An accessory doesn't inherit the parent's class

This blog has already covered how a chosen predicate has to survive review on its own record, not just its outcome. Before the 21st Century Cures Act, a manufacturer building a genuinely new accessory — a display-only companion to an infusion pump, a mounting bracket for an imaging system — often reached for a predicate built on distant technological characteristics, or routed the accessory through a full PMA, for a reason that had nothing to do with the accessory's own risk: FDA's historical practice was to treat an accessory as carrying the same classification as whatever device it plugged into, parent class included. Section 3060(c) of the Cures Act ended that practice by amending Section 513(b) of the FD&C Act directly, and a device file still reasoning from the parent's own class to the accessory's has been arguing from a premise the statute no longer allows.

A statute that overrides the old default

Section 3060(c) of the 21st Century Cures Act, enacted December 13, 2016, amended Section 513(b) of the Federal Food, Drug, and Cosmetic Act to direct FDA to classify an accessory based on its own intended use, notwithstanding the classification of any other device with which the accessory is intended to be used. The word notwithstanding is the operative one: it doesn't just permit FDA to classify an accessory separately, it forecloses the older shortcut of reading the accessory's class off the parent device's own listing. FDA's guidance, Medical Device Accessories — Describing Accessories and Classification Pathways for New Accessory Types, describes what industry had treated as the prior norm plainly: absent a classification of its own, an accessory had generally taken on the classification, and the regulatory burden, of the highest-risk parent device it was intended to work with. A Class III parent meant a Class III accessory, whatever the accessory itself actually did.

The test is still risk — just the accessory's own

The Cures Act doesn't ask FDA to invent a new classification methodology for accessories. The guidance applies the same risk-and-controls test Section 513 already uses for any device — whether general controls alone provide a reasonable assurance of safety and effectiveness, whether special controls need to be added, or whether the device needs premarket approval — it just requires that test to run against the accessory's own performance and failure modes rather than the parent's. FDA states the consequence directly: an accessory to a Class III parent device may pose a lower risk that general controls, or general and special controls, can adequately mitigate, and can be regulated as Class I or Class II accordingly. A mounting accessory whose own worst failure mode is a dropped device presents a different risk than the life-sustaining function of the system it's built to hold, and the statute now requires that difference to reach the classification, not just the marketing story built around it.

Being an accessory is itself a threshold question

Before a device gets the benefit of its own classification test, it has to clear a definitional gate the guidance sets first: an accessory is a finished device intended to support, supplement, or augment the performance of one or more parent devices, a status determined from labeling, promotional claims, and other evidence of intended use. That test cuts both directions. A product marketed to complete a parent device's own intended use, rather than merely support it, is answering a harder question than the accessory pathway is built for. And a product that supports a device type in general, rather than one or more specific parent devices FDA can actually name, hasn't shown the guidance's own threshold question has been answered at all.

De Novo is FDA's own suggested door, not the only one

For an accessory type with no existing classification and no predicate to compare against, FDA's guidance recommends the De Novo classification process under Section 513(f)(2) as the route to a new classification, rather than treating the absence of a predicate as a reason to fall back on the parent device's own class. De Novo's own benefit-risk case, this blog has already covered, asks a sponsor to justify general controls, or general and special controls, as sufficient on the merits — exactly the showing an accessory's own classification now requires. A 510(k) built against a predicate accessory already classified in its own right remains available where one exists; what the statute closed off is treating the parent's own class as a predicate by proxy.

Where this meets the file

A device family's classification record needs the accessory's own intended-use statement and risk analysis carried as a field distinct from the parent device's own class, checked against Section 513(b)'s own test rather than copied down from whatever line the parent device occupies in the same file. A device-classification worksheet built around that split, alongside the classification tool this blog has already covered, is previewed in the launch catalog. If your program still classifies an accessory off its parent's own line, the shelf takes that correction directly.