The short answer
A non-inferiority trial asks whether a new treatment is not worse than a comparator by more than a pre-specified margin. Failing to demonstrate that is a statement about where a confidence interval fell. It is not the same as demonstrating the treatment is inferior.
Key facts
- Question asked
- Is it not worse by more than margin X?
- Set in advance
- The margin, before any data
- Decided by
- Where the confidence interval falls
- Failing means
- Non-inferiority was not demonstrated
- Failing does not mean
- Inferiority was demonstrated
- Live example
- REDEFINE 4 (NCT06131437)
Why the design exists at all
Superiority trials ask whether a treatment beats a comparator. Sometimes that is the wrong question. A new option may be worth having because it is easier to take, better tolerated or cheaper, provided it is not meaningfully less effective. Non-inferiority formalises not meaningfully less effective into a number agreed before the trial runs.
The margin is the whole design
A non-inferiority margin is the largest difference that would still be considered acceptable. It has to be justified and fixed in advance, because a margin chosen after seeing the data could be set wherever it produces the desired conclusion. Everything about how such a trial reads depends on that number, and it is the first thing to look for.
Research material referenced
Retatrutide 10mg, third-party HPLC tested
What the result actually is
A confidence interval around the difference between treatments. If the whole interval sits within the margin, non-inferiority is demonstrated. If part of it falls outside, it is not. That is a statement about precision and position, not a verdict on the compound, and a trial can miss the threshold while still showing the treatment worked well.
The asymmetry most coverage misses
Three outcomes are possible, not two. Non-inferiority demonstrated. Non-inferiority not demonstrated. Or inferiority actually demonstrated, which requires the entire interval to sit on the wrong side. Reporting that lands on failed and therefore worse collapses the middle case into the third, and the middle case is where most failed non-inferiority trials land.
REDEFINE 4 as the live illustration
The ClinicalTrials.gov record for NCT06131437 states the primary outcome as confirming non-inferiority of CagriSema versus tirzepatide on relative change in body weight. That objective was not met. The reported figures were 23.0% against 25.5% on the efficacy estimand, a real difference, and also a result in which the smaller number is a large weight reduction.
How to read one properly
Find the pre-specified margin. Find the confidence interval. Check which of the three outcomes actually occurred. Then read the absolute numbers separately, because a compound can miss a comparative threshold and still be effective. Those are two different questions and a headline usually answers neither.
Quick reference
| Where the confidence interval sits | Conclusion |
|---|---|
| Entirely within the margin | Non-inferiority demonstrated |
| Crosses the margin | Not demonstrated (verdict open) |
| Entirely beyond the margin | Inferiority demonstrated |
Frequently asked questions
Does failing non-inferiority mean the drug is worse?
No. It means non-inferiority was not demonstrated. Showing inferiority requires the entire confidence interval to fall beyond the margin.
Why not just run a superiority trial?
Because a new option can be worth having for tolerability, convenience or cost if it is not meaningfully less effective, which is what non-inferiority tests.
What should I look for first?
The pre-specified margin. It must be fixed before the data, and everything about how the result reads depends on it.
Continue researching
Related guides and reference material.
Primary sources & clinical trials
Published papers, trial registrations, regulator pages and company announcements. Links open the original.
Written and published by UK Peptides, the trading name of AURYXRESEARCH LTD (company 17388727, registered in England and Wales).
Questions or a correction? Email hello@ukpeptides.org or call +44 7426 987848.
More GLP-1 & Incretin Science articles
- Satiety and Nausea Run Through Different CircuitsA 2024 Nature paper showed hindbrain GLP-1 receptor circuits for satiety and aversion are dissociable. Nausea may not be the price of appetite suppression.
- Three Head-to-Heads, One Common ComparatorSURMOUNT-5, REDEFINE 4 and TRIUMPH-5 all use tirzepatide. What it means when one compound becomes the field's reference point.
- What Open-Label Does to a Weight-Loss TrialREDEFINE 4's registry record lists masking as none. Why an unblinded weight trial is weaker evidence, and where the effect creeps in.
- Two Answers to the Same QuestionCagriSema pairs GLP-1 with amylin; tirzepatide pairs it with GIP. REDEFINE 4 was effectively the first direct test between those strategies.
- Semaglutide and Liver Disease: What ESSENCE TestedA 1,205-participant Phase 3 in metabolic dysfunction-associated steatohepatitis, published in NEJM in June 2025. What it measured and why liver trials are hard.