New FDA preemption decision in Celebrex litigation

In re Bextra and Celebrex Marketing Sales Practices and Product Liability Litigation, 2006 WL 2374742 (N.D.Cal.)

Judge Breyer is handling this multidistrict litigation, consisting of putative class actions arising from the marketing and sale of Celebrex. Non-steroidal anti-inflammatory drugs ("NSAIDs") have been widely used for pain relief for several years. NSAIDs, however, pose a risk of gastrointestinal toxicity, which results in thousands of deaths every year. Pfizer developed Celebrex, a NSAID known as a COX-2 inhibitor, with the hope that it would have fewer gastrointestinal side effects than traditional NSAIDs. Pfizer marketed Celebrex both to consumers and to medical professionals as a solution to chronic pain.

Plaintiffs allege that Celebrex’s marketing was deceptive because Pfizer (1) suppressed data showing the cardiovascular risks associated with the use of Celebrex; (2) falsely claimed that the use of Celebrex had fewer gastrointestinal side effects than traditional NSAIDs; and (3) falsely claimed that Celebrex provided superior pain relief and safety over traditional NSAIDs. Celebrex was a very successful drug, accounting for billions in sales, at a price premium that was an order of magnitude greater than the price of traditional NSAIDs. Plaintiffs allege that, had Pfizer not falsely advertised Celebrex, they would have purchased equally effective and less expensive drugs. The putative class seeks to represent all U.S. end-payors, including consumers and third-party payors who bought or paid for Celebrex not for resale, from December 1, 1998 until now.

Plaintiffs have four claims: (1) RICO; (2) state consumer protection laws; (3) unjust enrichment; and (4) breach of warranty. Pfizer argued that all these claims were preempted by the FDCA and the FDA’s regulatory authority, and made other arguments as well. This opinion addresses only preemption.

The FDA required Celebrex to be marketed in accordance with its approved label, which included a warning for aggravated hypertension but did not otherwise warn of cardiovascular risks. The FDA had access to study results that might have raised a red flag, but the agency didn’t act. Plaintiffs allege, however, that Pfizer also had results that would have reinforced the concerns raised by that study, and that the FDA would have taken the problem more seriously had Pfizer not concealed this until 2001.

In 2001, an FDA Advisory Panel reviewed the cardiovascular risks of Vioxx and Celebrex. The result was a requirement of a new warning on the Vioxx label, but the FDA specifically determined that the overall rate of serious adverse cardiovascular events for patients taking Celebrex was no higher than in patients taking other NSAIDs, and the FDA-approved revised label reflected this. In 2005, the FDA concluded that the benefits of Celebrex outweigh the risks in appropriate patients and therefore Celebrex should remain on the market as a prescription drug, but ordered Pfizer to include a black boxed warning on the Celebrex label that highlights the potential increased risk of serious adverse cardiovascular events.

Back when the FDA first approved Celebrex, it warned Pfizer that "any promotional activities 'that make or imply comparative claims about the frequency of clinically serious GI events compared to NSAIDs or specific NSAIDs will be considered false and/or misleading....' " In 2005, the FDA required an upgrade in the GI toxicity warning on the label to a black box warning that NSAIDs, including Celebrex, “cause an increased risk of serious gastrointestinal adverse events including bleeding, ulceration, and perforation of the stomach or intestines, which can be fatal.”

How to apply preemption principles to the complaint? The complaint alleges that Celebrex’s marketing was unlawful because it didn’t disclose increased cardiovascular risk. Essentially, plaintiffs argue that Pfizer should have included an additional warning not rquired by the FDA on the Celebrex label and advertising.

Many courts have held that “failure to warn” claims don’t conflict with FDA regulations and are therefore not preempted, as FDA requirements impose only minimum standards. Other courts, however, have found preemption, at least where the FDA actually considered and rejected a similar warning.

The FDA recently issued a preamble to a Final Rule on labeling that reversed its longstanding position on preemption. The FDA now "believes that State laws conflict with … Federal law when they purport to compel a firm to include in labeling or advertising a statement that FDA has considered and found scientifically unsubstantiated." The FDA expressly disagrees with those cases that have held that state-law failure to warn claims are not preempted. The FDA has ultimate authority over a label, so the manufacturer can’t change it without permission. Moreover, the FDA’s label is both a floor and a ceiling. Additional disclosures of risk information can make a manufacturer liable under the FDCA if the additional statement is unsubstantiated or otherwise false or misleading. (Has anyone ever heard of the FDA telling a manufacturer that it’s disclosing too many risks?)

The FDA was concerned that additional disclosures could, instead of protecting patients, confuse the benefit/risk profile necessary to make informed judgments about drug use. And such disclosure requirements would be imposed by nonexpert judges and juries, often on behalf of a single individual or group without consideration of others’ interests. The result would be defensive labeling, which could result in underuse of good drugs. (Ah, the old Type I/Type II error problem. If the FDA is certain it doesn’t make many mistakes approving beneficial drugs, then it’s wise to worry about underuse.)

