A headline gives two weight-loss figures: 20.2% and 13.7%. Does that mean one medicine will help every person lose an extra 6.5% of their weight? Those numbers come from SURMOUNT-5, a randomized comparison of tirzepatide and semaglutide, but they describe estimated group averages after 72 weeks, not individual forecasts.[1] Understanding the experiment turns an eye-catching comparison into something more useful: evidence about particular treatments, in a defined population, measured against specific outcomes.
One experiment makes the comparison more informative
SURMOUNT-5 randomly assigned 751 adults with obesity but without type 2 diabetes to tirzepatide or semaglutide. It was a phase 3b, open-label, controlled trial, with equal allocation between the two groups. Both treatments were given by subcutaneous injection, meaning under the skin, using the protocol's maximum tolerated treatment levels over 72 weeks.[1]
Random assignment helps make treatment groups comparable at the outset. That gives a direct comparison an advantage over placing results from separate trials beside one another: differences in enrollment and study design are less likely to masquerade as a drug difference.
“Open-label” means participants and investigators knew the assigned treatment. Randomization and blinding solve different problems; knowing the treatment can influence expectations and behavior. The design therefore supports a direct comparison while leaving potential effects of that knowledge relevant to interpretation.[1]
The primary endpoint, or main planned outcome, was percentage change in body weight from baseline to week 72. Key secondary endpoints included waist circumference and the proportions reaching specified weight-loss thresholds.[1]
The difference was measured in percentage points
The abstract reports least-squares means: averages estimated through the statistical analysis, rather than simply a list of individual results. Weight change was −20.2% with tirzepatide and −13.7% with semaglutide. The corresponding 95% confidence intervals were −21.4% to −19.1% and −14.9% to −12.6%.[1] These intervals express uncertainty around the estimated group means; they are not ranges containing 95% of individual responses.
| Outcome at week 72 | Tirzepatide | Semaglutide | How to read it |
|---|---|---|---|
| Estimated mean weight change | −20.2% | −13.7% | A 6.5-percentage-point difference between reported means.[1] |
| Estimated mean waist change | −18.4 cm | −13.0 cm | A 5.4 cm difference between reported means.[1] |
| Reaching at least 10%, 15%, 20% or 25% weight reduction | More likely at every listed threshold | Less likely at every listed threshold | The abstract does not give the individual threshold percentages.[1] |
The weight and waist comparisons each had P<0.001.[1] That provides strong statistical evidence of a difference under the analysis assumptions. It does not describe the probability that a particular person will respond better to one treatment.
Subtracting the weight percentages gives 6.5 percentage points, not a promise of an extra 6.5% loss for everyone. The abstract supports greater average reductions with tirzepatide for these endpoints; it does not supply a personalized prediction.[1]
Where the conclusion stops
The population matters. Adults with obesity but without diabetes were the people studied; this result cannot simply be assigned to people with diabetes, adolescents or every other population.[1] The treatments matter too. This was a comparison of specified injectable regimens, not every formulation, route or newer regimen carrying either molecule's name.[1]
The endpoints also set boundaries. A larger reduction in weight or waist circumference does not, by itself, establish a larger benefit for every health outcome. The reported comparison does not answer which treatment prevents more cardiovascular events, which is best tolerated by a particular person, or what happens after treatment ends.[1]
Nor does it isolate why the treatments differed. Semaglutide activates the GLP-1 receptor; tirzepatide activates both GIP and GLP-1 receptors.[2][3] Comparing two complete molecules is not the same experiment as adding only GIP activity while holding everything else constant. SURMOUNT-5 therefore cannot assign its average difference exclusively to GIP.
Eli Lilly funded the trial.[1] That is relevant context for evaluating design, analysis and reporting, alongside the randomized results. This article uses the original abstract, which does not provide enough detail to independently assess missing-data handling, every sensitivity analysis or the full adverse-event tables. Those limits should remain visible when interpreting the headline.
Common side effects are only part of safety
Gastrointestinal events were the most common adverse events in both groups. Most were mild to moderate and occurred primarily while treatment levels were being increased.[1] The abstract does not provide detailed comparative event rates, so that description cannot establish that the treatments were equally safe or equally tolerable.
The official Wegovy and Zepbound labels warn about severe gastrointestinal reactions, acute pancreatitis, gallbladder disease and acute kidney injury associated with volume depletion.[2][3] Volume depletion means loss of body fluid; the labels describe kidney-injury reports occurring largely in people whose gastrointestinal symptoms led to dehydration.[2][3] “Mostly mild to moderate” at the trial level does not erase these clinically important possibilities.
Both labels also carry boxed warnings about thyroid C-cell tumors observed in rodents. Whether the drugs cause these tumors in humans remains unknown; rodent findings are not proof of human cancer. Both are contraindicated with a personal or family history of medullary thyroid carcinoma or Multiple Endocrine Neoplasia syndrome type 2.[2][3]
A research vial does not inherit the trial result
SURMOUNT-5 tested defined treatments, not anything sold under the same molecule name.[1] FDA states that unapproved versions do not undergo its premarket review for safety, effectiveness and quality. It has received reports of dosing errors involving compounded injectable semaglutide, including cases requiring hospitalization, and warns about products sold for human use under false research labels.[4]
Laboratory research materials should not be consumed or self-administered. Even documented identity or purity would not establish a finished formulation's suitability for human use, clinical effectiveness or safety. Reconstitution does not supply that missing evidence. This distinction does not require assuming every reagent is contaminated or untested: analytical characterization and suitability as a medicine answer different questions. A trial result belongs to the treatments and conditions studied, not automatically to a laboratory vial.
Source note: This article uses the original SURMOUNT-5 abstract, not its full paper; relevant mechanism and warning sections of current Wegovy and Zepbound prescribing information; and FDA safety information retrieved October 23, 2026. Abstract-level access limits assessment of detailed methods and adverse-event tables. This is editorial source review, not medical review or testing by Chudik Scientific.