By nuViteo Labs · Published · Editorial review
Scientific source verification through 2026-11-04. Publication and editorial-review dates describe preparation of these articles, not a new scientific reappraisal. The scientific source verification remains limited to the canonical release verified through 2026-11-04. This is a directed evidence selection, not a systematic review or GRADE assessment.
Editorial interpretation
start with the question, not the biggest percentage
A clinical-trial result is an answer to a specified question in a specified population. Before asking which number is larger, ask who entered the trial, what treatments were compared, which outcome was measured, when it was measured, and how treatment discontinuation was handled. These details determine the meaning of an average. Removing them can make an accurate number support an inaccurate conclusion.
The selected cards concern semaglutide and tirzepatide, with adults without diabetes in the weight-change populations discussed here. They are not a complete catalogue of treatments or an individual treatment recommendation. The GLP-1 treatments journey keeps each finding attached to its design. A larger average weight change may be relevant, but it is not a personal forecast or a complete account of benefits, burdens, and safety.
What the evidence shows
what the placebo-controlled weight trials report
In STEP 1, adults without diabetes with obesity or overweight and a weight-related condition received semaglutide or placebo, with lifestyle intervention in both groups. From baseline to week 68, the reported mean weight changes were −14.9% and −2.4%. The published adjusted between-group difference was −12.4 percentage points. These canonical results use the primary treatment-policy estimand, assessing effects regardless of treatment discontinuation or rescue interventions; they should not be replaced by subtraction of rounded arm means.
In SURMOUNT-1, adults without diabetes with obesity or overweight and a weight-related complication had reported mean changes from baseline to week 72 of −15.0%, −19.5%, and −20.9% with tirzepatide 5, 10, and 15 mg, respectively, versus −3.1% with placebo. These are the publication's treatment-regimen estimand results, addressing effects regardless of treatment discontinuation in the intention-to-treat population, not a relabeling of the registry's differently handled outcome values.
These trials establish comparisons against their own placebo groups. Reading their active-arm averages side by side does not create randomization between the drugs across studies. Differences in trial populations, procedures, follow-up, and analysis remain. The right question is whether a direct comparison exists and, if it does, what that comparison actually estimated.
Trace these results to their evidence cards
Exact result identifiers in the curated release:
What the evidence shows
the head-to-head comparison and its analysis population
SURMOUNT-5 directly compared tirzepatide with semaglutide in an open-label trial of adults without type 2 diabetes who had obesity or overweight with a weight-related comorbidity. Both arms received diet and physical-activity intervention. From baseline to week 72, least-squares mean weight changes were −20.2% and −13.7%, respectively, with a reported adjusted contrast of −6.5 percentage points. These results use the modified treatment-regimen estimand, not the separate efficacy estimand.
The SURMOUNT-5 card distinguishes 751 randomized participants, 750 exposed to at least one dose and included in the modified intention-to-treat parent population, and 636 evaluable participants in the registry's primary-outcome entry. These counts describe different roles. The registry outcome count must not replace the parent analysis population or be used to characterize the analysis as complete-case-only.
The modified treatment-regimen analysis includes discontinuation, but it is not a rule that every subsequent observation is included without exception. The canonical card describes exclusions after nonstudy crossover to the comparator and after weight-loss procedures, alongside specified missing-data handling. This is why estimand should not be translated into a vague promise of real-world effectiveness. The SURMOUNT-5 Trial Inspector makes the reported comparison, uncertainty, and denominator distinctions easier to inspect.
Trace these results to their evidence cards
Exact result identifiers in the curated release:
Editorial interpretation
weight change is not a body-composition or strength result
A weight endpoint does not distinguish fat from lean tissue. The SURMOUNT-1 DXA substudy adds measured body-composition information, but it is a selected substudy with pooled tirzepatide doses and its own analysis set. It cannot be treated as a body-composition analysis of SURMOUNT-5, nor can its pooled results be assigned to a single dose or every person in the parent trial.
Its DXA lean mass is lean soft tissue excluding bone mineral content. It is not a direct measurement of skeletal muscle, and the card contains no direct strength or performance outcome. This matters whenever a headline moves from weight loss to claims about muscle quality or preserved function. The Body Composition Decoder helps separate those concepts without estimating an individual's tissue loss or establishing a diagnosis.
The distinction is not an argument that weight change is unimportant. It is an argument for matching the outcome to the claim. A trial can provide a useful weight estimate while leaving important questions about function, long-term trajectory, or a particular person's experience unanswered. Asking for those missing outcomes strengthens a conversation rather than contradicting the result that was actually measured.
Limits and uncertainty
a treatment effect is not a complete benefit-risk assessment
The weight cards also include adverse-event context, but descriptive safety counts or percentages do not by themselves demonstrate comparative safety superiority. Different discontinuation outcomes and safety windows cannot be merged into a generic tolerability ranking. Where a safety denominator or exact observation window was not verified, the absence should remain visible rather than be filled using a randomized arm size or the efficacy timepoint.
Verification for STEP 1 and the primary SURMOUNT-1 card used original abstracts and official registries; the full NEJM articles were not accessed. SURMOUNT-5 additionally used its statistical analysis plan, but not its full NEJM article. The DXA substudy full text was consulted. These are scoped verification statements, not interchangeable full-text appraisals, and no GRADE rating or systematic review of the entire treatment question was performed.
Finally, outcomes measured during a trial's treatment period do not establish weight maintenance after withdrawal. They also do not tell an individual what dose to use, whether treatment is suitable, or how to respond to adverse effects. Those decisions require individual assessment and additional information. The Understanding weight loss journey is a route into the research, not a substitute for that assessment.
Questions to take forward
make the trial relevant without making it personal prediction
Ask: How closely does this trial population resemble my situation? Are we discussing an active-versus-placebo comparison, a head-to-head comparison, or an informal comparison across studies? What happened to participants who discontinued treatment, and what does the reported analysis leave out? Which outcomes matter to my goals beyond the scale?
Ask: What adverse-effect information is relevant to me, and which questions does this evidence set not answer? How will we discuss follow-up and maintenance separately from the initial treatment response? Questions for Your Visit can help turn those uncertainties into an organized discussion without turning group averages into a promise.