Photo of Michelle Yeary

June 25, 2026 was a very good day for preemption. While one hand of the judiciary was busy with the Supreme Court handing down its 7-2 ruling in Monsanto Co. v. Durnell — establishing that FIFRA expressly preempts Roundup state-law failure-to-warn claims, the other hand was quietly doing the same thing with toothpaste. Specifically, the Southern District of New York issued Brower v. Colgate-Palmolive Co., 2026 WL 1831253 (S.D.N.Y. June 25, 2026), a partial dismissal of an OTC economic loss class action that drew a clean line between preempted omission claims and surviving affirmative misrepresentation claims under the FDCA.

Plaintiffs — a geographically scattered group of consumers from New York, Minnesota, Illinois, New Jersey, Arizona, and California — sued Colgate-Palmolive on behalf of a putative nationwide class, alleging that Colgate’s toothpastes contained heavy metals, specifically lead and mercury, that Colgate knew about but chose not to disclose to consumers. Id. at *1. The complaint ran to thirteen claims: eleven state consumer protection counts, a common law fraud by omission claim, and unjust enrichment. Defendant moved to dismiss on four grounds: (1) express preemption under the FDCA; (2) lack of standing; (3) failure to state a claim; and (4) statute of limitations. The court granted the motion in part, and the preemption analysis is where the real action is.

Before diving in, we have to acknowledge the timing — and its significance. As we discussed in our post on Durnell, the Supreme Court’s ruling established that FIFRA’s express preemption clause — which the Court described as “materially identical” to the FDCA’s device preemption provision at 21 U.S.C. § 360k(a) — bars state-law claims that would impose requirements “in addition to or different from” what federal law demands. The Brower court wasted no time citing Durnell in its preemption analysis of the FDCA’s OTC drug preemption clause, 21 U.S.C. § 379r(a). That is no coincidence. As we noted in our Durnell post, § 379r(a), appears on the Supreme Court’s own list of FDCA provisions with preemption language “similar or identical” to FIFRA’s. See Durnell, 2026 WL 1825691, at *8.

The upshot: the Supreme Court’s interpretation of what “in addition to or different from” means in the FIFRA context — that it means exactly what it says — applies with equal force to the FDCA’s express preemption clauses. Brower v. Colgate is, in part, a same-day proof of concept.

Plaintiffs alleged that defendant’s voluntary safety-related statements — including that its products promote “Whole Mouth Health,” “Protect teeth, tongue, cheeks [and] gums,” and are “Responsibly Made” — were false and misleading because the products actually contained lead and other heavy metals. Id. at *3. The court held these claims were not preempted. The FDCA itself prohibits false or misleading labeling under 21 U.S.C. § 352(a)(1), and plaintiffs’ state-law claims ran parallel to that federal prohibition rather than imposing requirements different from or in addition to it. Because the challenged statements were voluntary marketing claims not addressed by the FDA’s monograph, they were fair game for challenge under state consumer protection law. Id. at *4.

The key piece of the court’s analysis is that the existence of the monograph does not give manufacturers blanket immunity over every statement that appears on their label. Here, the monograph authorizes specific decay-prevention claims and prescribes labeling requirements for the anti-cavity drug component. It does not, as the court put it, “broadly immunize all advertising or labeling statements made by manufacturers.” Id.

The story is different for omission claims. Plaintiffs also argued that defendants should have disclosed the presence of heavy metals on its labeling. The court held those claims expressly preempted under § 379r(a). Id. at *5. The reasoning is straightforward. Because the FDCA imposes no obligation to disclose the presence of heavy metal contaminants anywhere on a toothpaste label, any state-law claim premised on that omission creates requirements “in addition to” what federal law demands. And that is precisely what § 379r(a) forbids.

Central to this conclusion was the court’s characterization of the FDA monograph as “comprehensive and granular.” The monograph prescribes, with considerable specificity, the required active ingredient disclosures, warning language, directions for use, indications, and format requirements for anti-cavity drug products. This level of regulatory detail, the court found, reflects a considered exercise of the FDA’s comprehensive authority over OTC drug labeling. When the FDA has constructed an elaborate labeling framework and heavy metal disclosures appear nowhere in that framework, the inference that the agency made a deliberate judgment about label content is “substantially stronger” than in cases where the regulatory scheme only lightly touches the relevant product category. Id. As the court memorably put it:

To allow state omission claims to supplement this comprehensive scheme would be to substitute judicial judgment for agency judgment in precisely the domain Congress assigned to the FDA.

Id. In other words, when the FDA has spoken comprehensively about what belongs on a label, a state court cannot supplement that judgment by compelling additional disclosures the agency never required.

The court also took care to distinguish (and politely part ways with) the District of Connecticut’s analysis in Clinger v. Edgewell Personal Care Brands, 2023 WL 2477499 (D. Conn. Mar. 13, 2023), which had addressed omission claims in a similar OTC drug context. The Clinger court acknowledged the full text of § 379r(a) — with its “different from” and “in addition to” prongs — but then analyzed only the former in applying the standard. By asking only whether federal law prohibited the omitted disclosure (rather than whether it required it), Clinger effectively collapsed the two prongs into one, treating the absence of a federal prohibition as sufficient to defeat express preemption. Brower, 2026 WL 1831253, at *5. That, the court explained, drains the “in addition to” language of its independent force. The two prongs are distinct, and the Brower court rightly gave both their due.

This is good ammunition for future preemption arguments. The “in addition to” prong — which asks whether a state requirement exceeds the federal floor rather than whether it conflicts with federal law — is an independent basis for preemption that plaintiffs’ counsel has too often been allowed to paper over.

Having preempted the omission claims, the court turned to standing and trimmed the plaintiff roster further. The problems were predictable, but instructive for those fighting contamination class actions. Several plaintiffs were dismissed for failure to meaningfully link their testing allegations to the products they personally purchased. Plaintiffs relied on two sources of testing data. Both were found wanting for deficiencies like failing to state the number of samples tested, the geographic proximity of testing to plaintiffs’ purchases, the date of testing, or who performed it. Id. Without those details, the court could not infer that heavy metals were systematically present in the products each plaintiff actually bought. “Some study found lead in something” does not establish that your tube of toothpaste had lead in it.

One additional standing point that deserves emphasis: the court rejected plaintiffs’ argument that one named plaintiff’s standing confers standing on other named plaintiffs. Id. Each plaintiff in a class action must establish their own Article III standing. That principle is sometimes honored more in the breach, and it was good to see it applied here with rigor.

So as to provide a complete picture of the decision, here are also the key points on the other issues addressed by the court:

  • Failure to state a claim. On the surviving affirmative misrepresentation claims, plaintiffs sufficiently alleged that the challenged statements were false or misleading and that there was a material safety risk from the trace amounts of heavy metals at issue. Whether statements like “Whole Mouth Health” and “Responsibly Made” are misleading to a reasonable consumer is a question of fact for the jury, and the court declined to resolve it on a 12(b)(6) motion. Id. at *9. The court similarly refused to look behind plaintiffs’ “any exposure to heavy metals is harmful” theory of injury; whether those trace amounts actually pose a health risk is reserved for summary judgment.
  • Consumer fraud elements by state. The New York GBL §§ 349 and 350 claims survived, as did the California CLRA and UCL claims, for the same reasons. Id. at *9-10. On the Minnesota claims: the MFSAA and MCFA survived (the public benefit requirement was satisfied given the breadth of the alleged marketplace deception), but the MDTPA claim was dismissed because plaintiff Brower — the only Minnesota named plaintiff — failed to adequately allege irreparable injury or a genuine threat of future harm necessary for injunctive relief under that statute. Id. at *11.
  • Unjust enrichment. Dismissed as duplicative. The unjust enrichment claim rested on the exact same facts and sought the exact same relief as the remaining statutory claims, without identifying any independent quasi-contractual basis for recovery. Id. at *12.
  • Statute of limitations. Defendant argued that New York’s three-year statute of limitations for consumer protection claims bars any recovery for purchases predating 2022. The court declined to dismiss on that basis because the complaint did not plead facts sufficient to determine when particular purchases were made. The issue is preserved, so it’s a fight for another day. Id. at *13.

