We recently came across the law review article, E. Lindenfeld, “Clear Evidence Clarified,” 75 Food & Drug L.J. 346 (2020). Since it cited and critiqued a number of our blogposts, we thought it was appropriate to reply.

Our initial impression is that the Lindenfeld article is comparatively reasonable – that is, compared to some prior P-side analysis of Merck Sharp & Dohme Corp. v. Albrecht, 139 S. Ct. 1668 (U.S. 2019), we’ve seen. We do note a significant omission, however. The phrase “newly acquired information” appears only once in the article, 75 Food & Drug L.J., at 348, and only in passing. Thus, the article tacitly buys into the other side’s largely rejected position that there is no separate basis for preemption based on this prong of the FDA’s changes being effected (“CBE”) regulation. However, since the article does not analyze this issue, we won’t either, except to say that courts have just about universally treated lack of “newly acquired information” as an independent basis for preemption, also decided as an issue of law under Albrecht, with the plaintiff having the initial burden of identifying the “information” that allegedly satisfied this requirement. E.g., Knight v. Boehringer Ingelheim Pharmaceuticals, Inc., ___ F.3d ___, 2021 WL 41897, at *5-6 (4th Cir. Jan. 6, 2021); In re Avandia Marketing., Sales & Products Liability Litigation, 945 F.3d 749, 760 (3d Cir. 2019); Boone v. Boehringer Ingelheim Pharmaceuticals, Inc., 239 A.3d 1175, 1194 n.33 (Conn. 2020); see generally our post here.

Turning to what the Lindenfeld article does discuss, it focuses (p. 355) on the Supreme Court’s explication of the “clear evidence” requirement in Albrecht: that in cases “like” Wyeth v. Levine, 555 U.S. 555 (2009), a defendant must have evidence “show[ing] that it fully informed the FDA of the justifications for the warning required by state law and that the FDA, in turn, informed the drug manufacturer that the FDA would not approve changing the drug’s label to include that warning.” 139 S. Ct. at 1678. Except for one thing. . . .

That thing is the Lindenfeld article’s complete omission of Albrecht’s limitation of this reasoning to cases “like” Levine. Thus, the article conclusively assumes that all “clear evidence” cases are “like” Levine – which would be OK for a P-side piece, except we object to this being done sub silentio. Maybe that approach will prevail, but right now, the only on-point precedent is to the contrary. Cerveny v. Aventis, Inc., 783 Fed. Appx. 804, 808 n.9 (10th Cir. Aug. 9, 2019) (case was not “like” Levine because of third-party regulatory activity that independently caused the FDA to reject the same warning the plaintiff had proposed. Admittedly, that’s a only footnote in a non-precedential decision, but it’s the only thing so far on this particular point.

The Lindenfeld article parses Albrecht’s language into five parts: “[(1)] the drug manufacturer [(2)] fully informed the FDA [(3)] of the justifications for the warning required by state law [(4)] and that the FDA, in turn, informed the drug manufacturer [(5)] that the FDA would not approve changing the drug’s label to include that warning,” which it characterizes as “extremely positive” from the plaintiff’s point of view. 75 Food & Drug L.J., at 356. Since the Blog designated this part of Albrecht our sixth worst case for 2019, we don’t entirely disagree.

The Lindenfeld article’s first proposition is that Albrecht effectively threw out Levine’s requirement of “clear evidence” that the FDA “would have rejected” the label change that any particular plaintiff is advocating – calling “hypothetical preemption” “nearly dead,” and arguing instead that the defendant must “actually” have proposed, and had rejected, the same label change that the plaintiff wanted. 75 Food & Drug L.J. at 356. That, of course, is akin to saying no “clear evidence” preemption ever – because what plaintiff’s counsel will be so inept not to propose something different than what the FDA is already on record as having nixed? If the Court in Albrecht had intended there to be no preemption, then it would have said so, and the opinion would have been a lot shorter. This overstatement ties in, we think, to the article’s prior 86ing of Albrecht’s prefatory “like Levine” limiting language. A lot of cases aren’t “like” Levine and thus can involve the FDA “informing” manufacturers (including defendants) of rejected warning language even in the absence of a particular entity’s submission. Two such examples are rejections of citizen’s petitions (what happened in the Cerveny case cited above) and FDA implementation of classwide labeling covering multiple similar drugs in one fell swoop.

Another example is the FDA’s “own” inaction under 21 U.S.C. §355(o), which the Lindenfeld article mentioned earlier (pp. 349-50), but then does not discuss again until much later. If the FDA was sitting on “new safety information” that would have authorized it under §355(o) to “promptly notify” a defendant manufacturer of the need for a label change and did nothing, then that would seem like a good basis for concluding, in the absence of an actual submission, that the FDA “would have” rejected a label change under the original Levine version of the “clear evidence” test. There’s no law on this fact pattern yet, and it might not be forthcoming, because the cleanest route to preemption on this set of facts appears to be the “newly acquired information” prong of the CBE regulation, rather than making essentially the same argument based on §355(o).

