T.H. v. Novartis Pharmaceuticals Corporation
This text of 407 P.3d 18 (T.H. v. Novartis Pharmaceuticals Corporation) is published on Counsel Stack Legal Research, covering California Supreme Court primary law. Counsel Stack provides free access to over 12 million legal documents including statutes, case law, regulations, and constitutions.
Opinion
Cuéllar, J.
*154
Under California law, a brand-name drug manufacturer has a duty to warn of known or reasonably knowable adverse effects arising from an individual's use of its drug. (See
Stevens v. Parke, Davis & Co.
(1973)
Plaintiffs' mother, J.H., was prescribed terbutaline, a generic form of the brand-name drug Brethine, to suppress premature labor during her pregnancy. Plaintiffs T.H. and C.H. were born full term, but were diagnosed with developmental delays at three years of age and autism by the time they turned five. Through their father as guardian ad litem, the minors allege that those responsible for the terbutaline label knew or should have known-based on studies of the drug's effects in rats and in humans-that the drug posed a serious risk to fetal brain development. They further allege that the drug's label unreasonably failed to include a warning about this risk.
*340 Federal law explicitly conveys to the brand-name manufacturer-and only that manufacturer-the responsibility to provide an adequate warning label for both generic terbutaline and its brand-name equivalent, Brethine. As explained in more detail below, only the brand-name drug manufacturer has unilateral authority to modify the drug's label by adding to or strengthening a warning. Generic drug manufacturers are required to follow the brand-name manufacturer's label to the letter. Accordingly, the manufacturer of Brethine controlled both the form and content of the terbutaline warning label.
Plaintiffs brought suit against defendant Novartis Pharmaceuticals Corporation (Novartis), which manufactured Brethine until December 2001, and aaiPharma Inc. (aaiPharma), which purchased the rights to and manufactured Brethine thereafter-using the same label Novartis had used-when plaintiffs' mother was prescribed the generic bioequivalent in 2007. Plaintiffs claim that Novartis knew or should have known that its warning label failed to alert pregnant women or their physicians to the risk Brethine posed to fetal brain development; that manufacturers of terbutaline were compelled by federal law to include Brethine's deficient label on their own products; that it was foreseeable Novartis's successor (aaiPharma) would not change or update Brethine's deficient label; and that in reliance on the deficient warning label, plaintiffs' mother was prescribed terbutaline, which adversely affected plaintiffs' developing brains in utero. What Novartis asserts in response is that its duty to provide a safe and adequate warning label for Brethine did not encompass those who were prescribed terbutaline in reliance on the Brethine label. Novartis further contends that any such duty should not extend to those who were exposed to terbutaline after Novartis ceased manufacturing Brethine and sold its rights in the drug to aaiPharma.
Such contentions, and the case in which they arise, reach us at a very early stage in the litigation. In reviewing a demurrer, we ask only whether the
*156
plaintiff has alleged-or could allege-sufficient facts to state a cause of action against the defendant. (
Schifando v. City of Los Angeles
(2003)
I. BACKGROUND
From a certain perspective, the claim underlying this lawsuit is quite straightforward. Plaintiffs T.H. and C.H., who are fraternal twins, sued defendant Novartis for negligence and negligent misrepresentation arising from Novartis's failure to warn of the risks of Brethine, an asthma *341 drug sometimes prescribed "off label" to stop or slow preterm labor. Plaintiffs allege that Novartis knew or should have known that Brethine carried a substantial risk of causing developmental and neurological damage to the fetus, yet failed to warn of this risk.
What removes this case from the realm of the ordinary is that plaintiffs' mother was never prescribed Brethine. Rather, she-like many pregnant women experiencing premature labor -was prescribed terbutaline sulfate (terbutaline ), the generic bioequivalent drug. Moreover, Novartis stopped manufacturing Brethine and sold all rights to the drug in 2001, six years before plaintiffs' injury. During the period it was the brand-name manufacturer, however, Novartis had the legal duty to disclose Brethine's known and reasonably knowable risks in the drug's warning label. All generic manufacturers, in turn, had a specific legal responsibility regarding the label: to ensure the terbutaline label was identical to the Brethine label. We therefore examine plaintiffs' allegations against the backdrop of the distinctive legal framework governing labeling for brand-name and generic pharmaceuticals.