Two decisions on the same day (one by the Supreme Court) reinforcing the same concept: “in addition to” and “different from” mean exactly what they say, and courts should not let plaintiffs collapse those two prongs into one. And, while this is not a complete preemption win, for defendants it ultimately leaves a manageable battlefield — the surviving affirmative misrepresentation claims are limited to specific voluntary marketing puffery, not the full universe of possible labeling challenges. A good day indeed.

Photo of Steven Boranian

New Jersey law requires that trial courts conduct a “rigorous gatekeeping” analysis when assessing the admissibility of expert opinions.  That’s the Accutane standard, named for the New Jersey Supreme Court’s opinion in In re Accutane, 234 N.J. 340 (2018).  To the extent there was any ambiguity on whether Accutane applies to all civil cases, the New Jersey Supreme Court has now revolved it  Accutane applies in every civil case.  Moreover, the “rigorous analysis” requirement applies to all purported methodologies, including differential diagnoses.

The case is Beavan v. Allergan U.S.A., Inc., 355 A.3d 1201 (N.J. 2026), which involved an FDA-approved steroid pellet used to treat certain eye conditions. In 2018, the manufacturer advised the FDA that during assembly a tiny silicone particulate could detach from the needle’s sleeve and be ejected with the pellet.  This issue ultimately led to a recall covering twenty-two lots in which sample testing showed that two of ninety units—2.2 percent—contained a silicone particulate.

The plaintiff was treated with an injection from a recalled lot and alleged that she lost vision in one eye as a result.  To prove this, she relied on two experts, whose opinions were based on . . . well, not much.  Both reasoned that because the plaintiff had received prior injections without complications, the recalled unit must have been defective and most likely caused the injury.  But notably, neither could cite evidence that a silicone particulate was actually generated or injected during the plaintiff’s procedure.  As one expert (the treating physician) acknowledged, “the only thing we were going by” was that “it was a recalled lot.”  Beavan, 355 A.3d at 1210. 

The trial court found these opinions sufficiently reliable to raise triable issues of fact, but did so without conducting the Accutane gatekeeping inquiry.  The Appellate Division reversed and ordered entry of summary judgment for the defendant, in an opinion that we reported on here. 

The New Jersey Supreme Court took a different tack and remanded for the trial court to conduct the proper “rigorous gatekeeping” under Accutane.  According to the Supreme Court any dispute about the reliability of expert testimony in a civil case must be resolved by the trial court acting as gatekeeper, applying the Accutane framework.  Here is the quote you will want to use:

Our view of proper gatekeeping in a methodology-based approach to reliability for expert scientific testimony requires the proponent to demonstrate that the expert applies . . . scientifically recognized methodology in the way that others in the field practice the methodology. When a proponent does not demonstrate the soundness of a methodology, both in terms of its approach to reasoning and to its use of data, from the perspective of others within the relevant scientific community, the gatekeeper should exclude the proposed expert testimony on the basis that it is unreliable.

Beavan, 355 A.3d at 1216 (quoting Accutane).  There are no exceptions.  Nor is Accutane’s mandate confined to the mass tort context, which the Supreme Court made clear:

The procedural framework adopted in Accutane is not limited to the specific setting in which that case arose; it applies to all civil cases in which the parties dispute the reliability of expert testimony under N.J.R.E. 702 and 703.

Id. (emphasis added). 

Accutane likewise carved out no exception for differential diagnosis. To the contrary, when an expert uses a differential diagnosis to opine on the cause of an injury, the gatekeeping inquiry must be applied to both steps of the analysis: the “rule in” step, in which the expert compiles plausible causes, and the “rule out” step, where the expert eliminates factors to reach a conclusion as the most likely cause. The trial court must scrutinize each component of the opinion to confirm it rests on a methodology satisfying Accutane.

Applying these principles, the Supreme Court held that the trial court erred by admitting the testimony without “rigorous gatekeeping,” and the Appellate Division got it wrong too by directing entry of summary judgment.  The proper outcome, in the Supreme Court’s view, was to remand for the trial court to properly assess the admissibility of plaintiffs’ experts.

So the plaintiff gets a do over.  The Supreme Court encouraged the trial court to hold an evidentiary hearing, and it also held that the plaintiff’s retained expert’s opinion was not a “net opinion,” which is a New Jersey thing.  A “net opinion” is one that is unsupported by facts or data, but the retained expert here adequately explained the “why and wherefore” of his opinions in his report.  That does not, however, mean that the retained expert’s opinions will be admissible after scrutiny under Accutane, which is a separate inquiry. 

Finally, the Court also flagged that a treating physician designated to offer expert causation testimony must, like a retained expert, serve a written report when an adversary requests one through interrogatories. The treating physician here did not prepare a report.  On remand, the trial court will determine whether the plaintiff will be allowed to serve one. 

The New Jersey Supreme Court vacated a good outcome for the defendant in the Appellate Division, but the good news is that it solidified the trial court’s gatekeeping role and closed off any uncertainty on whether and how broadly Accutane applies.  Rigorous gatekeeping is required, in all civil cases. 

Photo of Bexis

Ever since we reviewed the Environmental Protection Administration’s (“EPA”) 2019 decision reaffirming its prior conclusion that glyphosate does not cause cancer and, therefore, the “Roundup” herbicide did not warrant any cancer warnings, we have been convinced that state-law claims demanding such warnings were (or should be) preempted – regardless of how pro-plaintiff decisions might seek to muddy the waters.

It took a while, but yesterday the United States Supreme Court, in a strong 7-2 decision held precisely that:

[Plaintiff’s] state tort claim would require [defendant] to add a cancer warning to Roundup’s label even though federal law requires [defendant] to use the EPA-approved label without a cancer warning.  Because [plaintiff]’s state tort claim would impose a pesticide labeling requirement “in addition to or different from” the label required by EPA, FIFRA expressly preempts [plaintiff]’s claim.

Monsanto Co. v. Durnell, 609 U.S. ___, 2026 WL 1825691, at *3 (slip op.) (U.S. June 25, 2026) (emphasis added) (hereafter “Durnell”).

We add “emphasis” to the statutory language because, as the Supreme Court observed, the express preemption clause that Congress enacted in FIFRA, 7 U. S. C. §136v(b), is “materially identical” to the FDCA’s preemption clause for medical devices, 21 U.S.C. §360k(a).  Durnell, 2026 WL 1825691, at *10.  See Id. at *8 (“nearly identical”); cf. id. at *8 (listing 21 U. S. C. §§678 (meat); 1052(b) (eggs); 467e (poultry); 379s(a) (cosmetics), 379r(a)(2) (OTC drugs); and 343-1(a)(2)-(4) (food)) as having “similar or identical labeling preemption provisions”).  The only difference from §360k(a) is inversion of the order of the phrases “different from” and “in addition to.”  The ordering of this identical language meant nothing in Durnell.  Thus, the numerous pro-preemption holdings in Durnell should apply with equal force to medical devices and other federally regulated consumables.  It becomes our job as defense counsel to ensure that this happens.

Turning to those pro-preemption holdings, we first address something that’s not found in the majority opinion at all – that is any reference to a “presumption,” or even an “assumption” against preemption.  That’s important because a 20+-year-old FIFRA opinion, Bates v. Dow Agrosciences LLC, 544 U. S. 431 (2005), had relied on such a presumption, despite the statute’s express preemption clause.  Id. at 449-50.  In the interim (as the Blog has repeatedly discussed), the Court abolished any such presumption in express preemption cases.  Puerto Rico v. Franklin-California Tax-Free Trust, 579 U.S. 115, 125 (2016).  But in Durnell, despite what Bates had held about preemption presumptions, the majority opinion said not a word about it, even while discussing Bates itself.  Instead, the “textual conclusion” in Durnell gave force to the literal meaning of §136v(b)’s language:

In sum, federal law requires [defendant] to sell Roundup with the label that EPA approved at the initial registration and that EPA has subsequently re-approved on multiple occasions − that is, the label without a cancer warning.  [Plaintiff]’s state tort claim, by contrast, would require [defendant] to add a cancer warning to its labels.  That [state]-law requirement is “in addition to” and “different from” [defendant]’s federal-law labeling obligations.

2026 WL 1825691, at *8.  Hence the title of this post, that “‘different from’ and ’ in addition to’ mean what they say.”  Particularly in the recalcitrant Third Circuit, Durnell is further proof that, yes, the Supreme Court’s abolition of any presumption against express preemption applies to product liability litigation.

