Reading the evidence
How to Read a ClinicalTrials.gov Record: A Field-by-Field Guide
AR
· 10 min read
A summary of published research, with citations. It states no amount, schedule or route, and makes no claim about what this compound does in a person. Supplied for laboratory research use only.
On this page · 11 sections
- Why are trials registered at all?
- The worked example: TRIUMPH-4
- What does the status field mean?
- Phase, and what it does not tell you
- Primary and secondary outcomes: the fields that matter most
- Estimated versus actual enrolment
- Results posted versus results published
- Sponsor, collaborators and responsible party
- Version history
- A quick checklist
- Frequently asked questions
A ClinicalTrials.gov record is a trial's plan, registered before results exist, plus updates as the trial runs. To read one, check the status and its verification date, the phase, the pre-specified primary and secondary outcomes, whether enrolment is estimated or actual, whether results are posted on the registry or only published elsewhere, who the sponsor is, and what changed between versions. Comparing the record with the paper is how outcome switching and unreported trials get caught.

Why are trials registered at all?
Because the published literature alone has repeatedly been shown to be a biased sample of the research actually done. In a cohort of Danish trials, 62 percent had at least one primary outcome that was changed, introduced or left out between the protocol and the published paper. Statistically significant outcomes were more likely to be reported in full: the pooled odds ratio was 2.4 for efficacy outcomes and 4.7 for harms (PMID 15161896). A systematic review of such cohort studies found that 40 to 62 percent of trials had a primary outcome changed, introduced or omitted (PMID 18769481).
The antidepressant literature is the best-known example of a whole drug class distorted this way. Of 74 trials registered with the FDA, 31 percent were never published. The published literature made 94 percent of trials look positive, while the FDA's own analysis found 51 percent positive. Effect sizes in the journals were inflated by 32 percent overall (PMID 18199864). Registration addresses this by fixing the plan in public before the results can influence what gets reported. How to read a peptide study covers pre-specification from the methods side.
The worked example: TRIUMPH-4
The record used here is NCT05931367, TRIUMPH-4. It is a phase 3 trial of retatrutide, an investigational triple agonist that no regulator has approved, in adults with obesity or overweight and knee osteoarthritis. The data below comes from the ClinicalTrials.gov API, retrieved in October 2026. Retatrutide's phase 3 TRIUMPH trials covers the whole programme, and gray-market retatrutide explains why its trial results do not transfer to unregulated products sold under the same name.
| Field | What TRIUMPH-4's record says | What to read into it |
|---|---|---|
| Overall status | Completed | Data collection has ended. It says nothing about results |
| Status verified | January 2026 | When the sponsor last confirmed the record |
| Phase | Phase 3 | The confirmatory stage, not approval |
| Design | Randomised, parallel, double-blind (participant and investigator), placebo-controlled | Compare with what the paper reports |
| Enrolment | 445, actual | The number actually enrolled, not the target |
| Primary outcomes | WOMAC pain subscale change; percent change in body weight; both at week 68 | Two co-primary endpoints, fixed in advance |
| Secondary outcomes | Eleven, including WOMAC physical function, waist circumference, lipids, blood pressure and pain intensity scores | Supporting endpoints, not the main test |
| Arms | Two retatrutide arms and a placebo arm | Arms are labelled by number; doses are not shown here |
| Sponsor | Eli Lilly and Company (industry), also the responsible party | Who designed, paid for and reports the trial |
| Dates | Start Aug 2023; primary completion and completion Nov 2025, both actual | The clock for reporting results starts at primary completion |
| Results | None posted | No paper on PubMed by October 2026 either |
| Other IDs | Sponsor protocol code and an EU trial number | Use them to find the same trial in other registries |
What does the status field mean?
Status describes recruitment and data collection, not success. The common values are "not yet recruiting", "recruiting", "active, not recruiting" (participants are still being followed), "completed", "terminated" (stopped early), "withdrawn" (stopped before anyone enrolled) and "suspended". "Unknown" means the record has not been verified within the registry's required window.
Always read status together with the "status verified" date. A trial listed as active whose record was last verified years ago tells you very little. For example, NCT02637284, a phase 1 study of a BPC-157 product registered in 2015, shows "unknown" status, last verified in October 2015, with "active, not recruiting" as its last known status and estimated completion dates in 2016. Nothing in the record says whether it ran, finished or reported.
If a trial was terminated, read the stated reason. "Business decision" and "futility" mean very different things.
Phase, and what it does not tell you
The phase field tells you what question the trial was designed to answer: first-in-human safety, dose finding, or confirmation in a larger population. It does not tell you the drug works or is approved. Clinical trial phases explained covers what each phase is for. Many records also list "not applicable", which is used for trials that do not test a drug, such as device, behavioural or supplement studies.
Primary and secondary outcomes: the fields that matter most
The outcome fields are the reason the registry exists. Note exactly what was named as primary, how it was measured and at what time point, then check the paper against it. TRIUMPH-4 names two primary outcomes at week 68. A paper that led with a different endpoint, a different time point or a subgroup would need explaining.
Three things to watch for:
- Vague outcomes. An analysis of the registry found that 61 percent of a sample of outcome measures did not say clearly which metric the planned analysis would use (PMID 21366476). "Change in pain" is less binding than "change from baseline in WOMAC pain subscale at week 68".
- Many primary outcomes. The same analysis found that 20 percent of results records listed more than two primary outcomes. Each extra primary outcome gives the trial another chance to report a win.
- Primary outcomes that change. If the primary outcome changed after the trial started, the version history shows when. A change made after results could have been seen is the pattern registration is meant to expose.
