Not too long ago, we tried to extrapolate from a doctoral thesis on quantum dots to lessons for litigation.  That “[q]uantum dots are between one-billionth of a meter and one-hundred-millionth of a meter in size” emphasized that “appreciating the scope and scale of what is being discussed can be critical.”  In the spate of litigation over alleged contaminants and degradation products in a variety of over-the-counter and prescription drugs, it seems rare that the plaintiffs want to talk about scale beyond that a particular chemical can be detected at all.  Not us.  Trace chemicals in a sample are often discussed in terms of parts per million (“ppm”).  The amount of a chemical that is known to be harmful (e.g., carcinogenic) in a particular organism might be discussed in terms of grams of exposure per kilogram of body weight of the organism, sometimes with a frequency included.  The number of molecules in a mole of a substance is about 602,214,076,000,000,000,000,000.  The number of molecules in a gram of substance depends on the molecular weight of the substance, but it will typically be in the sextillion range.  For a little context, 1 ppm would be one person out of ten stadiums full of 100,000 people each.  If that person weighs 75 kg, then a risk threshold (or lowest observable effect level) of 1 gram per kilogram of body weight per day would require 75 grams per day.  If we are talking about 200 mg of an NSAID, then that person would need to take 375 of them per day.  If we are talking about a trace contaminant or degradation product that occurs at 2 ppm in that NSAID, then that person would need to take 187,500,000 of those NSAIDs per day to hit that threshold.  We could probably go on discussing the number of molecules, the molecular weight of benzene, or other math-adjacent stuff that would make Thomas Dolby proud.  We will not, but this exercise is actually relevant to a case based on a claim that the medical product at issue has a detectable trace amount of benzene or some other contaminant or degradation product in it.

This week, we have two separate purported class actions for consumer protection over different OTC drugs that were dismissed by two very different courts on preemption grounds.  Kouyate v. Harvard Drug Grp., LLC, No. 1:24-cv-6223-GHW, 2025 U.S. Dist. LEXIS 190152 (S.D.N.Y. Sept. 26, 2025), concerned an acne face wash that contained benzoyl peroxide.  Leonard v. CVS Pharmacy, Inc., No. 5:24-cv-06280-EJD (N.D. Cal. Sept. 29, 2025), concerned cold and flu medicines that contained guaifenesin.  Although we used the past tense above, each plaintiff was given the chance to replead non-preempted claims.  Each also based his/her claims on the purported presence of small but unspecified amounts of benzene in their respective products and the respective manufacturer’s failure to deviate from the FDA-required labeling to describe the purported benzene.  We think that these decisions are close enough to discuss together, adding to our many prior posts about similar claims, including with these same drugs.  See here, here, and here for a sampling.  Kouyate came out first and has a more extensive preemption discussion, so we will address it first.

The plaintiff in Kouyate bought his OTC acne wash knowing it had benzoyl peroxide and purportedly believing it did not have any benzene in it.  (The complaint actually alleged that plaintiff filled a prescription, but that seems like the same sloppiness the court noted from the lawyer re-using complaints.  2025 U.S. Dist. LEXIS 190152, *9 n.3.  This is an OTC product.) He alleged that benzoyl peroxide can degrade into benzene over time, and that an “independent laboratory” found his acne wash had “unsafe levels of benzene” in it.  Id. at *1.  Based on this, he brought a class action under New York law for false advertising, common law fraud, and negligence per se based a state criminal statute banning misbranded and adulterated drugs.  To us, these allegations scream preemption, both express preemption and under Buckman.  Yet, the court offered a careful analysis of the preemption issues, an analysis that we think would have supported not giving plaintiff a third shot to plead non-preempted claims.  We also think that the court’s conclusions that plaintiff had not pleaded the false advertising claim—which required an affirmative misrepresentation—or a common law fraud claim—which required the basic fraud elements pled with particularity—also supported not giving plaintiff another shot on those claims.  Id. at *47-56.  (Evaluating the adequacy of pleading first, as we usually advise, also might have helped.)  It also should have been pretty clear that any attempt at negligence per se, even though based on a violation of state law, would be preempted because it amounted to a stop selling claim.  We are getting a little bit ahead of ourselves, though.

