
- 36
- Ending study week
- 88
- Time reference
- randomization after open-label lead-in
- Reported statistic
- least-squares mean
- Unit
- % change from randomization
- Open-label lead-in enrollees
- 783
- Lead-in participants who received at least one dose
- 782
- Randomized participants
- 670
- Estimand and analysis rules
- Name
- treatment-regimen estimand
- Definition
- Mean treatment effect for all randomized participants regardless of adherence, using the full analysis set.
- Statistical analysis
- ANCOVA for continuous endpoints; Figure 2 describes hybrid imputation. Detailed imputation beyond the published Methods was not reverified from the supplement.
- Secondary estimand distinction
- Efficacy estimand excludes post-discontinuation data and uses MMRM. Registry values -6.7% versus +14.8% and difference -21.4 pp belong to this different analysis and are not mixed into this primary card.
Extracted outcomes
| Outcome and group | Reported value | Assessment period | Sample and denominators | Uncertainty |
|---|---|---|---|---|
| Percent change in weight from week 36 to week 88 Continued tirzepatide Effect estimate | -5.5 % change least-squares mean Unit: % change | Randomization at week 36 to week 88 Study weeks 36–88 | Reported N: 335 — randomized full analysis population
| 95% confidence interval: -6.8 to -4.2 |
| Percent change in weight from week 36 to week 88 Switch to placebo at week 36 Effect estimate | 14 % change least-squares mean Unit: % change | Randomization at week 36 to week 88 Study weeks 36–88 | Reported N: 335 — randomized full analysis population
| 95% confidence interval: 12.8 to 15.2 |
| Percent change in weight from week 36 to week 88 Continue tirzepatide minus switch to placebo Effect estimate | -19.4 percentage points adjusted mean difference Unit: percentage points | Randomization at week 36 to week 88 Study weeks 36–88 | Reported N: 670 — combined randomized full analysis population
| 95% confidence interval: -21.2 to -17.7 Reported p-value: <0.001 (Between-group comparison) |
| Participants with a serious adverse event Continued tirzepatide Descriptive result; not an effect estimate | 3 % reported event percentage Exact numerator: 10 Unit: % | Weeks 36–88 and safety follow-up | Reported N: 335 — safety analysis population
| Event-frequency confidence interval not verified. |
| Participants with a serious adverse event Switch to placebo at week 36 Descriptive result; not an effect estimate | 3 % reported event percentage Exact numerator: 10 Unit: % | Weeks 36–88 and safety follow-up | Reported N: 335 — safety analysis population
| Event-frequency confidence interval not verified. |
| Discontinuation due to adverse events Continued tirzepatide Descriptive result; not an effect estimate | 1.8 % reported event percentage Exact numerator: 6 Unit: % | Weeks 36–88 and safety follow-up | Reported N: 335 — safety analysis population
| Event-frequency confidence interval not verified. |
| Discontinuation due to adverse events Switch to placebo at week 36 Descriptive result; not an effect estimate | 0.9 % reported event percentage Exact numerator: 3 Unit: % | Weeks 36–88 and safety follow-up | Reported N: 335 — safety analysis population
| Event-frequency confidence interval not verified. |
Percent change in weight from week 36 to week 88 — Continued tirzepatide: extraction context
- Outcome role
- Not specified in the consulted scope
- Scale reference
- week_36_body_weight
- Estimand and analysis rules
- Name
- treatment-regimen estimand
- Definition
- Mean treatment effect for all randomized participants regardless of adherence, using the full analysis set.
- Statistical analysis
- ANCOVA for continuous endpoints; Figure 2 describes hybrid imputation. Detailed imputation beyond the published Methods was not reverified from the supplement.
- Secondary estimand distinction
- Efficacy estimand excludes post-discontinuation data and uses MMRM. Registry values -6.7% versus +14.8% and difference -21.4 pp belong to this different analysis and are not mixed into this primary card.
- Extraction location
- JAMA Table 2 (treatment-regimen estimand)
Percent change in weight from week 36 to week 88 — Switch to placebo at week 36: extraction context
- Outcome role
- Not specified in the consulted scope
- Scale reference
- week_36_body_weight
- Estimand and analysis rules
- Name
- treatment-regimen estimand
- Definition
- Mean treatment effect for all randomized participants regardless of adherence, using the full analysis set.
