The Trial Ended. Who Made It Into the Open-Label Extension?
November 4, 2026 · 5 min read · Aqrab editorial team
A randomized trial ends, and participants are invited to keep receiving the study treatment in an open-label extension. Years later, the extension reports reassuring outcomes. Before interpreting them, trace who entered, who stayed, and who supplied the outcome. The extension may add useful long-term exposure and safety observations, but its participants are no longer automatically the randomized group.
Follow the Denominator, Not Just the Follow-Up Time
Consider this entirely fictional flow. It illustrates missing information, not the performance of any treatment.
| Stage | People | What is known? |
|---|---|---|
| Randomized in parent trial | 100 | Original allocation known |
| Entered extension | 60 | 40 did not enter |
| Had long-term outcome recorded | 45 | 15 more have no endpoint here |
| Met a response definition | 36 | 36/45 = 80% of those assessed |
The 80% describes the assessed subset. It does not establish an 80% long-term response among the 100 originally randomized, because outcomes for 55 people are not shown. Nor is 36/100 an estimate of that response rate: it is only the fraction with a documented response in this fictional flow. Why people did not enter or remain matters, especially if tolerability or earlier benefit influenced continuation.
The practical takeaway
For every long-term percentage, write its numerator and denominator beside the original randomized count. Then ask what happened to everyone between those counts.
Four Questions Before Repeating the Claim
- Who was eligible and who enrolled? Ask for a flow from each original treatment arm, including reasons for non-entry.
- Who remained under observation? Report withdrawal, loss to follow-up, and outcome availability by time point; describe reasons when known.
- What comparison survives? If all participants receive the same treatment without a concurrent randomized control, a within-extension outcome is not a randomized long-term treatment effect.
- What can safety data still tell us? Longer exposure can reveal potential adverse-event signals. Event rates still require exposure time and careful interpretation of dropout, visit schedules, and the absence of a concurrent control.
An extension is not useless because it is uncontrolled. It answers a narrower question. Keep signal detection and observed outcomes separate from a claim that treatment caused a long-term benefit or is safer than an alternative.
Where Aqrab Fits
Use Aqrab to organize a methods critique around the parent-trial population, extension entry, missing outcomes, and the comparison actually available. It cannot recover outcomes that were never collected or turn an uncontrolled extension into a randomized contrast.
Source Anchors
The FDA Standard Safety Tables and Figures guide, section 2.6 describes the signal-detection value and interpretive limits of open-label extensions. Hemming and colleagues analyze patient-selection bias, while Bowers and colleagues document variation in extension design and retention. The numerical flow above is invented for teaching and is not drawn from these sources.
Keep reading
Don't stop at one method.
Good methods judgment comes from contrast. Read the neighboring guides, see where the assumptions diverge, and avoid treating every observational problem like it needs the same hammer.
Twenty Patients, Thirty Events: Read the Harms Denominator
A worked trial example separates the proportion of patients harmed from recurrent events per person-time, then shows why unequal follow-up needs more than a rate correction.
A 12-Month Benefit and a 30-Day Harm Are Not One Net-Benefit Number
A fictional trial timeline shows why benefit and harm counts measured over different windows cannot be silently collapsed into one net-benefit claim.
Clinical Importance: When a Significant Result Is Still Too Small to Matter
A practical guide to reading confidence intervals against a prespecified clinical-importance threshold. Separate evidence against the null from evidence that an effect is large enough to matter.
This is the newest guide so far.