Continuing review is the IRB's own clinical hold

Two stop authorities over the same running study

A study can be free of any FDA action — no hold, no warning letter, no untoward finding — and still be unable to enroll a single new subject, because the local IRB's approval has lapsed. 56.109(a) gives an IRB the authority to approve, require modifications in, or disapprove research; that authority doesn't expire once granted, and it isn't a one-time vote the way an initial approval can feel to a sponsor filling out the file. It's an ongoing responsibility the IRB discharges on its own schedule, independent of anything CDRH is or isn't doing on the same protocol.

The IRB's clock is a ceiling, not a target

56.109(f) requires an IRB to conduct continuing review of covered research at intervals appropriate to the degree of risk, but not less than once per year. The once-a-year figure is the outer limit, not the default: an IRB reviewing a higher-risk protocol — a new device with limited human experience, a vulnerable subject population, a design that leans on investigator judgment mid-study — can and often does set a shorter interval on its own initiative. A file that assumes “annual” as a fixed cadence, rather than reading the actual date the IRB set at the last approval, will miss a review that's already overdue.

What happens when the year runs out unrenewed

FDA's 2012 guidance on IRB continuing review after clinical investigation approval addresses exactly the failure mode this creates: a study whose continuing review comes due and isn't completed in time. The guidance's expectation, consistent with how IRBs and sponsors have long treated a lapse, is that research activities have to stop when approval lapses — new enrollment, new procedures on existing subjects, anything beyond what's needed to protect subjects already in the study — until the IRB acts again. That's a harder stop than most trackers give it credit for, because nothing about it requires FDA to say a word.

A separate written record, from a separate authority

56.109(e) requires the IRB to notify the investigator and the institution in writing of its decision — approval, required modifications, or disapproval — and, on a disapproval, to state its reasons and give the investigator a chance to respond. That written determination sits in a different chain than FDA's own written basis for a 520(g)(8) hold: different author, different trigger, different addressee. A tracker that keeps one field for “regulatory status” and updates it whenever either party writes something has collapsed two authorities that don't move together and don't lapse or lift on the same terms.

The paper trail behind the date

56.115 requires the IRB and the institution to retain records of continuing review activities, correspondence between the IRB and investigators, and the minutes showing the vote and basis for each decision, for at least three years after the research completes — and 56.108 requires the IRB to have written procedures for how continuing review actually gets conducted, including how it determines which studies need more than the annual minimum. A continuing-review date on a tracker is only as good as the record behind it; an auditor asking to see the basis for a given interval is asking a question 56.108's procedures are supposed to answer in writing, not a question the sponsor's file can answer on the IRB's behalf.

Where this meets the file

A study-conduct file needs the IRB's continuing-review date carried as its own row — the interval the IRB actually set, not an assumed annual default — alongside FDA's hold status as a separate row entirely, because either one can stop the study without the other moving at all. A clinical-study tracking worksheet built around that separation is previewed in the launch catalog. If your program tracks IRB continuing review differently, the shelf takes that correction directly.