In any event, the Lindenfeld article’s almost interment of the Levine version of “clear evidence” seems to us, at minimum, to be premature. “Albrecht is better understood as a clarification of the impossibility standard in [Levine] rather than as a repudiation of it.” Dolin v. GlaxoSmithKline LLC, 951 F.3d 882, 888 (7th Cir. 2020.

[I]t should not always be the case that simple inaction by the FDA in light of submitted data will always be “clear evidence” that the FDA would reject a particular warning. In this case, however, in light of the known issues and the ongoing give-and-take between [the defendant] and the FDA on these issues . . . the FDA’s continued inaction does represent clear evidence under these facts.

Ridings v. Maurice, 444 F. Supp.3d 973, 998 (W.D. Mo. 2020) (emphasis original). We think there’s life in this part of Levine yet.

The next proposition of the Lindenfeld article is that the preemptive “force” of an FDA rejection of a citizen petition “is uncertain.” 75 Food & Drug L.J. at 357. That’s true, and always has been. Some courts have allowed plaintiffs another bite at the apple, even after some plaintiff-side group tried and failed – under identical scientific standards − to get the FDA to make the same label change. But as the article conceded, id. at 358 & n.117, several appellate decisions (including the aforementioned Cerveny decision) have not. That post-Albrecht courts have largely “sided with defendants” is not surprising. The citizen petition process is a formal one, 21 C.F.R. §10.30, and thus provides the “official administrative record for an FDA decision” that Albrecht finds to be an appropriate basis for preemption. 139 S. Ct. at 1680.

The Lindenfeld article asserts that this view “runs against the plain language of Albrecht” that “it must be the manufacturer who proposes the change.” 75 Food & Drug L.J. at 358 (emphasis original). But that position ignores both Albrecht’s “plain language” that the Albrecht description of “clear evidence” is applicable only to cases “like” Levine and the facts of Albrecht itself – which involved class labeling. See In re Fosamax (Alendronate Sodium) Products Liability Litigation, 852 F.3d 268, 278 (3d Cir. 2017) (in 2010, “FDA announced that it would require all bisphosphonate manufacturers to add information regarding the risk of atypical femoral fractures”) (emphasis added), vacated, 139 S. Ct. 1668 (2019). That the Supreme Court in Albrecht remanded, rather than reversed, alone establishes that this assertion is incorrect. The manufacturer need not have itself proposed the label change for it to be preemptive.

The Lindenfeld article next addresses the language in Albrecht about defendants “fully infom[ing]” the FDA. 75 Food & Drug L.J. at 359-61. The article recognizes that “a number of courts have held that Buckman [v. Plaintiffs Legal Committee, 531 U.S. 341 (2001),] precludes, at the clear evidence stage, inquiries into sufficiency of a manufacturer’s submissions to FDA.” 75 Food & Drug L.J. at 360 (footnote omitted). This is, frankly, the hardest part of Albrecht to reconcile, since neither the majority nor the concurrence ever cite Buckman. The article takes the position that “it is abundantly clear that plaintiffs emerged from the high Court triumphant on this issue” and that “plaintiffs who believe that defendants were less than forthright with FDA” are free to ignore Buckman’s preemptive effect. 75 Food & Drug L.J. at 360-61.

We don’t think so, and so far, neither have the courts. Albrecht does not give plaintiffs free reign to discard prior FDA rejections of their positions. After Albrecht “the FDA, and only the FDA, can determine what information is ‘material’ to its own decision to approve or reject a labeling change.” Avandia, 945 F.3d at 759 (emphasis original). In our view that goes for both sides. We think that what this “fully informed” language means is much less than plaintiffs being “triumphant.” Rather, it means that, where the FDA has rejected a label, it is for the FDA to decide − perhaps by citizen petition, perhaps by other means – whether the information on which it acted was complete. We have already mentioned that happening in one case. By contrast, the Lindenfeld article has over 200 footnotes, but none supporting its plaintiffs were “triumphant” statement. While we believe that the Buckman/Albrecht interface is an open question, plaintiffs haven’t won anything yet.

The next section of the Lindenfeld article is not as controversial. It is that a defendant seeking preemption must have informed the FDA of the “justifications” for a label change strengthening a warning. To the extent that preemption doesn’t turn on routine submission of “adverse event reports,” 75 Food & Drug L.J. at 361, we’re not going to disagree. “[S]tacks” of unanalyzed documents, id.at 362, don’t cut it after Albrecht, if they ever did. But this aspect of Albrecht is not that important, because it overlaps substantially with the other CBE requirement that there be “newly acquired information” sufficient to support any supplemental label change. The question of how “fully informed” the FDA was is more properly – particularly given the Buckman issue – decided by answering the question of whether newly acquired information exists. See, e.g., Gayle v. Pfizer Inc., 452 F. Supp.3d 78, 88 (S.D.N.Y. 2020) (“adverse event reports, without any analysis indicating causality, cannot constitute ‘newly acquired information’”). The preliminary CBE requirement of newly acquired information is where we see issues of adequate factual basis for a label change being hammered out – not “clear evidence.”