Estimated versus actual enrolment
Every enrolment number has a label. "Estimated" is a target, set while the trial is planned or recruiting. "Actual" is what happened. TRIUMPH-4 shows 445, actual. The retatrutide cardiovascular outcomes trial, NCT06383390, still showed 10,000, estimated, in October 2026, because it had not finished. A trial whose actual enrolment fell far short of its estimate may have lacked the statistical power it was designed for.
Small differences between the registry and the paper are normal, but reconcile them anyway. TRIUMPH-1 (NCT05929066) lists 2,335 participants, actual. Its New England Journal of Medicine paper reports 2,339 randomised (PMID 42814954). The two counts may have been taken at different points or defined differently, and a careful reader notes the difference rather than assuming either number is wrong.
Results posted versus results published
These are two separate events, and either can happen without the other.
Posted means summary results were entered into the registry in a fixed structure: participant flow, baseline characteristics, every pre-specified outcome, and adverse events. The phase 2 retatrutide obesity trial (NCT04881760) reached primary completion in May 2022 and posted results in September 2023. Its record links to the journal paper.
Published means a peer-reviewed paper exists. TRIUMPH-1 has a paper in the New England Journal of Medicine, linked from its record, but had no results posted on the registry as of October 2026. TRIUMPH-4 had neither.
Posted results have one advantage over papers: they include every pre-specified outcome and the adverse-event tables in a standard format, whether or not the paper discussed them. Reporting has often been late. An analysis of trials likely to be covered by US reporting law found that only 13.4 percent reported summary results within the required 12 months of completion, and 38.3 percent had reported by the end of the five-year study window (PMID 25760355). An earlier analysis found that 52 percent of a sample of results records had an associated publication within two years of posting (PMID 21366476).
Sponsor, collaborators and responsible party
The sponsor designs and funds the trial. The responsible party submits the record and the results. In industry trials, such as TRIUMPH-4, they are usually the same company. Sponsorship does not make a result wrong. But it does matter for how you read it: you should know who designed the outcome list and who controls when results appear. Check the "collaborators" field too, and compare the sponsor with the funding statement in the paper.
Version history
Each record keeps an archive of every submitted version, reached through the record's history tab on ClinicalTrials.gov. That archive is where outcome switching becomes visible. Compare the first version, posted before enrolment, with the version current when results appeared, and look at the dates on each change to the primary outcome, enrolment target or completion date. Changes to a completion date or enrolment number while recruiting are routine. A change to a primary outcome near or after primary completion is the one to examine.
A quick checklist
- Status, and the date it was verified
- Phase and design: randomisation, masking, control
- The primary outcome, word for word, with its time point
- Enrolment, and whether it is estimated or actual
- Primary completion date, and whether results are posted
- Linked publications, and whether they match the outcomes
- The history tab: what changed, and when
Nothing here is medical, dosing, or therapeutic guidance.
Frequently asked questions
What does "completed" mean on ClinicalTrials.gov? It means the trial has finished enrolling and collecting data. It says nothing about whether the treatment worked or whether results are available. TRIUMPH-4 has been listed as completed since November 2025, with no results posted on the registry and no paper found on PubMed by October 2026.
What is the difference between primary completion and study completion? Primary completion is when the last participant is assessed for the primary outcome. Study completion is when all data collection, including secondary and follow-up measures, ends. Results reporting obligations are usually counted from primary completion, so that date matters more for judging whether results are late.
Why would a completed trial have no results? The sponsor may be within a permitted reporting window, may have been granted a delay, may not be legally required to report, or may simply not have reported. Check the dates and any linked publications. A long gap after primary completion, with no results and no paper, is worth noting.
Can a trial change its primary outcome after registration? Yes, and sometimes for good reasons. But every change is kept in the record's version history with a date. Studies comparing protocols with papers found primary outcomes changed, introduced or omitted in 40 to 62 percent of trials, so check the history before trusting a paper's choice of headline outcome.
Selected Research References
- Chan AW, Hróbjartsson A, Haahr MT, et al. Empirical evidence for selective reporting of outcomes in randomized trials: comparison of protocols to published articles. JAMA. 2004;291(20):2457-2465. pubmed.ncbi.nlm.nih.gov/15161896
- Dwan K, Altman DG, Arnaiz JA, et al. Systematic review of the empirical evidence of study publication bias and outcome reporting bias. PLoS One. 2008;3(8):e3081. pubmed.ncbi.nlm.nih.gov/18769481
- Turner EH, Matthews AM, Linardatos E, et al. Selective publication of antidepressant trials and its influence on apparent efficacy. New England Journal of Medicine. 2008;358(3):252-260. pubmed.ncbi.nlm.nih.gov/18199864
- Zarin DA, Tse T, Williams RJ, et al. The ClinicalTrials.gov results database--update and key issues. New England Journal of Medicine. 2011;364(9):852-860. pubmed.ncbi.nlm.nih.gov/21366476
- Anderson ML, Chiswell K, Peterson ED, et al. Compliance with results reporting at ClinicalTrials.gov. New England Journal of Medicine. 2015;372(11):1031-1039. pubmed.ncbi.nlm.nih.gov/25760355
- Jastreboff AM, Kaplan LM, Davies MJ, et al. Retatrutide, a Triple Hormone Receptor Agonist, for Treatment of Obesity. New England Journal of Medicine. 2026. pubmed.ncbi.nlm.nih.gov/42814954
All products sold by NovaEvum are for research and laboratory use only. They are not intended for human or animal consumption, diagnosis, treatment, cure, or prevention of any disease.