As for Bates, Durnell distinguished its adverse aspects as involving a claim rarely seen in product liability – challenging the product’s “efficacy” rather than its safety.  Id.  Far more pertinent, the Court held, was its later preemption decision in Riegel v. Medtronic, Inc., 552 U. S. 312 (2008) (see, e.g., our discussions of Riegel here, here, and here).  Riegel, as all regular Blog readers well know, is the basis for broad preemption of tort claims against FDA pre-market approved (“PMA”) medical devices.  Seven justices in Durnell held that “Riegel further confirms that [plaintiff]’s failure-to-warn claim is expressly preempted,” 2026 WL 1825691, at *8, because §360k(a)’s nearly identical language likewise preempted claims challenging the FDA’s “review [of] the device’s label and [its] determin[ation] that the label is neither false nor misleading.”  Id.  The Durnell court viewed an FDA medical device PMA as equivalent (“just as”) to the EPA’s approval (called “registration”) of regulated chemicals.  Id.  Thus,

Riegel is dispositive here.  If FDA’s premarket approval of medical devices preempted additional state-law requirements, so too must EPA’s registration of pesticides and approval of pesticide labels.

Id. at *9.  Both statutes “reflect Congress’s judgment that the ability to sell a product throughout the country with a single label can be important to maintaining an efficient nationwide market.”  Id.  That’s an important concept for us to remember – uniformity in labeling is an economic virtue that Congress has chosen to protect through preemption.

The Court’s view of Riegel as “dispositive” also has implications for non-warning claims in Roundup litigation.  We expect the other side to disagree, but they are whistling past the preemption graveyard.  Why?  Riegel.  Based on the same “different from”/”in addition to” preemption language, Riegel held that numerous non-warning claims were also preempted by the FDA’s device approval.  552 U.S. at 320-21 (preemption extends to “strict liability; breach of implied warranty; . . . negligence in the design, testing, inspection, distribution, labeling, marketing, and sale of the [product; and] . . . negligent manufacturing”).  To the extent that Roundup plaintiffs try to fall back on non-warning-related claims, the “dispositive” Riegel decision should bar those claims as well.

As to Riegel, we also were also heartened by Durnell’s footnote 9, which reiterated something that the lower courts have been forgetting in medical device litigation – Riegel did not create an exception for so-called “parallel claims”:

Riegel did not consider state-law claims that were parallel to an applicable federal requirement, apart from the FDA’s premarket approval process. . . .  It is true that Riegel did not assess whether the state-law claims at issue were “different from, or in addition to” the relevant federal-law requirements.

Id. at *12 n.9 (emphasis added).  Too many courts seem to view Riegel as affirmatively authorizing “parallel” claims, when in fact the Court did no such thing, since it never considered that issue at all.  Durnell did, and the existence of purportedly “parallel” claims did not defeat preemption, because “EPA’s analogous registration and label approval determinations” were nevertheless “‘requirements’ under FIFRA.”  Id.  We should be reminding courts, first, that Riegel did not endorse the waived concept of parallel claims, and second, that such claims are still “requirements” that can be “additional to” governmentally imposed requirements.

The good news for defendants does not stop with the Court’s applying Riegel and finding preemption.  Durnell went on to demolish a number of anti-preemption arguments that plaintiffs in prescription medical product liability litigation have also frequently raised.

Number 1 – That tort law and FDA regulation are compatible because both promote safety.  Not good enough.  “[T]hat argument operates at far too high a level of generality and disregards the central and comprehensive role that [the agency] performs in making labeling determinations under [the statute’s] registration provisions.”  Durnell, 2026 WL 1825691, at *9.

Number 2 – Reliance on general “misbranding” or “adulteration” provisions to attack product-specific labeling.  The general does not control the specific.  “Reliance on [the statute’s] general standard for misbranding rather than the specific requirements imposed under federal law would nullify [the] preemption clause and the uniformity that Congress sought for safety warnings on [product] labels.”  Id.

[Plaintiff’s] failure-to-warn claim does not fault [defendant] for using a label different from the [agency]-approved labeling.  [Plaintiff] instead faults [defendant] for not using a label different from the [agency]-approved label.  But [the statutory] preemption clause expressly preempts any state tort claim that would require a [regulated] manufacturer to use a label “in addition to” or “different from” federal requirements imposed under [the statute], which, as explained above, include the [agency]-approved label.

Durnell, 2026 WL 1825691, at *10.  This holding in Durnell should be particularly useful in OTC drug litigation, where some courts have erroneously credited such arguments.

Number 3 – Product-specific requirements not having preemptive force.  That argument was worse than wrong; it was “implausible.”  Id.  “[B]y its text, [the statute] affords preemptive force to federal requirements imposed ‘under’ [it], not merely those imposed ‘by’ the actual statute.”  Id.  The same is true of §360k(a), thus this argument also “contravenes Riegel.”  Id. at *10.  The statute also authorizes the agency to “prescribe regulations” and “make registration and labeling decisions.”  Id. at *9.  “There is no good argument for treating FDA’s premarket approval as ‘requirements’ ‘under’ the Medical Device Amendments, but not treating EPA’s regulations and registration determinations as requirements under FIFRA’s materially identical preemption clause.”  Id. *10.

Number 4 – Loper Bright.  Given the statute’s express delegation of power to prescribe regulations and to “determine” labeling compliance, Loper Bright Enterprises v. Raimondo, 603 U. S. 369 (2024), is completely inapplicable to preemption.

[The statute] empowers [the agency] to “prescribe regulations. . . .”  And [it] expressly directs [the agency] to register [products] and determine that the [their] labeling complies with [the statute’s] many specific requirements.  During that extensive registration process, [the agency] critically evaluates the [product] label to ensure that the label contains all warnings necessary to protect human health.  After [the agency] makes a determination about the appropriate warnings for a [product’s] label, a manufacturer is legally required to use that label unless and until [the agency] subsequently approves or requires a new label.

Durnell, 2026 WL 1825691, at *10 (citations and quotation marks omitted).  Change “register” to “approve” and, again, you have the FDCA (for both drugs and devices).  As we’ve argued before, Loper Bright only applies to agency statutory interpretation (like that wrongly credited in Medtronic, Inc. v. Lohr, 518 U.S. 470 (1996)), and not to delegated agency power to regulate particular products. The Supreme Court so held in Durnell.

Number 5 – Hindsight-based “unilateral” label change arguments.  That didn’t cut it either.  It doesn’t appear that the EPA has any equivalent to the FDA’s “changes being effected” regulations, so this aspect of Durnell does not map exactly to the FDCA.  But Durnell reiterated that federal law preempts any state-law demand for a label change when a manufacturer has no ability to change it unilaterally:

[I]f the manufacturer unilaterally changed the label, as [plaintiff] says [defendant] should have done, the manufacturer would be flouting [agency] regulations and exposing itself to potentially severe federal penalties.  So [plaintiff’s] argument also triggers potential retroactivity and estoppel questions.  The law is not ordinarily read to retroactively penalize persons for doing what the Government had required them to do.

Id. at *11 (citations omitted).  More broadly, Durnell rejected arguments that “would negate [the] express preemption clause, expose manufacturers to potentially massive tort liability for doing what [the agency] required them to do, and eviscerate the ‘uniformity of [agency] labeling determinations.”  Id. (citation omitted).

Number 6 – Failure to report new information.  That the government “was statutorily authorized to withdraw premarket approval for a medical device based on newly reported data or existing information” did not prevent preemption.  Id. at *12.  That power was solely vested in the government, not private plaintiffs.  That the government might determine that an “approved” label “had in essence become misbranded over time due to new evidence does not deprive [existing] registration decisions of their preemptive force.”  Id.  “[E]xtensive processes” for gathering and reporting of new risk information did not reduce the scope of preemption.

[Plaintiff’s] policy concern about regulatory lag is amply addressed by the extensive processes that [the statute and its] implementing regulations have established to respond to new or evolving safety information.  For example, manufacturers must apprise [the agency] of new information ‘regarding unreasonable adverse effects’ of their [products].  That obligation is backed by civil and criminal penalties.

