This is a guest post by Luther Munford, Butler Snow LLP, who comes to us fresh from the pelvic mesh wars. He is grateful to Jin Yoshikawa of that firm for his help with this post. As always, our guest posters do their own work and thus deserve 100% of the credit (and any blame) for what they write.
As previously stated, Medtronic, Inc. v. Lohr, 518 U.S. 470 (1996), has nothing useful to say about the FDA’s current system for clearing medical devices. The 1982 world it describes disappeared decades ago. This author’s article in the Food and Drug Law Journal, L. Munford, “Courts v. FDA: A Lesson from Pelvic Mesh Litigation on Relative Competence to Decide a Legal Question,” 76 Food & Drug L.J. 6 (2021), provides the details, and it is hardly alone in its conclusions.
The reasons for Lohr’s irrelevance go deeper than the reason most commonly given, i.e., that the 1990 amendments to the Medical Device Act strengthened the 510(k) process and, for the first time, required the FDA to clear devices before they could be marketed.
The truth is that Congress never intended for the process described in Lohr to last, and it did not. From the very beginning, Congress created a process of device classification that drew a legal distinction between the temporary process at issue in Lohr and the permanent 510(k) process that came after it.
The device classification process is something Congress took seriously. Congress dictated the composition of medical panels that FDA used for the initial classification of device types into Classes I, II and III according to the risk they presented. The panels published their findings, with supporting citations, in the Federal Register. After opportunity for notice and comment, the FDA promulgated literally hundreds of regulations classifying device types. See 21 C.F.R. Parts 860-892. This process took place, and was completed, in the 15 years after the Medical Device Amendments (“MDA”) were enacted in 1976.
Of particular importance to modern device litigation is that the FDA’s placement of a device type in Class II meant that the agency determined, as a matter of law, that the device did not “present a potential unreasonable risk of illness or injury” that would require it to be in Class III. 21 U.S.C.§ 360c(a)(1)(C)(ii)(ll). That is a safety determination that applies to the entire type.
Once the FDA initially established device types, 510(k) clearance of a Class II device requires equivalence in safety and effectiveness to a predicate Class II device. When the FDA now makes a judgment as to equivalence to a predicate device, it examines what “significant changes” the manufacturer may have made, 21 C.F.R. § 807.81 (a)(3), and has before it not only the contents of the 510(k), but also the regulatory history of the predicate device or devices and the record for other devices of the same type. Clearance of the device is an individualized safety determination, just as a judge’s decision in a case based on case precedents is an individualized legal determination.
Lohr speaks to none of this. The device at issue in Lohr was cleared in 1982 pursuant to the transitional provisions in the statute that lasted only until devices were initially classified in Classes I or II, or met the standards for premarket approval in Class III. 21 U.S.C. § 360c(f)(1)(A)(i)(I). The device in Lohr was in the most risky class, Class III, which meant that the FDA had determined that it might present an unreasonable risk of illness or injury. Despite that classification, the statute allowed it to be cleared even before approval based on its equivalence to a device on the market in 1976, when MDA device regulation began. All the manufacturer had to do to be able to market such a device was to file a “510(k) notice.” The FDA reviewed the notices but did not then have to grant permission before the device could be sold. This procedure was temporary, intended by Congress to prevent the onset of device regulation from creating artificial monopolies.
The Lohr opinion held that this form of regulatory review did not impose any “requirement” that would preempt state tort “requirements.” 518 U.S. at 495. The Court emphasized the lax nature of the transitional clearance process Congress had created to handle the temporary problem of artificial monopolies. The court called the process “grandfathering” and said it simply required equivalence to a device sold in before 1976 that the FDA had never specifically reviewed for safety. Id. at 478. Clearance, Lohr stated, depended on “equivalence, not safety,” and, said the device had “never been formally reviewed under the MDA for safety or efficacy.” Id. at 493. For good measure, it cited evidence that the FDA review of 510(k) grandfathering notices on average took only 20 hours, and contrasted that with a much greater time needed for review of Class III approval applications, which required independent evidence of safety and effectiveness. Id. at 479.
But none of this has ever been true with respect to FDA clearance of non-grandfathered Class II devices, especially after 1990 when the Safe Medical Devices Act amended the statute to require FDA clearance before marketing. 21 U.S.C. § 360c(f)(1)(A)(i):
- When the FDA now clears a device as in Class II, there has been a formal safety review – twice. First, the congressionally required medical panel reviewed the safety of the device type as a whole, which the FDA endorsed after notice and comment rulemaking. That medical review necessarily found no potential unreasonable risk. There was no such determination in Lohr.
- Second, the manufacturer of any new device must obtain FDA permission before marketing that device by showing its safety and equivalence to a predicate Class II device − a process that involves another formal medical review. The FDA considers not only the 510(k) submission for that device, but also the regulatory history of the predicate device(s) and others of the same device type. Nothing even remotely like that happened in Lohr. Mere equivalence to a device on the market in 1976 is no longer sufficient.
- When all these steps are taken into account, it is obvious that the 20-hour estimate for the cursory transitional process at issue in Lohr is meaningless.
- As a matter of federal law, both before and after the SMDA, the classification of the device is an FDA determination of reasonable assurance of safety and effectiveness. The device in Lohr had never satisfied the approval requirements needed for the class to which it belonged, which in any event was the riskier class, Class III.