- Statistical analysis
- ANCOVA for continuous endpoints; Figure 2 describes hybrid imputation. Detailed imputation beyond the published Methods was not reverified from the supplement.
- Secondary estimand distinction
- Efficacy estimand excludes post-discontinuation data and uses MMRM. Registry values -6.7% versus +14.8% and difference -21.4 pp belong to this different analysis and are not mixed into this primary card.
- Extraction location
- JAMA Table 2 (treatment-regimen estimand)
Percent change in weight from week 36 to week 88 — Continue tirzepatide minus switch to placebo: extraction context
- Outcome role
- Not specified in the consulted scope
- Scale reference
- difference_in_percent_change_from_week_36
- Estimand and analysis rules
- Name
- treatment-regimen estimand
- Definition
- Mean treatment effect for all randomized participants regardless of adherence, using the full analysis set.
- Statistical analysis
- ANCOVA for continuous endpoints; Figure 2 describes hybrid imputation. Detailed imputation beyond the published Methods was not reverified from the supplement.
- Secondary estimand distinction
- Efficacy estimand excludes post-discontinuation data and uses MMRM. Registry values -6.7% versus +14.8% and difference -21.4 pp belong to this different analysis and are not mixed into this primary card.
- Extraction location
- JAMA Table 2; reported adjusted contrast
- Published difference −19.4 pp; do not replace it with −19.5 from rounded arm means.
Participants with a serious adverse event — Continued tirzepatide: extraction context
- Outcome role
- Descriptive safety outcome
- Extraction location
- JAMA Table 3
- Safety after lead-in selection; it does not describe everyone starting tirzepatide.
Participants with a serious adverse event — Switch to placebo at week 36: extraction context
- Outcome role
- Descriptive safety outcome
- Extraction location
- JAMA Table 3
- Safety after lead-in selection; it does not describe everyone starting tirzepatide.
Discontinuation due to adverse events — Continued tirzepatide: extraction context
- Outcome role
- Descriptive safety outcome
- Extraction location
- JAMA Table 3
- Safety after lead-in selection; it does not describe everyone starting tirzepatide.
Discontinuation due to adverse events — Switch to placebo at week 36: extraction context
- Outcome role
- Descriptive safety outcome
- Extraction location
- JAMA Table 3
- Safety after lead-in selection; it does not describe everyone starting tirzepatide.
What the curated evidence supports
- After a 36-week tirzepatide lead-in, 670 participants were randomized. Over the next 52 weeks, weight change from randomization was −5.5% with continued tirzepatide and +14.0% after switching to placebo; adjusted difference −19.4 percentage points (95% CI −21.2 to −17.7).
- During the SURMOUNT-4 randomized phase and safety follow-up, serious adverse events occurred in 10 of 335 participants in each group. Adverse-event discontinuation was 1.8% with continued tirzepatide and 0.9% with placebo.
- Weight results during 68 or 72 treatment weeks and a two-week sleep intervention answer different questions from a withdrawal comparison after initial treatment. SURMOUNT-4 provides that specific comparison for 52 weeks after randomization.
Limitations
- The percentage reference is body weight at week 36, not initial week-zero weight.
- The randomized population was selected after reaching a tolerated dose: 670 of 783 lead-in enrollees.
- A 14% increase relative to week-36 weight does not mean regaining 14% of the weight previously lost.
- This does not evaluate gradual tapering, other maintenance strategies or individual treatment decisions.
Verification scope
Verified through 2026-11-04 within the consulted scope.
- Bibliographic metadata checked
- Yes
- Original article abstract consulted
- Yes
- Full-text consultation
- Consulted
- Official trial registry consulted
- Yes
- Statistical analysis plan consulted
- No
- Verification boundary
- Verification is limited to the specific fields and sources consulted; it is not a full systematic appraisal or GRADE rating.
Source record
Continued Treatment With Tirzepatide for Maintenance of Weight Reduction in Adults With Obesity: The SURMOUNT-4 Randomized Clinical Trial (2024). JAMA.
Source identifiers
- Article DOI: 10.1001/jama.2023.24945
- PubMed record 38078870
- Source record: jamanetwork.com/journals/jama/fullarticle/2812936
- PubMed Central: PMC10714284
- Trial registration: NCT04660643
- Structured registry: NCT04660643