FDA's HDE Has Two Gates — and the Annual Distribution Number Is the One Teams Miss

FDA can grant a Humanitarian Device Exemption only after a separate office has granted something else first. The Office of Orphan Products Development issues a Humanitarian Use Device designation under its own review; the HDE itself is a distinct marketing application reviewed under 21 CFR Part 814, Subpart H. Teams that treat the HUD letter as the finish line discover, often mid-application, that the review answering the real commercial question — how many units can ever be sold for a profit — has not started yet.

Two gates, not one

The sequence is easy to compress in planning documents and hard to compress in practice. A Humanitarian Use Device determination establishes that a device is intended to treat or diagnose a condition affecting no more than 8,000 individuals in the United States per year — a threshold question FDA's Office of Orphan Products Development answers on its own timeline, separate from any marketing application. Only once that determination exists can a sponsor file the HDE itself, which Subpart H reviews on a different standard entirely: not population size, but whether the probable benefit to health outweighs the risk of illness or injury, supported by a smaller evidentiary record than a PMA requires. Filing the HDE application before the HUD determination is final is a common sequencing error that stalls review rather than accelerating it.

The ADN is not a market-size estimate

FDA's HDE guidance describes the Annual Distribution Number as the number of devices reasonably necessary to treat or diagnose eligible patients in a year, calculated by multiplying the number of devices one patient needs per year by the population cap. A device used twice per patient per year against the 8,000-person ceiling yields an ADN of 16,000 — a figure that has nothing to do with how large the addressable market actually is, and everything to do with keeping distribution proportionate to the rare population the designation was built around. Distributing beyond the ADN in a given year does not make the device illegal to sell; it means the units above the ADN cannot be sold for a profit for the rest of that year.

• Treating the ADN as a sales forecast. It is a regulatory ceiling on profit-eligible volume, not a commercial projection — finance teams that model it as the latter overstate addressable revenue.
• Calculating it once and leaving it. A change to the indication, the dosing regimen, or the number of devices used per patient changes the ADN math and may require a recalculation with FDA.
• Assuming it caps distribution. It caps the units eligible to be sold for a profit — a sponsor can still distribute beyond the ADN, just not profitably.
• Setting it per company instead of per device. The ADN is calculated for the specific HUD, not for a sponsor's broader rare-disease portfolio.

An Annual Distribution Number is not a license to project revenue. It is the arithmetic FDA uses to keep a device's reach proportionate to the rare population it was designated to serve. Why the ADN has to be recalculated, not just filed once

Profit is prohibited by default

Section 520(m) of the FD&C Act starts from a default prohibition: an HDE-approved device cannot be sold for a profit. The statutory exception, created by the Pediatric Medical Device Safety and Improvement Act and carried forward since, applies only where the device treats a disease or condition in pediatric patients or a defined pediatric subpopulation and is labeled for them, or where it treats an adult-only disease that does not occur in pediatric patients, or occurs in numbers that make pediatric device development impossible, highly impracticable, or unsafe. A sponsor assuming HDE approval automatically opens a profit-eligible channel is working from the wrong default; the exception has to be established, not presumed, and it narrows who can build a commercial case around a rare-population device at all.

None of this is unusual as device regulatory work goes — it is simply sequenced differently from a standard premarket pathway, and treating the HUD letter, the HDE approval, and the ADN as one continuous event rather than three distinct gates is where timelines and revenue assumptions both go wrong. For devices where the rare-population route turns out not to fit, De Novo classification is the pathway most often considered instead, and sponsors building a wider rare-disease program strategy or a device with a pediatric-labeled indication need the ADN and profit analysis built in from the start, not reconciled after HDE approval.