Clinical Trial Registration: What ICMJE and Journals Actually Require Before They'll Consider Your Manuscript

ICMJE's own Recommendations state it directly: registration must occur at or before the time of first patient consent for enrollment, as a condition of consideration for publication. This is what "prospective registration" means. It is not about when you submit the manuscript. It is about when the first participant agreed to take part.

The rule has applied since 2005: trials that began enrolling on or after July 1, 2005 must have been registered prospectively to be eligible for ICMJE-following journals. For trials that began enrolling before that date, ICMJE journals will accept retrospective registration, meaning the registry entry was created after enrollment had already started.

ICMJE accepts registration in any WHO International Clinical Trials Registry Platform (ICTRP) primary registry, or in ClinicalTrials.gov, which is itself a WHO ICTRP data provider. The registry has to be publicly accessible at no charge, and it has to capture the WHO Trial Registration Data Set (currently 24 items, version 1.3.1) before the first participant is enrolled, not added retroactively once someone asks for it.

Source: WHO International Clinical Trials Registry Platform, primary registry network (accessed September 2026). This is a sample; WHO's primary registries page lists the full network.

If your trial only needs to satisfy a journal's ICMJE-aligned policy, registering once, in whichever primary registry fits your country or funder, is enough. You do not need to register the same trial twice.

Section 801 of the FDA Amendments Act is easy to conflate with the ICMJE journal policy, but it is a different obligation with a different enforcer. FDAAA 801 requires certain "applicable clinical trials" of drugs (including biologics) and devices to register on ClinicalTrials.gov, generally within 21 days of enrolling the first participant, and to submit summary results, typically within one year of the trial's primary completion date. This applies whether or not the trial is ever submitted to a journal.

Noncompliance is enforced by the FDA, not an editor. After a Notice of Noncompliance, the responsible party has 30 calendar days to correct the violation before the FDA can seek civil monetary penalties of up to $10,000 per day the violation continues, under the statutory cap (the FDA adjusts this figure for inflation in its current regulations). If your trial is an applicable clinical trial under FDAAA 801, meeting the ICMJE journal requirement does not automatically satisfy the federal one, and the reverse is also true. Check both separately.

Source: ICMJE Recommendations and ClinicalTrials.gov / FDA civil money penalty guidance (accessed September 2026).

A registry entry is not "done" just because it exists. The WHO Trial Registration Data Set requires 24 items before a trial counts as registered, including (not an exhaustive list):

• primary registry name and the trial's identifying number
• date of registration
• secondary identifying numbers, if any
• source of funding and the responsible party
• the health condition or problem studied
• the intervention and comparator, described specifically enough to distinguish arms
• key inclusion and exclusion criteria
• study type (interventional, observational subtype where relevant)
• anticipated start date and target sample size
• recruitment status
• the primary outcome and its timepoint
• key secondary outcomes

For trials that began enrolling participants on or after January 1, 2019, ICMJE also requires a data sharing plan to be stated in the registration itself, not just in the eventual manuscript.

Registering the trial is step one. Reflecting it correctly in the manuscript is step two, and it is where editors actually check your work.

Registration number in the abstract. ICMJE recommends that the registration number and registry name appear at the end of the abstract. Separately, authors should give the registration number the first time they use a trial acronym anywhere in the manuscript, including for other trials they reference in the Discussion.

The registered and reported outcomes have to match. Editors compare the primary outcome in the registry against the primary outcome in the manuscript. If they differ without an explanation, that reads as selective outcome reporting, one of the exact problems registration exists to catch. Any post-hoc change to the outcome or analysis plan needs to be disclosed and justified in the manuscript, not left for a reviewer to notice.

The reporting checklist item is not the same as the registry entry. CONSORT's trial registration reporting item asks for the registry name, the identifying number with a URL, and the date of registration, stated in the manuscript itself, matching what the registry shows. For the full CONSORT 2025 checklist and how it applies to trial manuscripts beyond registration, see pre-submission review for clinical trial manuscripts.

This is the situation that generates the most anxious searches, so here is the decision logic rather than a vague warning.

Your trial began enrolling before July 1, 2005. ICMJE journals accept retrospective registration for these trials without treating it as an exception. This is now a narrow and shrinking category of active submissions.

Your trial began enrolling on or after July 1, 2005, and it was registered after enrollment started. ICMJE's own language is specific here: exceptions to the prospective rule should be rare, and if an editor allows one, the published article must state when registration was completed and why it was delayed, and the journal should publish a statement explaining why the exception was granted.

That is not a rubber stamp. Some journals will not grant the exception at all and will treat retrospective registration as grounds for rejection without external review, which is within their editorial discretion. Others will consider it case by case.

What this means practically: do not assume your journal will accept a retrospective registration, and do not wait for an editor to ask about it. Check the specific target journal's instructions for authors before you submit, and if you cannot find a clear answer, draft the explanation (when registration was completed, why it was delayed) as part of your cover letter rather than hoping the gap goes unnoticed. An editor who has to ask why a trial was registered late is already primed to be skeptical of the rest of the manuscript.

In our pre-submission review work, registration problems rarely show up as a missing registry entry. They show up as a mismatch that the authors did not notice because they were looking at the manuscript and the registry as two separate documents. The registered primary outcome says one thing and the results section reports a different one, with no explanation anywhere in the text.

The registry lists five secondary outcomes and the manuscript reports twelve, several of which read like they were added after the data came in. The registration date sits a few weeks after the stated enrollment start date, close enough that it could be a rounding error in the manuscript or could be a real retrospective registration nobody flagged.

None of these are necessarily evidence of misconduct. They are usually oversights: a protocol amendment that never made it back to the registry, or an outcome that got renamed for clarity without anyone checking that the change was documented. But an editor cannot tell the difference between an oversight and selective reporting from the manuscript alone, and by policy they are supposed to assume the more skeptical reading until it is explained.

A manuscript readiness check is useful here specifically because registration mismatches are easy to miss when you already know what you meant. The free readiness scan and the $39 Full Review review methodology and citations, but neither replaces comparing your registry entry line by line against your manuscript. That comparison is on you, or on a co-author who was not involved in writing the outcomes section, before you submit.