Durnell, 2026 WL 1825691, at *12.  Conspicuously absent from the Court’s discussion of the “many ways of ensuring a [product’s] continued compliance” is any allowance of private litigation concerning purported failures to report.  Id.  Instead, “if third parties (like [plaintiff]) want to bring new information to [agency’s] attention or if they believe that [the agency] has failed to consider relevant information, those third parties are free to petition [the agency].”  Id. at *13.  The same is true of the FDA.  Thus:

[The agency] possesses a variety of tools to learn of and address new safety information.  And as a matter of law, state tort law may not impose labeling requirements “in addition to” or “different from” federal requirements imposed under [the statute].

Id.

That’s a lot of valuable new preemption ammunition that defendants in prescription medical product liability litigation can now rely upon.

Obviously, Durnell is a huge win for the defendants in the Roundup litigation.  But almost as obviously, it is also a big win for our medical device, OTC drug, and other clients who manufacture and market products that Congress intended to be regulated by the FDA, not by the states, and that are accordingly protected by the various preemption clauses in different parts of the FDCA.

Photo of Eric Alexander

Like many, many people around the world, we have spent a good chunk of the last two weeks following the World Cup.  Not just the U.S. team or even the clips of first-time visitors to the U.S. reveling in its regional cuisine and cultural quirks.  We have found ourselves watching intently as teams from two countries about which we otherwise have no strong feelings clash.  Naturally, the disinterested viewer takes on at least a mild rooting interest for one of the teams, usually the underdog.  In part because the field has expanded to 48 teams/countries, there are some real underdogs in the field.  There also have been some very surprising results that might make you decide that you are going to support one of those underdog teams for the rest of its run in the tournament.  (We eschew the international soccer announcing/reporting convention that a team takes on a plural verb form.)  In other words, you might choose to hop on the bandwagon of, for example, tournament “debutantes” Cape Verde or Curacao.  These island nations each tied a soccer power from a nation with more than 90 times as large a population.  So, feel free to hop on one of these bandwagons without hurting yourself.  However, please do not try to convert your new interest into the implausible claim that you were “always a fan” of the soccer team of a country you may have had trouble finding on a map.  If you want to claim longstanding fandom or even a pre-tournament prediction for a longshot’s success, then you better bring receipts.

We offer this public service message in part because today’s decision from Leon v. Zydus Pharms. Inc., No. 22-cv-7636 (JGK), 2026 WL 1758825 (S.D.N.Y. June 17, 2026), concerns implausible product identification claims that ended a product liability suit after it lingered around much longer than it should have.  First, the facts without commentary.  The plaintiff took some version of the prescription antihypertensive carvedilol for his hypertension for roughly two years before he had a stroke.  He may or may not have skipped his once daily dose of the medication the morning of his stroke.  Id. at *1.  Around the same time, the defendant learned that one lot of its carvedilol tablets somehow had a small percentage of tablets of the antidepressant paroxetine, which it also manufactured, included in what was released for distribution, so it recalled that lot.  Id.  Plaintiff, or his lawyers, connected the stroke to the recalled lot and sued the manufacturer.  After almost four years, including fact and expert discovery, the defendant moved for and won summary judgment because plaintiff never had any evidence that he unwittingly took paroxetine because of the issues with the recalled lot.  Id. at *3.  To the contrary, defendant offered uncontradicted evidence that the recalled lot was not distributed to where plaintiff lived and there were 299 non-recalled lots distributed.  Id.  So, defendant won, as it should have.  We will get back to the Leon court’s analysis, but first some commentary on the facts, particularly facts that are missing from the decision and we assume were brought out over the course of a case that should not have survived motions to dismiss.  (We found this basic information in minutes, so it surely was part of the record.)

Carvedilol has been around for a long time, so there are branded and generic versions.  The defendant is one of ten companies with an approved ANDA to sell generic carvedilol tablets of different dosages; it got its ANDA approved in 2007.  Other companies had ANDAs for generic tablets that were discontinued.  There is also an extended release carvedilol capsule in branded and generic forms, but the defendant never got an ANDA to sell those.  All of these are prescription medications in the U.S., so the plaintiff should have been able to specify the pharmacy or pharmacies where he filled his prescriptions in the weeks leading up to his stroke.  He also had to have a doctor or other licensed healthcare provider with prescribing authority who wrote and refilled his prescriptions for carvedilol over the course of two years.  The record of these prescriptions and dispensations should have indicated 1) what dose and version of carvedilol plaintiff was prescribed, 2) what company made the tablets or capsules he was dispensed leading up to his stroke, and 3) the lot information for them.  None of this information was included in the Leon decision.  In regard to the statement that plaintiff was supposed to take the medication once per day, it should be noted that the tablets defendant made were indicated to be taken twice a day and only the extended release capsules that defendant did not make were indicated to be taken once a day.  Plaintiff did not save his pill bottle or, apparently, any other pertinent records.  If plaintiff had any proof that he was taking defendant’s carvedilol tablets at all, then he would have had records to indicate if his tablets were dispensed from the one of the 300 lots with a recall for the pill mix-up, even though it was not distributed to where he lived and presumably got his prescriptions filled.  (If his prescriptions were filled by mail, then there also would have been clear records on manufacturer, dosage, lot, etc., also.)  Thus, plaintiff had no product ID evidence to offer because he likely was not dispensed defendant’s carvedilol tablets near the time of his stroke at all, let alone from the single lot with a known issue.  So, unless plaintiff’s complaint offered false factual assertions supporting that he received tablets from defendant’s recalled lot, then this case should have been dismissed at the pleading stage, perhaps with the aid of a motion for more definite statement.

Part of this, of course, is that the defendant made generic carvedilol (and paroxetine), although the Leon decision did not say that.  Any design and warnings claims, however framed, are preempted.  The only potentially viable non-preempted claims would have been manufacturing defect and an affirmative misrepresentation to the plaintiff outside of product labeling.  To satisfy TwIqbal, however, each of these claims required support from detailed factual assertions that plaintiff probably could not have offered in good faith.  There was also the problem that plaintiff claimed he had a stroke because he accidentally took one or more doses of paroxetine instead of his prescribed carvedilol, setting aside that his medical records indicated he skipped his carvedilol on the day of his stroke.  Nobody claims a few tablets of paroxetine causes stroke; this is a drug that has been widely used for more than thirty years without a stroke risk even with chronic use.  Rather, it seems that plaintiff claimed that not having the antihypertensive effect of carvedilol for one or more days caused his stroke.  However, carvedilol is not indicated to prevent stroke and lack of efficacy claims cannot be a basis for recovery for an injury under a fact pattern like this.  Indeed, plaintiff did not even name any experts on medical causation when he had the chance.  Yet, this case lasted for almost four years before defendant won summary judgment.  Such a waste of judicial resources.

We suspect that this case was always just the sort of bad bandwagoning that needs to be checked early.  Plaintiff probably never had a prescription filled with defendant’s carvedilol.  Or whatever carvedilol he was supposed to be taking at the time of his stroke was not made by defendant, possibly because he had been switched to the once daily extended release version.  Assuming he had a current script and was not taking someone else’s medications (a bar to recovery under New York law), records existed that indicated that whatever carvedilol he received leading up to his stroke was not from the recalled lot.  Instead, he merely referenced an apparently completely coincidental recall and muddied the waters for as long as he could.  This also seems like an ideal case for limited early discovery on product ID.  Make plaintiff “show his receipts” before he wastes plenty of time and money with a frivolous case.

On the bright side, the Leon court did stand firm on plaintiff’s argument that defendant’s motion required it to disprove the possibility that plaintiff could have received a tablet of paroxetine mistakenly included in a carvedilol lot, such as from the recalled lot.  Id. at *3 n.2.  This is a fairly straightforward application of the old summary judgment trilogy from 1986.  Defendant’s evidence that the lot was not distributed to where plaintiff lived, which was presumably buttressed by evidence about the circumscribed scope of the issue identified in the recall, was enough to shift the burden to plaintiff.  Id. (citing Celotex and the 2010 advisory committee notes to Fed. R. Civ. P. 56).  Plaintiff had no evidence in the record to carry his burden that he was dispensed defendant’s paroxetine instead of defendant’s carvedilol as he claimed.  That is how Rule 56 is supposed to work.  Plaintiff also lost on lack of proof of causation and just about every other element of his causes of action.  Id. at *4-5.  Plaintiff was shut out, but it could have been even worse as the court did not even address what would have been a winning preemption argument.