The court determined that the preamble was entitled to deference. Following the Supreme Court in Geier, the court found that Congress delegated implementing authority to the FDA; the subject matter is technical; and the relevant history and background are complex and extensive. The agency is uniquely qualified to understand the likely impact of state requirements. Plaintiffs argued that a preample isn’t a regulation or even an interpretive rule, but the court held that it was enough of an expression of agency opinion to justify deference. True, the consistency of the FDA’s position matters, and FDA used to hold that its standards were minimum ones rather than both minimum and maximum. In fact, in 2000, when the FDA published the proposed drug rule (the rule to which the preemption preable is attached), it took a no-preemption position. The change came in 2001, after a change in administration, when the FDA began to submit pro-preemption amicus briefs. But changing position after executive branch change is not inherently illegitimate, and the FDA’s position has been consistent since then.

The FDA's view is that a claim is preempted if the FDA determined that the warning the plaintiff seeks to impose is not supported by the evidence before the FDA. The FDA does not have to expressly determine that the warning would be false and misleading, although the FDA has suggested that an unsubstantiated statement is indeed false and misleading. (Of course, the DC Circuit’s First Amendment jurisprudence has been very hostile to this position when it’s actually used to regulate claims, for example on supplements, as opposed to being used defensively by drug companies. But, presumably, the logic behind the DC Circuit’s rule means there are also First Amendment problems with imposing liability for failure to warn. But Pfizer doesn’t want to make that argument if it can rely on less sweeping preemption claims.)

The allegation that Pfizer withheld material cardiovascular risk data from the FDA does not change the preemption analysis. The court pointed out that “[t]he law is well established that a claim premised on a drug manufacturer's failure to provide data to the FDA is preempted.” Thus, the cardiovascular risk claims were all preempted.

Pfizer also argued for preemption of the other theory of liability, that Pfizer falsely claimed that Celebrex had fewer GI complications than other NSAIDs and was more effective. Pfizer noted that the FDA requires drug companies to submit all advertising to the FDA's Division of Drug Marketing, Advertising, and Communications ("DDMAC"). DDMAC reviews the advertisements for compliance with the FDCA and FDA regulations on advertising, and has the authority to require a company to stop running a particular advertisement or to run a corrective promotion.

And how well does that work out? Money quote from DDMAC's director:

"We get complaints from consumers and physicians who call us up and say, 'Tom, how can you allow that TV ad to be on?'" Abrams says. "They're flabbergasted when we say, 'We didn't approve it before it went on TV.' Often, we're seeing it at the same time as the American public. DDMAC has limited resources and we use our limited resources as effectively as we can to do our job."

DDMAC usually takes long enough to act that many campaigns are over before DDMAC objects. DDMAC has been notably quiescent for the same period of time that the FDA has argued in favor of preempting state-law failure to warn claims. And the Washington Legal Foundation has a policy of responding to every DDMAC letter that issues on First Amendment grounds, just to show the antiregulatory flag.

Still, Pfizer submitted its challenged Celebrex advertisements to DDMAC and, with a few exceptions, the DDMAC did not object. According to Pfizer, this is necessarily equivalent to a FDA determination that the ads are accurate and strike a fair balance between the benefits and risks of Celebrex. Therefore, any claim that such ads were deceptive conflicts with the FDA's determination and are impliedly preempted.

The court was correctly unwilling to take preemption that far. There was no record evidence that the FDA actually reviewed all the submitted ads, let alone that FDA review means a determination that the ad isn’t misleading. The FDA’s silence on whether false advertising claims should be preempted was also significant in comparison to its position on failure to warn claims.

Pfizer also argued that particular ads identified in the complaint are consistent with the FDA-required label and therefore any claims based on those advertisements are preempted. Plaintiffs' claims, however, are that the ads implied that Celebrex is superior to other NSAIDs because it causes fewer gastrointestinal symptoms, a claim which the FDA expressly determined would be false and misleading. Given that Pfizer was arguing that its ads did not imply GI superiority or greater efficacy and were thus not misleading as a matter of law, the court was unwilling to make a determination at this stage of the case.

Last, Pfizer argued that the FDA, not a court or a jury, should initially decide whether Pfizer's ads are misleading because plaintiffs' claims fall within the "primary jurisdiction" of the FDA. The primary jurisdiction doctrine applies to issues within the special competence of an administrative agency, and allows courts to route threshold decisions to that agency. The court easily rejected this argument. The issue isn’t whether Celebrex has fewer GI complications than other NSAIDs. The FDA already determined it doesn’t. The issue is whether Pfizer falsely claimed that Celebrex was superior, and courts and juries routinely decide false advertising cases.

With failure to warn off the table, false advertising may be increasingly important in pharmaceutical cases, even as issues of reliance and causation make consumer protection class actions difficult to maintain. (Third-party payors may have an easier time aggregating claims, as long as they have standing.) In another judicial system, one might expect a probabilistic or “fraud on the market” theory to emerge, but I hardly expect it will in these times.