The product at issue in Kouyate was governed by a monograph that set out specific requirements for the labeling of a benzoyl peroxide containing acne wash, including the percentage of benzoyl peroxide, a list of inactive ingredients, specific warnings, and directions.  The monograph does not require or otherwise address any disclosure of the possibility that some small amount of benzene might be created by the degradation of the benzoyl peroxide.  The labeling for defendants’ products uncontestably complied with the monograph requirements and did not include any affirmative misrepresentations (e.g., “This has no benzene,” “None of the ingredients will ever degrade into another chemical substance at all.”).  Keeping in mind that the plaintiff did not assert any product liability claims that would have been exempted from the express preemption provision for OTC drugs, the simple conclusion was that any change in labeling plaintiff claimed was required to avoid violating the state false advertising statute and/or common law fraud requirements was going to be “different from or in addition to [or] otherwise not identical with” the FDCA requirements imposed through the monograph.  Id. at *24-26.  Relying heavily on the Second Circuit’s decision in Critcher, which evaluated the express preemption provision for cosmetics in the context of a case about the labeling of the amount of skin cream in a container, the court rejected the argument that the FDA requirements were insufficiently specific to preempt plaintiff’s claims about additional labeling language.  Id. at *27-31.  Indeed, benzene, whether present at 2 ppm or more, is not “an active or inactive ingredient that needs to be disclosed in the label” because it was not “a purposefully added component of the drug.”  Id. at *27.  (Not long ago, we discussed a decision that postulated that a quasi-manufacturing defect claim could exist because the allegedly harmful substance was not purposefully added to the product.)

Not long before plaintiff bought his acne wash, FDA had advised that benzoyl peroxide products with more than 2 ppm of benzene should not be sold.  However, the monograph had not yet changed, and the acne wash complied with current requirements.  That did not change the analysis.  Nor did the related argument that the manufacturer should have updated its labeling through the same CBE mechanism that confounded the Supreme Court in Levine.  But Kouyate recognized that monograph drugs like the acne wash could not use the CBE regulations.  Id. at *34.

That left the negligence per se claim, which was designed to avoid preemption based on the carveout from preemption under 379r(f) when state requirements identical to FDCA requirements are the basis for liability.  Plaintiff argued that the same labeling issues preempted under his other theories were not preempted because the drug was misbranded under state law; he also argued that the drug was adulterated under state law because having benzene above 2 ppm rendered it harmful, worthless, and illegal to sell.  The problem for plaintiff was that trace benzene from the degradation of benzoyl peroxide does not make the acne wash misbranded or adulterated under the FDCA.  Id. at *38-41.  If, as plaintiff claimed, degradation into benzene is a known, inevitable process, then no acne wash could contain benzoyl peroxide.  Yet, FDA considers these benzoyl peroxide drugs, otherwise compliant with the monograph, to be safe and effective.  Id. at *40.  The mere allegation that the trace amount in plaintiff’s particular product exceeded 2 ppm does not matter because the monograph does not impose any such limit.  The only sources of any such limit are FDA guidances and other non-binding authority.  Id. at *40-43.  Plaintiff’s other arguments, which did not really tie to recognized state law anyway, were unavailing.  We do think that plaintiff’s theories of liability sounded like the stop selling arguments rejected in Bartlett and many cases since.  So, it is hard for us to see how plaintiff can plead non-preempted non-products claims based on the mere presence of benzoyl peroxide and labeling compliant with the current monograph.

The preemption analysis in Leonard was shorter, but no less correct.  The generic OTC drug had an approved ANDA, but it also had monographs relevant to plaintiff’s theory that the defendants had to disclosure that an inactive ingredient, carbomer, was manufactured with benzene.  Like Kouyate, there was certainly no FDA requirement to disclose how the carbomer was made or to refrain from using carbomer made with benzene.  The regulatory requirements were also in some flux when plaintiff purchased her medication in July 2024.  At that time, per FDA guidances, carbomers used as inactive ingredients in drugs were supposed to be in the process of transitioning from a monograph limit of “up to 5,000 ppm” of benzene to a limit of no more than 2 ppm.  Any new requirements, however, were not effective until August 2026.  Unlike in Kouyate, the plaintiff in Leonard did not allege specific testing of benzene in her particular products, just that its choice of carbomer allowed for the possibility of benzene at some level.    She also did not claim that the products failed to comply with actual FDA requirements.  Based on the Dougherty case from the same court, which involved similar claims about a similar product, that meant that any claims based on alleged adulteration or misbranding were preempted.  Id. at *5.  Plaintiff tried to use the decidedly bad decision from the same court in Holley, but the Leonard court recognized that impossibility preemption and express preemption are different.  Id. at *6.  The plaintiff’s claims ran afoul of express preemption by trying to impose requirements different than what the FDCA imposed.  The court also rejected the argument that plaintiff’s claims under California’s version of the federal Sherman Act were automatically parallel claims not subject to express preemption.  But the challenge to “the FDA-approved content and label as adulterated and deceptive” under any claim did attempt “to impose a requirement that is different from or in addition to the federal requirement.”  Id.  As long as plaintiff continues to assert non-product liability claims, it is hard to see how they will not be preempted in Leonard as in Kouyate. 

Given how many times we have written about the preemption of claims based on alleged benzene in FDA-regulated products, including those containing benzoyl peroxide and guaifenesin, we doubt that these will be the last two complaints asserting slightly different versions of the same preempted claims.