While plaintiff did not bring the receipts to show product ID, defendant surely saved some receipts that it would need to seek costs as the prevailing party.  That would help some, but it would probably take fee shifting to deter the next potential plaintiff who considers bringing a suit without any semblance of proof of product ID.  This would be a good case for that too.

Photo of Stephen McConnell

Few issues are as currently controversial as gender affirming care. For some people, the incongruence between gender identity and sex assigned at birth can cause clinically significant distress, recognized as “gender dysphoria” by the American Psychological Association. Treatment can come in various forms, such as social transitioning or medical intervention, which might include chemical treatments or surgery.

There is no doubt where the Trump administration stands on the controversy. Eight days into his second term, the President issued an Executive Order (EO) that sought to end such medical treatment. The Department of Justice (DOJ) followed up with memoranda (DOJ Memos) that set forth how, with certain interpretations of existing law, including the Food, Drug, and Cosmetics Act (FDCA), the EO goal would be achieved. The EO and DOJ Memos refer to gender affirming care as “mutilation.”

In Massachusetts v. Trump, 2026 WL 1584837 (D. Mass. June 3, 2026), fifteen states, the Governor of Pennsylvania, and the District of Columbia sued the President, the then United States Attorney General (AG), and the DOJ challenging the EO and the DOJ Memos. Most of those states had enacted laws mandating insurance for gender affirming care and shielding patients and providers from criminal and civil liability for the provision of such care. The plaintiffs alleged that the EO and DOJ Memos were arbitrary and capricious under the Administrative Procedure Act (APA), and contrary to constitutional right, power, or privilege under the Tenth Amendment. The plaintiffs sought a declaratory judgment.

The defendants moved to dismiss the action, arguing that the plaintiffs lacked standing, that the Memos were not final agency actions, that those Memos announced policies committed to agency discretion by law, and that the claims failed under the Tenth Amendment. The court denied the government’s motion to dismiss. Thus, the federal government lost this initial skirmish involving the off-label use-related showdown over gender affirming care.

Standing

The standing issue arose in the context of the government’s motion under Rule 12(b)(1) challenging the court’s subject matter jurisdiction. Jurisdictional matters come first, so standing led off the festivities. The court held that the plaintiffs had standing to challenge, inter alia, the government’s novel interpretations of the FDCA. The government was not merely articulating enforcement policies, but announced a new interpretation of the FDCA. For example, the EO prioritized investigations and actions against “deception” of consumers about the “long-term side effects of chemical and surgical mutilation.” In one of the Memos, the AG directed investigations of any violations of the FDCA by “manufacturers and distributors engaged in misbranding by making false claims about the on- of off-label use of puberty blockers, sex hormones, or any other drug used to facilitate a child’s so-called ‘gender transition.’” We have discussed the government’s novel attack on off-label use here and here. Those earlier matters took place in the context of subpoenas, whereas this case is a direct challenge to the entire underlying policy.

The government’s EO and DOJ memoranda assert that healthcare providers’ statements regarding gender dysphoria treatments are false claims prosecutable under the FDCA, regardless of whether they are made in conjunction with on- or off-label uses of medication. The federal government has never before made it illegal for healthcare providers to consult with their patients about, and prescribe medication for, treatment of a recognized medical condition.

The court held that the plaintiffs had established an injury in fact. First, the plaintiffs had a legitimate fear of future prosecution (which establishes standing under the APA) because the risk of prosecution here was readily apparent in the record, since the government has already initiated enforcement actions. Second, the plaintiff governments also alleged direct harm that confers standing. Healthcare facilities in the states were expending significant time and resources responding to patient questions, the Memos prompted many facilities to limit the provision of gender affirming care to adolescents, and the delay of such healthcare would lead to more expensive care down the road. Where plaintiffs challenge an FDCA interpretation that would prohibit their conduct, the injury can be traced to the government’s actions. The court also held that the plaintiffs had standing to seek a declaratory judgment and to vindicate their rights under the Tenth Amendment (more on the latter momentarily).

Final Agency Action

The Memos are “rules” under the APA because they announce a new interpretation of statutes. They also “bind and direct inferior employees.” Further, the government’s own statements, which direct enforcement, establish the finality of its actions. There has never been anything remotely clever or subtle about anything the Trump Administration has said or done. Call it transparency or something else, but the Administration’s bluster certainly has the effect of making things easier for the courts. Indeed, the court reasoned that “[b]ased upon the government’s inflammatory rhetoric … and the Memos’ pronounced goals, the purpose of the Memos is to intimidate and harass.” Mission accomplished.

Agency Discretion

Under APA section 701(a)(2), a plaintiff may not obtain judicial review for an agency action “committed to agency discretion by law.” But the Memos’ announce statutory interpretation that would “provide clear standards for enforcement.” The defendants’ “actions are not committed to agency interpretation by law.” Again, lack of subtlety has a price.

The Tenth Amendment

The Tenth Amendment provides: “The powers not delegated to the United States by the Constitution, nor prohibited by it to the States, are reserved to the States respectively, or to the People.” To put it mildly, “Tenth Amendment jurisprudence ‘has traveled an unsteady path.’” But it is now well established that the Tenth Amendment is violated when the federal government acts in excess of its statutory authority in an area traditionally reserved to the states. Medical practice is an area traditionally reserved for the states. Interestingly, the court supported this proposition by citing the recent SCOTUS Skrmetti case, which permitted a state to limit gender affirming care. Turnabout is fair play. Or something like that. The federal government’s FDCA interpretations directly conflict with the plaintiff states’ laws concerning the same type of medical procedures. “Accordingly, Plaintiffs have stated a claim under the Tenth Amendment.”

Photo of Michelle Yeary

Baby formula is not a prescription drug. No doctor hands you a slip of paper for it. The FDA does not regulate it that way. And yet — when a premature newborn receives cow-milk-based preterm formula through a tube in a neonatal intensive care unit, administered by hospital staff, ordered by neonatologists, selected after individualized review of the infant’s gestational age, birth weight, comorbidities, and the risks and benefits of every available feeding option — what is that if not a prescription? Watson v. Mead Johnson & Company, LLC, 2026 IL App (5th) 240936, 2026 WL 1706992 (Ill. App. Ct. June 12, 2026), answers that question. In what we believe is the first ruling by any appellate court on the issue, an Illinois intermediate appellate court reversed a multi-million-dollar verdict and held that the learned intermediary doctrine applies to infant formula used exclusively in a hospital setting under physician oversight. The formal absence of an FDA prescription requirement does not determine the analysis. The clinical context does.

Plaintiff’s son was born more than two months premature. He spent his entire 25-day life in neonatal intensive care. His physicians discussed the risks of prematurity with his mother, including necrotizing enterocolitis (NEC), a potentially fatal inflammatory intestinal disease, before and after birth. During his hospitalization, the baby’s neonatologists ordered his feedings be supplemented with defendant’s cow-milk-based formula for preterm infants. The child developed NEC and underwent three unsuccessful surgeries. Plaintiff sued the formula manufacturer for strict liability design defect, strict liability failure to warn, and negligence. The jury found for plaintiff on the failure-to-warn and negligence counts. Defendant’s posttrial motions were denied.  Id. at *1-3.

Before the appellate court could reach the learned intermediary question, it had to address plaintiff’s threshold procedural argument—the general verdict rule. The general verdict rule provides that when a jury returns a general verdict on multiple theories of liability, the verdict will be upheld if there was sufficient evidence to support any one of those theories—and a defendant who failed to request special interrogatories cannot complain on appeal. Id. at *17. Plaintiff argued that the learned intermediary doctrine applied only to her strict-liability failure-to-warn claim, not to her four theories of negligence, and that because the jury returned a general verdict encompassing negligence as well, defendant was foreclosed from raising the issue on appeal.

The court was not persuaded. It reviewed plaintiff’s four negligence theories: (a) defendant knew or should have known its formula significantly increased NEC risk; (b) it placed a harmful product into commerce; (c) it marketed the product for premature infants despite the risk; and (d) it failed to disclose NEC risks to parents and healthcare providers. The court dissected each in turn. Theory (a) mirrored the strict-liability failure-to-warn claim. Theory (b) was a reframing of the design-defect theory. Theory (c) was a restatement of the failure-to-warn claim—Illinois does not recognize negligent marketing as an independent cause of action. Theory (d) was “effectively identical” to the strict-liability failure-to-warn count. Id. at *19. In the court’s words, plaintiff’s “negligence theories materially overlap with, and in several instances merely restate, the same factual grounds underlying her strict liability design-defect and failure-to-warn claims.” Id. at *18. The negligence count did not set forth four distinct theories but rather “repeated iterations of the same operative facts framed under multiple labels.” Id. at *20. Where negligence is re-packaged strict liability, it is not treated as a separate basis for recovery, and the general verdict rule does not bar appellate review.

