Considering Old and New 510(k) Predicates

The 510(k) route to market for Class II devices involves the comparison of the candidate device to a legally marketable (already cleared) device. Note that the predicate device need not actually be on the market. There are several guidances on doing this comparison. At one time it was possible to use more than one predicate device and borrow attributes from each. For example, a mesh used in the groin could be combined with a mesh designed for tissue ingrowth for use in another part of the body and a combination then created for use in abdominal surgery. This “split predicate” approach was discouraged via a 2014 guidance.

Since a new device need not be identical to its predicate, from the early days following the Medical Device Amendments of 1976 there was concern about “device creep” in which through a series of changes and 510(k)s a device could become quite unlike its original predicate. Creep could be so extensive that the end device could approach needing a PMA, especially if there had been an attempt to bring all of the changes to market at once which would catch the FDA’s attention. This could lead to a regulatory strategy of incremental change that does not raise red flags at the agency. PMA Supplements offer the same opportunity to create a series of changes in order to reach a new design that might not have been approved if all the changes were implemented at once.

A different perspective on predicates and change, and perhaps a lack of possibly desirable cumulative creep, was announced by FDA on November 27, 2018 and addressed here. In this announcement FDA Commissioner  Scott Gottlieb raised the issue of disallowing the use of old predicates. This optimistically assumed that newer devices are necessarily better than older ones, and therefore a comparison to a newer cleared device would necessarily be beneficial. A side issue here is the irony that an older device might still be on the market, and therefore presumably still judged to be safe and effective, yet not safe enough to be a predicate. The commissioner also proposed to create a web listing of devices using predicates that were more than 10 years old in order to damn them by inference.

We might pause here to remember Class II devices that have been recalled or withdrawn from the market because it turned out they were not as good as their predicates. It is also possible to use the lower safety of a newer device in a regulatory strategy. Suppose a device has been cleared against an older predicate and then found to not live up to expectations that it was actually better than the predicate. Rather than admit this problem, a manufacturer might use the newer but lower-performing device as its next predicate in order to avoid having to compare to the better performance of the original predicate.

It is a nice assumption that devices get continually better, and that therefore new predicates would always be preferred. But unfortunately, this is not always the case.