Finally, Albrecht requires sufficient FDA “communicat[ion]” of any denial. Here is where we are constrained to agree with the Lindenfeld article. Loosy-goosy preemption arguments based on informal FDA statements, or telephone calls, aren’t likely to succeed after Albrecht. Not that this kind of argument had great success before. E.g., Reid v. Johnson & Johnson, 780 F.3d 952, 964 (9th Cir. 2015); Fellner v. Tri-Union Seafoods, L.L.C., 539 F.3d 237, 250 (3d Cir. 2008). But there is still a contradiction here that the Lindenfeld article has difficulty dealing with. In listing the kinds of FDA activity that were preemptive, Albrecht, 139 S. Ct. at 1679, included a citation to 21 U.S.C. §355(o)(4)(A), which as mentioned above, includes the FDA’s power to require label changes on its own initiative. We’re not sure what this means either, so we agree with the article that there is “ambiguity” here that potentially supports preemption. See Ridings, 444 F. Supp.3d at 998-99 & n.24 (“FDA’s continued inaction does represent clear evidence under these facts”).

In discussing Albrecht’s “policy implications,” the Lindenfeld article repeats its over-enthusiastic conclusion that “hypothethcal preemption” is “no longer viable.” 75 Food & Drug L.J. at 363. Again, that’s a function of ignoring the Court’s “like” Levine qualifier. Where the sole regulatory actor was the defendant, and the sole activity was whether to change the label – the facts of Levine – “would have rejected” arguments may well be precluded by Albrecht. Extending such a preclusion to instances of a much fuller regulatory record, however, is an aspirational, not actual, reading of Albecht. The article goes on to state, however:

Albrecht does leave many important questions unanswered. As explained, the preemptive force of citizen petitions remains unclear. And while the Court did state defendants must have provided FDA with “full information,” it is unclear exactly what that necessarily entails. It is also unclear if courts will consider FDA inaction as sufficient under the Albrecht test.

75 Food & Drug L.J. at 363-64 (footnotes omitted).

Agreed. But does Albrecht take a “bright line” id. at 365-66, approach to preemption? Here we part ways. Frankly, we wish it had. But for the reasons discussed above it doesn’t. Even at its most definite, when discussing FDA actions with “force of law,” 139 S. Ct. at 1679, there is still the citation to §355(o). Where that regulation is in play, is it still a case “like” Levine? We can’t say for sure, and we can think of good arguments why it’s not, and shouldn’t be.

Will the Lindenfeld article’s position on Albrecht, if accepted, “overwhelm” the FDA? The article, of course, says no. Food & Drug L.J. at 367-68. But the FDA has throughout its history operated informally, on the basis of administrative back and forth. Albrecht’s emphasis on “formal” rejection of label changes as a barrier to preemption certainly could encourage liability-adverse manufacturers to push for final FDA action, which would put more strain on the FDA’s limited resources. Cf. 75 Food & Drug L.J. at 364 (FDA has “only 100 professional employees to monitor over 11,000 drugs) (citing 12-year-old law review article). It certainly won’t help for the Agency to be pushed to make more “formal” administrative dispositions than it has had to in the past.

Finally, the Lindenfeld article observes that defendants retain other “avenues” of preemption. 75 Food & Drug L.J. at 368-69. We agree. However, for the reasons already discussed in this post, we think those avenues are more substantial than major/moderate changes (which include all design changes), highlights, black box warnings – all of which the article acknowledges are “virtually always” or “automatically” preempted. Id. at 368. We think that newly acquired information, citizen petitions, class labeling, and other matters discussed herein also provide fertile preemption grounds.

Nor can we agree that, because “virtually across-the-board” preemption applies to generic drugs, id. at 369, that it “it makes sense to impose a difficult preemption regime on manufacturers of brand name drugs.” Id. That argument stands FDA regulation on its head. Branded drugs have to go through extensive clinical trials, and all the other steps required to obtain a full New Drug Applications. Generic drugs have none of that, and simply piggy-back on the efforts of branded manufacturers affirmatively to prove their products’ safety and effectiveness. The NDA isn’t an exception to the FDA’s safety requirements. Rather, “it is federal safety review.” Riegel v. Medtronic, Inc., 552 U.S. 312, 323 (2008) (emphasis original). Indeed, this closing contention sounds more like an argument for innovator liability – which we detest – than one against preemption.