With that threshold cleared, the central question was to whom defendant’s duty to warn ran. Defendant said: to the physicians. Plaintiff said: to the mother.

Plaintiff’s position had surface appeal. Baby formula is not an FDA-regulated prescription drug. Parents can, in principle, purchase it online. The learned intermediary doctrine, the argument went, belongs to prescription drugs and prescription devices. Baby formula is a food.

The appellate court disagreed. It started with the rationale underlying the learned intermediary doctrine as articulated by the Illinois Supreme Court—prescription drugs are complex; the prescribing physician weighs a drug’s propensities against the patient’s susceptibilities; the choice the physician makes is an “informed one, an individualized medical judgment bottomed on a knowledge of both patient and palliative.” Id. at *21. That rationale, the court observed, does not turn on whether the FDA has stamped “prescription only” on a label. It turns on whether individualized medical judgment was exercised in selecting and administering the product.

The undisputed evidence put the point beyond argument. The formula was obtained by the hospital directly from defendant. No infant in the NICU could receive it without a physician’s order. The neonatologists who ordered it for plaintiff’s son did so after evaluating his gestational age, birth weight, comorbidities, growth trajectory, the NEC risk associated with formula versus human milk, and the available alternatives. That is exactly “the individualized medical judgment of the type upon which the learned intermediary doctrine is based.” Id. at *22. The formal absence of a prescription requirement did not change what was actually happening in that NICU. The formula was, in the court’s words, “squarely within the class of products for which the learned intermediary doctrine applies.” Id.

Plaintiff pointed to the online availability of the product as evidence that the doctrine should not apply. The court dispatched that quickly. Plaintiff’s son never left the hospital. Plaintiff did not purchase the formula. “Considering the nature of [the formula] and the context in which the product was administered,” id., the court held that the learned intermediary doctrine applied as a matter of law. The duty ran to the physicians. We agree.

The trial court’s refusal to give the learned intermediary instruction was therefore reversible error. “By rejecting the doctrine, the trial court fundamentally skewed the legal lens through which jurors evaluated the evidence because the jury instruction misstated to whom [defendant’s] duty was owed.” Id. at *23. The error was not a technicality. It directed the jury to evaluate whether plaintiff was adequately warned when, as a matter of law, the duty ran to her son’s physicians. Illinois precedent is clear that when a case is tried under an erroneous theory of law, reversal and a new trial are required. Id.

The court also addressed one evidentiary issue it expected to recur on remand—the admission of extensive financial evidence regarding defendant’s revenues, profit margins, and executive compensation, despite the fact that plaintiff sought only compensatory damages and asserted no claim for punitives.

Plaintiff had argued the evidence was relevant to motive, knowledge, and corporate priorities. The trial court agreed and admitted it. The appellate court did not. Under Illinois law, when only compensatory damages are recoverable, a party’s financial condition is irrelevant and often prejudicial — it appeals to the jury’s sympathy and invites a verdict based on the defendant’s ability to pay rather than on evidence of liability. Id. at *24. The financial evidence here was anything but incidental. It was introduced before the first witness was called, highlighted repeatedly during trial, and emphasized in opening statements and closing argument. The jury heard about defendant’s revenues, its parent company’s billions in global sales, and executive compensation packages — none of which bore on whether the product was defective, whether the warnings were adequate, or what the compensatory damages should be. Motive, in any event, was not an element of plaintiff’s burden of proof. On remand, such evidence is inadmissible absent a punitive damages claim. Id.

The learned intermediary doctrine has never been a creature of FDA classifications. It is a doctrine built on clinical reality—a medical professional stands between the manufacturer and the patient, exercises independent expert judgment, and makes an individualized decision to use a product for a specific patient. When a preterm infant formula is ordered by a neonatologist in a NICU, administered through a tube, and selected through precisely that kind of individualized medical assessment, calling it “food” rather than a “prescription drug” does not change the analysis. The label on the can does not determine the law.

Photo of Bexis

We’ve posted twice recently about the potential jurisdictional benefits of the congressionally expanded “federal officer” basis for federal jurisdiction in tort litigation:  28 U.S.C. §1442 (a)(1).  Our first post cited in passing Griffin v. Optum, Inc., ___ F.4th ___, 2026 WL 1239289 (8th Cir. May 6, 2026), as an example of courts already applying Chevron USA Inc. v. Plaquemines Parish, 146 S.Ct. 1052 (2026), “to cases involving other industries.”

On second thought, though, the “other industry” in Griffin − pharmacy benefits managers (“PBMs”) – is sufficiently adjacent to our sandbox that it warrants a post of its own.  Indeed, Griffin is an opioid case, which makes it even more relevant, since plaintiffs make similar allegations against others in the chain of distribution of prescription drugs.  One other salient feature of Griffin is that the plaintiff here was a state attorney general, and opioid cases brought by states have been particularly difficult to remove.

So here’s Griffin.

The state sued four PBMs for supposedly “contributing to an opioid epidemic.  2026 WL 1239289, at *1.

Pharmacy benefit managers enter into service agreements with their clients − federal and non-federal sponsors of health insurance plans − to administer prescription drug programs.  As part of their services, the managers develop formularies, which are lists of prescription drugs covered by a health plan.

Id.  A “central” function of a PBM is negotiating rebates – “post-sale discounts calculated based on the number of consumers that fill a prescription for the manufacturer’s drug.”  PBMs get a cut of these rebates and pass “the remainder” to the various insurance plans that they service.  Id.  PBMs also “determine[ each] beneficiary’s coverage and copayment information.”  Id.  Critically, a couple of the PBM defendants performed these functions under a contract with a federal agency.  Id.

For that reason, they removed the entire state AG action to federal court on the basis that some of their challenged activities were taken under the supervision of a federal officer.  Id.  The state claimed that, by “disclaim[ing] all federal claims, including claims against ‘any federal officer or person acting under any office of the United States for or relating to any act under color of such office,’” in its complaint, it had immunized the suit from federal jurisdiction.  Id. at *2.  The district court fell for that ruse, but the court of appeals did not, and ordered the suit to remain in federal court.  Id.

Federal officer removal was “liberally construed” and “the typical presumption against removal does not apply.”  2026 WL 1239289, at *2 (citations and quotation marks omitted).  The PBMs were “persons” within the meaning of the statute, “acted under the direction of a federal officer,” in a way that had “sufficient connection between those acts and the[plaintiff’s] claims for relief,” and had “a colorable federal defense.”  Id. (citation omitted).

The key in Griffin was “acting under.”  That statutory phrase was “broad, and must be liberally construed.”  Id. (citations and quotation marks omitted).  It includes assisting the government in conducting “basic governmental tasks” such as “providing “the government with a product that it needed or performing a job that the government would otherwise have to perform.”  Id. (citation and quotation marks omitted) (emphasis added).

A PBM working for a government health program qualifies.

[The removing PBMs], at a minimum, acted under the direction of [a federal agency].  The [relevant federal statute] established a comprehensive program of health insurance for federal employees.  Congress [intended] to improve the position of the government with respect to private companies in recruiting the best talent [by] . . . establish[ing] a partnership between [that agency] and private carriers.  The [agency] contracts with private carriers and authorizes the carriers to subcontract with [PBMs] to provide coverage.  The [PBMs] negotiate with manufacturers to obtain discounts, often in the form of rebates, for the drugs that are ultimately included on the plan’s formulary.  When the [PBMs] negotiate with drug manufacturers, they play a key role in the [agency’] effort to carry out its duties . . . to provide prescription drug benefits.  Accordingly, the [PBMs] perform a basic governmental task by performing a task that the government itself would otherwise have to perform.

Id. at *3 (citations and quotation marks omitted).  The removing PBMs thus were subject to a variety of federal “regulations” and “guidelines” concerning “a variety of topics.”  Id. at *4.  These included mandatory contract terms providing federal “access to information at each claim and aggregate level” between the [PBMs] and pharmacies” so that the government could conduct audit “quarterly rebate guarantees, annual reconciliation and payments, actual billing and allocation of rebates, administrative fees, claim payments, fraud and abuse standards, performance guarantees, [and] pharmacy rebates.”  Id. at *4.