No court has ever rejected this analysis and in an early case [ed. note, this was Bexis’ case] the Fourth Circuit appeared to accept it. See Talley v. Danek, 179 F.3d 154, 161 (4th Cir. 1999) (reclassification into Class II meant spinal screws were “safe to market”).
But in the pelvic mesh cases, the MDL judge misinterpreted Lohr, looked only to Lohr’s discussion of the grandfathering process, and held that the FDA’s non-transitional clearance of those devices did not “go to” safety. Cisson v. C.R. Bard, Inc., 86 F. Supp.3d 510 (S.D. W. Va. 2015), aff’d sub nom. In re C.R. Bard, Inc., Pelvic Repair Sys. Prod. Liab. Litig., MDL No. 2187, 810 F.3d 913 (4th Cir. 2016). No mention was even made of those devices’ individual FDA placements in Class II. Other appellate courts have followed that decision. See e.g. Eghnayem v. Bos. Sci. Corp., 873 F.3d 1304, 1310-11 (11th Cir. 2017); Kaiser v. Johnson & Johnson, 947 F.3d 996 (7th Cir. 2020).
This is an error that needs to be corrected if, for no other reason, because it threatens to undermine the defense of future Class II medical device cases.
As discussed in the FDLJ article, these opinions are replete with error. They assume, based nothing more than the common presence of the numerals and letter “510(k),” that Lohr applies to all Class-II devices. That is not the case. Very few Lohr-type transitionally classified devices are still on the market. They opine that clearance of a Class II device could be based on a 1976 unregulated predicate. The law does not allow that. They act as if 510(k) is some kind of short cut the manufacturer can simply choose. It is not. The FDA makes that choice, acting on the advice of medical panels. The FDA, not the device manufacturer, decides which device types are eligible for 510(k) review.
And none of the decisions accurately describes what the FDA did to clear the devices in individual cases. The only opinion to address the classification process in detail, Kaiser, 947 F.3d at 1005-06, 1018, [ed. note, the Blog discussed Kaiser here] takes an FDA statement that the device type was too risky to be in Class I, chops out qualifying language, and turns it into a declaration that placement in Class II was not a finding of reasonable safety and effectiveness, which is not true and cannot, as a matter of law, be true.
To dig even deeper, these and other cases frequently and wrongly assume, based on the now-irrelevant hours of consideration comparison in Lohr, that a Class II “cleared” device is not as safe as a Class III “approved” device, simply because, for an individual device, the FDA approval process is more rigorous.
But the safety standard − the conclusion the FDA is required to make for each class − is identical. Statutorily, it is whether there is “reasonable assurance” of “safety and effectiveness.” 21 U.S.C. § 360c(a)(1)(A), (B), (C). That is what the statute says, and what the Supreme Court has affirmed, in FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 134 (2000): “Regardless of which category FDA chooses, there must be a ‘reasonable assurance of the safety and effectiveness of the device’” Id.
The classification process explains why this is so. The reason Class III devices require a more rigorous FDA process is that Class III is for riskier devices. And the reason Class II devices get less scrutiny is that they are safer devices. The end result – “reasonable assurance of safety and effectiveness” – is the same. In fact, some figures (see, the article) show proportionally fewer recalls for Class II 510(k) devices than for Class III premarket approved devices.
A comparison to hospital triage illustrates the mistake of linking safety to the degree of regulation. In an emergency room, some complaints merit no treatment, some require medicine, and others must have surgery. In each case, the goal is to return the patient to health. No one would say that the patient sent to surgery was healthier than the patient sent home with medicine, but that is precisely the kind of reasoning that is used when it is argued that more heavily regulated devices must somehow be safer than those that have been chosen, because of their lower risk, for less regulation.
Sometimes plaintiffs claim that the clearance of a Class II device using 510(k) is inferior to approval of a riskier Class III device because a regulation precludes describing 510(k) clearance as “approval.” 21 C.F.R. §807.97. The FDA adopted that regulation in 1976, when all clearance was transitional, and required only the filing of the previously described 510(k) notice. In 1976, the FDA did not in any sense “approve” such devices, since it did not have to grant permission to market. But since 1990, FDA permission is required, and the grant of permission to do something falls within the ordinary dictionary meaning of “approval.”
As a result, the only way to make sense of the regulation today is to say that it uses the word “approval” in a technical sense, i.e., the name the statute gives to the process used to gain permission to market a Class III device. The process used to gain permission, however, has no bearing on the “safety and effectiveness” of the device because the same standard, that of “reasonable assurance,” applies across the board.
It appears, however, that it is going to take a judge willing to declare that the Emperor has no clothes to correct these errors. A district court willing to take an independent look at the statutes and regulatory history of a particular device could do that. It might be encouraged to do so if it sought the opinion of the FDA on the issue. The FDA cannot and will not deny that “reasonable assurance of safety and effectiveness” is the standard that it requires any device to meet before that device can be marketed.
Reliance on Lohr to judge the FDA clearance process today is like relying on a home video of a child dancing with Barney the Dinosaur to judge the same person as an adult on Dancing With the Stars. The name of the contestant may be the same, but the video will say nothing of value when it comes to judging the adult’s ballroom dancing skills. Section 510(k) may still be a relevant citation, but Lohr says nothing about the methods the FDA uses today to determine the safety and effectiveness of devices, which involve far more than just a citation to 510(k).