Regulatory stop signals

Regulatory clinical trial stops are rare and not the same as safety failures

Only 91 of 23,617 stopped records (0.4%) are classified as regulatory. They deserve source review rather than being folded into safety or biological failure counts.

2026-08-207 min readKeyword: regulatory clinical trial stops
Five facts from the dataset
  • The current database contains 91 regulatory-classified records, representing 0.4% of all stopped records.
  • 46 regulatory records are terminated, 40 are withdrawn, and 5 are suspended.
  • Phase II is the largest regulatory phase group with 28 records.
  • Other is the largest disease-area slice with 34 regulatory records.
  • Regulatory identifies a decision context and should not be converted into a safety or biological-failure claim without supporting source evidence.

A small category with outsized ambiguity

Regulatory language attracts attention because it can suggest a clinical hold, an authority request, an approval issue, or another formal intervention. Yet only 91 of 23,617 stopped records (0.4%) are classified as regulatory in the current database.

That small count makes careful interpretation more important, not less. A regulatory action can arise from safety, manufacturing, documentation, protocol, compliance, or information requirements. The label identifies the decision context; it does not by itself establish biological failure.

Regulatory and safety answer different questions

Safety describes the substantive concern identified in the stop language: adverse events, toxicity, tolerability, or benefit-risk. Regulatory describes the role of an authority or formal regulatory process in the pause or stop. A record can involve both ideas, but a single analytical bucket should reflect the strongest explicit evidence available.

Collapsing every regulatory stop into safety would inflate safety counts and erase cases driven by nonclinical, quality, procedural, or documentation issues. It would also imply more certainty about causation than the registry text may provide.

The status distribution is unusually balanced

Regulatory records are distributed relatively evenly between terminated (46) and withdrawn (40) studies, with 5 suspended records. This is a reminder that regulatory language does not map neatly to one trial status.

A withdrawn study may never begin enrollment, a suspended study may potentially resume, and a terminated study has stopped early. The regulatory context and the study status need to be read together.

Phase II is the largest phase slice

Phase II contributes the largest phase group among regulatory-classified records, followed by Phase I and Phase III. Those counts do not establish that one phase is more exposed to regulatory intervention because the database does not contain the denominator of all trials in each phase.

They are useful for triage. A Phase I regulatory stop may direct attention toward dose, early safety, manufacturing, or protocol issues. A later-stage stop may require review of endpoint, benefit-risk, authority correspondence, or program-level decisions.

What to verify in the primary sources

I would look for the exact authority action, the date it occurred, whether the hold was full or partial, the sponsor's stated cause, and whether the action was later lifted. ClinicalTrials.gov may not contain the entire regulatory history.

FDA notices, sponsor filings, trial updates, regulator databases, and subsequent protocol changes can materially change the interpretation. If the public record only says regulatory reasons, the analysis should not invent a safety mechanism.

The responsible conclusion

Regulatory stops are important signals for investigation, but weak standalone evidence for a claim that the drug failed. Their low frequency also means that isolated records or sponsor counts can be misleading without context.

The right wording names the registered status and regulatory classification, then describes the documented reason with its source. Safety, efficacy, and biological failure should be added only when the underlying evidence supports them.

Regulatory records by status

Trial statusRegulatory records
Terminated46
Withdrawn40
Suspended5

Largest regulatory phase groups

PhaseRegulatory records
Phase II28
Phase I19
Phase III19
Phase I + Phase II8
Phase IV6
Unknown5
Early Phase I3
Phase II + Phase III3

Largest regulatory disease-area slices

Disease areaRegulatory records
Other34
Oncology16
Infectious Disease14
Gastroenterology & Hepatology6
Cardiovascular5
Endocrine & Metabolic5
Neurology4
Renal & Urology2

FAQ

Is a regulatory clinical trial stop automatically a safety failure?

No. Regulatory actions can involve safety, but they may also involve manufacturing, protocol, documentation, compliance, or other requirements. The primary source must establish the reason.

Are regulatory stops common in the database?

No. The current database contains 91 regulatory-classified records, representing 0.4% of 23,617 stopped trials.

Can a regulatory hold be lifted?

Yes. Some regulatory pauses can be resolved. Researchers should verify the current trial status and subsequent authority or sponsor updates rather than relying on an older snapshot.

Source note: counts are generated from the current ClinicalTrials.gov-derived stopped-trial dataset used by ClinicalTrialFailures.com. These labels are analytical screening signals, not medical advice.