MDR Article 70: Why Your Device Class Sets the Start Date

Most sponsors planning an EU clinical investigation can recite the headline numbers: a 10-day validation check, then 45 days for the competent authority to authorise. Far fewer can say which clock actually controls their enrollment date — and under MDR Article 70, those are frequently not the same clock. For a large share of investigational devices, the real start date lands ten days after submission, not forty-five. For others, it does not land until the competent authority explicitly says so — and in at least one major Member State, the EU-wide default for which devices get the shortcut does not apply at all.

The two clocks Article 70 actually runs

The first clock is short and procedural. Once a sponsor submits a clinical investigation application, the Member State concerned has 10 days to tell the sponsor whether the investigation falls within the Regulation's scope and whether the application dossier is complete. If it is not, the sponsor gets a window to respond or complete the file, which the Member State may extend. Only once that check clears does the application receive a validation date — and that date, not the submission date, is what every later deadline counts from.

Why the 45-day number is the wrong one for most sponsors

The second clock is the one sponsors build their site-activation plans around: up to 45 days from the validation date for the competent authority to authorise the investigation, extendable by a further 20 days where the authority consults experts. Treating that as the earliest possible start date is the default assumption — and for a meaningful share of investigational devices, it is simply wrong. Article 70 lets a sponsor start investigating a Class I device, or a non-invasive Class IIa or IIb device, immediately after the validation date, without waiting for the authorisation decision at all, unless national law provides otherwise. The 45-day clock, for those devices, is not a gate on enrollment; it is a review period that runs in parallel with a study that is already open.

• Class I and non-invasive Class IIa/IIb devices. May start immediately after the validation date under the MDR's own text — the competent authority's 45-day review continues in the background rather than gating site activation.
• Invasive and higher-risk devices. Get no such carve-out. The sponsor waits for the competent authority's explicit authorisation, so the 45 (or 65) days are a real, binding floor on the start date.
• National deviations. BfArM has stated plainly that in Germany, Class I and non-invasive Class IIa devices may not start immediately after validation, and non-invasive Class IIb devices follow the full authorisation procedure rather than the abridged route — a documented departure from the MDR's EU-wide default that a multi-country protocol cannot ignore.
• Ethics committee review. Runs on its own, nationally governed timeline that Article 70 does not touch. A device that clears the regulatory start date early can still be blocked from enrolling by an ethics committee that has not yet opined.

The 45-day authorisation window is a real deadline for some devices and a background process for others. Building one enrollment date for every site in a multi-country study treats both as the same clock. They are not. Why the device-class read has to come first

Where the time actually gets lost

In practice, few investigations lose time to the 45-day assessment itself. They lose it earlier, at validation, or later, to a request for information that pauses the authorisation clock mid-review. An incomplete Annex XV dossier does not fail outright — it reopens a response window that pushes the validation date later, which pushes every subsequent deadline later with it, including the device-class early-start date the sponsor may have been counting on. And once the Member State is in its assessment window, any request for additional information suspends that 45-day clock entirely until the sponsor answers; a two-week delay in pulling together a response is a two-week delay added straight to the calendar, not absorbed inside a buffer that was never built in. The practical lesson sits upstream of both clocks: a dossier built to clear clinical evaluation and completeness review on the first pass is worth more than any argument about which 45-day extension applies. Sponsors already managing the Article 75 clock for substantial modifications mid-study should treat Article 70 as the same discipline applied one step earlier — get the clock-starting event right, and the rest of the schedule follows from it.

None of this requires second-guessing the regulation — it requires reading past the headline 10-and-45 framing to the device-class carve-out underneath it, and then checking that carve-out against how each Member State where you plan to enroll has actually implemented it. Our European regulatory strategy work builds that device-by-device, country-by-country timeline before the application goes in, so the start date in the protocol is the start date the competent authorities will actually honor. If your medical device clinical investigation plan currently assumes one authorisation clock for every site, that is the assumption worth testing first.