Since the state’s claims in Griffin broadly attacked the PBMs’ handling of various rebates, the claims obviously “sufficiently related” to the services the moving defendant was performing for the government.  Id. at *5.  “The State claims that the rebates and fees induced the alleged collusion” to promote opioid overprescription.  Id.  Indeed, the complaint went on and on about rebates.  Id.  So in that sense, the state was hoist on its own petard.

Nor could the state immunize itself from federal officer jurisdiction with boilerplate “disclaimers”:

[N]o disclaimer, however worded, can help the State avoid a causal nexus between [the removing PBMs’] conduct on behalf of the federal government and the State’s] claims.  [The removing PBMs] conduct[] rebate negotiations indivisibly for both plans under the [federal statute] and plans outside of [it].  [They] do[] not have separate rebate agreements with drug manufacturers for [federal] and non-[federal] plans, and none of the resulting discounts or rebates from the negotiations are exclusive to either [type of] plans.  If [the removing PBMs] were liable for negotiating rebates on behalf of private clients, [they] would necessarily also be liable for negotiating rebates on behalf of the federal government—because it is the same negotiation.”  Therefore, the State’s disclaimers do not, and could not, waive claims based on [the removing PBMs’] rebate negotiations undertaken on behalf of both the federal government and private clients.

Griffin, 2026 WL 1239289, at *5 (citations and quotation marks omitted) (emphasis added).  Thus, the remand order was reversed, and the case remains in federal court.

Governmental involvement in the nation’s health care system is ubiquitous.  Thus, we invite our clients and other readers to look for other examples where defendants sued in state court supply either products or services (see our first use of emphasis above) on the same terms and under the same overall regulatory scheme to both federal and non-federal purchasers.  Plaintiffs cannot attack defendants’ (and not just PBMs’) conduct with respect to non-federal purchasers without also implicating federal procurement as well – and disclaimers “do not and [can]not” separate the two for federal jurisdictional purposes.

Photo of Bexis

This post is off-topic – it has nothing to do with prescription medical product liability litigation, or even the law.  Bexis recently took a vacation on the island of Kauai, in Hawai’i, for the sixteenth time.  He hiked, for the tenth time, the first two miles of the Kalalau Trail, to Hanakapiai Beach.  That’s nice, but it’s not really an adventure.  The actual adventure begins at the beach itself.  Bexis is memorializing it here because he is getting old for this kind of thing, and he has not seen any description of this particular adventure in any travel book – even the Ultimate Kauai Guidebook − or online.

Bexis does not want his discovery to disappear without a trace, so he is detailing it here, on his blog.

To replicate Bexis’ adventure, first, do not bother with the difficult stream crossing where the trail meets the Hanakapiai Stream.  Instead make your way down the well-established, albeit unmarked, somewhat rough (big rounded rocks) path along the near (eastern) side of the stream.  You will pass some pleasant day lunch sites and freshwater pools.  Then you must clamber over the last of the large rocks to reach the sandy beach.

Note:  Hanakapiai Beach exists only in the summer months.  The ocean is far too rough and dangerous at any other time of the year to attempt this adventure.  It is also best pursued at low tide.

Facing the ocean (makai in Hawai’ian), adventurers should turn right (east) and follow the Hanakapiai Beach.  The cliffs gradually get closer to the ocean.  Some 50 yards to the east is a large, horizontal cave, perhaps what’s left of a million-year-old or more lava tube.  From the beach you will not see the back of this cave.  There may be a shelf of sand that you have to climb onto before entering the cave.  The amount of sand in the cave varies from visit to visit.

Once inside the cave, your eyes will adjust, and you can see the roof of the cave above.  Walk about 15 yards into the cave.  The ceiling in this part of the cave is quite high, and there is no need to stoop (yet).

After your vision is fully adjusted to the darkness of the cave, turn to the left (again, an easterly direction), facing away from the ocean (mauka in Hawai’ian).  In the distance you should see sunlight from the back entrance to the cave.  The sunlight will be quite distinctive against the cave’s darkness.  If there is no sunlight – abort the adventure.  That means the other end of the cave is filled with sand, and thus passage is impossible.

If, as Bexis did, you can see sunlight, move cautiously towards it.  Keep your hiking stick, or hand, in front of and above your head, because the roof of this part of the cave is much lower.  Bexis had to crawl on his stomach to get through the cave’s back entrance. (here’s a 2018- vintage video of the cave – but the sand was higher in 2026).

The back entrance of the cave leads to a small beach close to the ocean.  If the waves completely cover this beach, don’t go further, because the ocean at Hanakapiai is dangerous.

If this beach at the back of the cave is passable, the next part of the adventure is to climb up the rocks to the shelf that runs along the cliff face to the east of Hanakapiai Valley.  You will need a pair of water-resistant shoes for this next part.  The shelf is rough, but walkable, for about the next 300 yards or so.  Other than occasional spray, the waves should not be much of an issue, although it is never a good idea to turn your back entirely on the ocean.  Most of this part of the adventure involves walking and climbing across bare, sometimes jagged, lava rock.  The rest is residual patches of sand.  The winter storms prevent any plants from growing on the shelf.

Picking your way across this moonscape by the ocean will take 15-20 minutes.

The adventure reaches its destination when the shelf ends at a good-sized – but not nearly as large as Hanakapiai – beach.  The descent from the shelf to this beach is relatively easy, and the sand is every bit as soft as the sand on Hanakapiai Beach.  Here’s a picture:

Depending on the waves, and your inclination, you can spend an hour on this beach.  Bring snacks.  The near side of the beach features another good-sized sea cave (a bit of which is visible on the right side of the picture), which is home to numerous small sea birds, that will undoubtedly be annoyed by any human presence.  Beyond the cave, at the other end of the beach, is the waterfall that drops from the cliff above and cascades down the rocks.  The waterfall lands a pile of rocks rather than on the beach itself, so you can’t stand under it (probably a good idea, since that’s where the rocks came from) but as the picture shows the rocks are covered with lush inter-tidal vegetation.  That area is also home to numerous good-sized, but harmless, black crabs.  Except possibly in the early morning, the towering cliffs provide plenty of shade.

Repeat the process in reverse to return to Hanakapiai Beach.

That’s Bexis’ adventure.  A trip through a cave and then around a rocky, but quite passable point, to a deserted beach with its own cave and a lovely waterfall.  Bexis has successfully done this four times, with his family.  Bexis is now 70 years old, so if he can do it, so can any reasonably fit person who would like to experience a truly deserted beach in Hawai’i.

Hiking through caves and clambering over rocks to secluded beaches may or may not make me a better lawyer, but it’s fun.

Aloha.

Photo of Eric Hudson

We’ve wondered if the widespread use of generative AI was going to produce an increase in the filing of pro se complaints and a marked change in their content. While the results are preliminary, at least one study shows a significant increase in pro se filings following the general availability of AI tools. And our own anecdotal experience suggests that, with the benefit of AI, pro se filings are moving from handwritten one-pagers to lengthy and more sophisticated complaints.

If AI is helping pro se plaintiffs prepare more sophisticated pleadings, we would hope that courts would become less deferential. After all, if generative AI helps pro se litigants bring their claims on a more level playing field, and if they’re bombarding courts with complaints, shouldn’t we see an end to the type of judicial indulgence in favor of pro se plaintiffs that leads to bizarrely pro plaintiff results? While we will continue to hold out hope, today’s case is an example of a court giving a pro se plaintiff the benefit of the doubt and, unfortunately, potentially expanding Texas product liability law into enterprise software.

Continue Reading Pro Se Blues in Texas
Photo of Bexis

Recently, towards the end of a post about preemption and the FDA’s Current Good Manufacturing Practices (“cGMPs”) – also known as Quality Management System Regulations (“QMSRs”) − we noted a “significant change” last February.  “[T]he FDA issued a final rule” that replaced all existing cGMPs “by incorporating by reference an international standard specific for device quality management systems.”  The purpose of the FDA’s changeover was “harmonizing to align more closely with the international consensus standard for devices.  Id.

Thus, as of February, 2026, the cGMPs are no more.  They are now solely known as QMSRs, and the FDA no longer drafts them (and no longer has to bother with the niceties of the Administrative Procedure Act).  Instead, the Agency has incorporated by reference the set of model medical device quality standards created by the International Standards Organization, specifically ISO 13485:2016:

The QMSR incorporates by reference the international standard, ISO 13485:2016, Medical devices – Quality management systems – Requirements for regulatory purposes and clause 3 of ISO 9000:2015, Quality management systems – Fundamentals and vocabulary.  This allows U.S. medical device quality management system requirements to align with internationally recognized standards and reduce burden on regulated industry.

Incorporating by reference ISO 13485:2016 as the foundational quality management system framework for medical device manufacturers promotes consistency in design, production, and lifecycle controls across global markets and now specifically requires risk management.

Incorporating by reference the terms and definitions set forth in Clause 3 of ISO 9000:2015 supports a common understanding of quality management system fundamentals.  However, to the extent any clause of ISO 13485 conflicts with any provision of the FD&C Act and its implementing regulations, the FD&C Act and/or its implementing regulations will control.

FDA, Quality Management System Regulation (QMSR) (updated Feb. 2, 2026).  The FDA also has – but has yet to finalize – one of its infamous “draft guidances” concerning “Quality Management System Information for Certain Premarket Submission Reviews.”  Among other things, this draft guidance makes clear that the QMSRs apply to PMA approved devices, and thus will be relevant to preemption.  Id. at Introduction.

Ordinarily, a copy of ISO 13485:2016 would cost you it’s $250 a pop, payable in Swiss Francs, from the ISO’s website.  But as our prior post helpfully mentioned, read-only copies are available for free here through ANSI (but you would have to register).  We suspect (but aren’t sure) that the availability of free copies is required by the government’s incorporation by reference, since the same website also offers copies of other ISO standards that have been federally adopted.  Ordinarily we would link directly to the document itself, but this one contains a copyright notice.  While there may well be some law out there that forbids American citizens from being forced to pay for copies of the laws that they are supposed to follow, we aren’t going to be the guinea pigs, so you’ll have to download from the above link yourselves.

We did that and have reviewed the actual ISO standards, and have the following mostly preemption-oriented comments:

  • The FDA’s incorporated by reference “requirements” for PMA (and related) devices are extensive, “including design and development, production, storage and distribution, installation, servicing and final decommissioning and disposal of medical devices, and design and development, or provision of associated activities (e.g. technical support).”  The inclusion of “associated activities,” such as “technical support,” should bring many if not all of the activities of “authorized representatives” within the ambit of the MDA’s broad “different from or in addition to” preemption clause, thereby closing a loophole that plaintiffs have exploited.
  • The standard’s inclusion of “other” parties’ that supply “raw materials, components, subassemblies, medical devices, sterilization services, calibration services, distribution services, [and] maintenance service” should reinforce Biomaterials Access Act preemption and perhaps expand preemption to new entities, depending on whether the ISO creates binding “requirements” as to them.
  • ISO 13485:2016 creates a “process approach” to device regulation, and “[a]ny activity that receives input and converts it to output can be considered as a process.”  Thus, the FDA’s blanket incorporation by reference of this standard would seem to create a very broad basis for preemption.
  • ISO 13485:2016 extends to a regulated entity’s “advisory notice” – “subsequent to delivery” – concerning the “use,” “modification,” “return,” or “destruction” of such devices.  Thus, its requirements extend to DHCP letters and recalls.
  • The definition of “labelling” (ISO 13485:2016 uses European spellings) is as broad as the FDA’s traditional definition:  “label, instructions for use, and any other information that is related to identification, technical description, intended purpose and proper use of the medical device.”
  • We’re not sure what to make of the standard’s application in multi-component off-label use situations.  A “note” to the definition of “manufacturer” states:  “Any person who assembles or adapts a medical device that has already been supplied by another person for an individual patient, in accordance with the instructions for use, is not the manufacturer, provided the assembly or adaptation does not change the intended use of the medical device.”  But whatever it means, the incorporated requirements certainly extend to this situation, and thus should be preemptive.
  • The definitions of “medical device” and “product” both expressly include “software.”  We don’t know how that will influence, if at all, the common-law definition of “product,” but for preemption purposes, the incorporated requirements clearly extend to device software.
  • ISO 13485:2016 includes a definition of “service”:  “Service is the result of at least one activity necessarily performed at the interface between the supplier and customer and is generally intangible.”  Services are not products for purposes of strict liability.  The standard includes examples, but they are unhelpful, since none of the examples relate to medical devices.
  • The standard’s “documentation” requirements are notably non-specific.  Take, for example, the “medical device file” (4.2.3).  That file is to include:  a “general description of the medical device, intended use/purpose, and labelling, including any instructions for use,” product “specifications,” “specifications or procedures for manufacturing, packaging, storage, handling and distribution,” “procedures for measuring and monitoring,” “as appropriate, requirements for installation,” and “as appropriate, procedures for servicing.”
  • The other ISO device requirements are just as generalized.  For instance, “design and development inputs (7.3.3)” consist of “functional, performance, usability and safety requirements, according to the intended use,” “applicable regulatory requirements and standards,” “applicable output(s) of risk management,” “as appropriate, information derived from previous similar designs,” and “other requirements essential for design and development of the product and processes.”  They are to “be reviewed for adequacy and approved.”
  • One last example, since it arises frequently in preemption cases, is reporting to the FDA.  Here is the sum total of the applicable ISO language (8.2.3):

If applicable regulatory requirements require notification of complaints that meet specified reporting criteria of adverse events or issuance of advisory notices, the organization shall document procedures for providing notification to the appropriate regulatory authorities.  Records of reporting to regulatory authorities shall be maintained.

That’s it – “document” the “procedures.”

Actually, that’s not it.  What’s most important is what we mentioned in the prior post:  “[A]s part of the PMA process, the FDA actually examines the manufacturer’s proposed processes and then mandates (through premarket approval) what those processes must entail.”  That’s how these general device requirements become specific to each device.  The FDA reviews and approves exactly what and how each device manufacturer proposes to satisfy those general standards for each submitted device.  That approval is preemptive.

The FDA’s online statement and draft guidance both reference “harmonization” with international standards.  Whether that will have any impact on preemption, we don’t know, but it might create an additional argument for preemption of bizarre state-law duties that don’t harmonize with anything – such as the purported duty to innovate one’s products claimed in the pending Gilead litigation.  The draft guidance states:

If FDA finds there has not been a demonstration that the methods used in, and the facilities and controls used for, manufacturing, processing, and packing a device conform to the QMSR, including because there is insufficient information submitted in the PMA, FDA is required to deny approval of the application under the Act.

Id. at p.3 (footnote omitted).  This is further confirmation that the FDA reviews each of these things during pre-market approval, and thus that any divergent state-law requirements are preempted.

Like the former cGMPs, the QMSRs themselves do not specify the details of any of the processes that applicants for approval must use.  To the contrary, they are intentionally very broad and capacious, allowing applicants to decide what each new device needs.  For example, the guidance section on “design and development requirements” is four pages long and – like the ISO itself − does not contain a single specific requirement, leaving that choice to the applicant.  Id. at 10-13.  These contain requirements for “identification,” “description,” “traceability,” keeping “records,” “reviews,” and the like.  While the applicant must “describe” “Any test methods, protocols, acceptance criteria, and as appropriate, statistical techniques with sample descriptions/sizes and supporting rationale,” id. at 11, the guidance never dictates what any of these items must be.  That is reserved for individual PMAs.  Thus, to our way of thinking, any attempt by a plaintiff to dictate specifics would necessarily be “different from or in addition to” what the QMSRs require, and therefore preempted.  All of the other requirements listed in the FDA draft guidance are similarly general.

It will no doubt be a while before medical devices subject to the FDA’s new QMSR/ISO 13485:2016 regime start appearing in prescription medical product liability litigation, but it is never too early to start thinking about how these regulatory changes will affect our efforts to defend those devices when litigation inevitably reaches them.

Drug & Device Resources

Bexis’ Book

FDA-regulated products now account for an estimated one-fifth of overall economic activity in the U.S. They have also been the focus of a litigation explosion. This timely guide covers all aspects of litigation involving drugs, medical devices, vaccines and other FDA-regulated prescription products.
Learn more about this book