Free access — add to your briefcase to read the full text and ask questions with AI
Cuéllar, J.
*154
Under California law, a brand-name drug manufacturer has a duty to warn of known or reasonably knowable adverse effects arising from an individual's use of its drug. (See
Stevens v. Parke, Davis & Co.
(1973)
Plaintiffs' mother, J.H., was prescribed terbutaline, a generic form of the brand-name drug Brethine, to suppress premature labor during her pregnancy. Plaintiffs T.H. and C.H. were born full term, but were diagnosed with developmental delays at three years of age and autism by the time they turned five. Through their father as guardian ad litem, the minors allege that those responsible for the terbutaline label knew or should have known-based on studies of the drug's effects in rats and in humans-that the drug posed a serious risk to fetal brain development. They further allege that the drug's label unreasonably failed to include a warning about this risk.
*340 Federal law explicitly conveys to the brand-name manufacturer-and only that manufacturer-the responsibility to provide an adequate warning label for both generic terbutaline and its brand-name equivalent, Brethine. As explained in more detail below, only the brand-name drug manufacturer has unilateral authority to modify the drug's label by adding to or strengthening a warning. Generic drug manufacturers are required to follow the brand-name manufacturer's label to the letter. Accordingly, the manufacturer of Brethine controlled both the form and content of the terbutaline warning label.
Plaintiffs brought suit against defendant Novartis Pharmaceuticals Corporation (Novartis), which manufactured Brethine until December 2001, and aaiPharma Inc. (aaiPharma), which purchased the rights to and manufactured Brethine thereafter-using the same label Novartis had used-when plaintiffs' mother was prescribed the generic bioequivalent in 2007. Plaintiffs claim that Novartis knew or should have known that its warning label failed to alert pregnant women or their physicians to the risk Brethine posed to fetal brain development; that manufacturers of terbutaline were compelled by federal law to include Brethine's deficient label on their own products; that it was foreseeable Novartis's successor (aaiPharma) would not change or update Brethine's deficient label; and that in reliance on the deficient warning label, plaintiffs' mother was prescribed terbutaline, which adversely affected plaintiffs' developing brains in utero. What Novartis asserts in response is that its duty to provide a safe and adequate warning label for Brethine did not encompass those who were prescribed terbutaline in reliance on the Brethine label. Novartis further contends that any such duty should not extend to those who were exposed to terbutaline after Novartis ceased manufacturing Brethine and sold its rights in the drug to aaiPharma.
Such contentions, and the case in which they arise, reach us at a very early stage in the litigation. In reviewing a demurrer, we ask only whether the
*156
plaintiff has alleged-or could allege-sufficient facts to state a cause of action against the defendant. (
Schifando v. City of Los Angeles
(2003)
I. BACKGROUND
From a certain perspective, the claim underlying this lawsuit is quite straightforward. Plaintiffs T.H. and C.H., who are fraternal twins, sued defendant Novartis for negligence and negligent misrepresentation arising from Novartis's failure to warn of the risks of Brethine, an asthma *341 drug sometimes prescribed "off label" to stop or slow preterm labor. Plaintiffs allege that Novartis knew or should have known that Brethine carried a substantial risk of causing developmental and neurological damage to the fetus, yet failed to warn of this risk.
What removes this case from the realm of the ordinary is that plaintiffs' mother was never prescribed Brethine. Rather, she-like many pregnant women experiencing premature labor -was prescribed terbutaline sulfate (terbutaline ), the generic bioequivalent drug. Moreover, Novartis stopped manufacturing Brethine and sold all rights to the drug in 2001, six years before plaintiffs' injury. During the period it was the brand-name manufacturer, however, Novartis had the legal duty to disclose Brethine's known and reasonably knowable risks in the drug's warning label. All generic manufacturers, in turn, had a specific legal responsibility regarding the label: to ensure the terbutaline label was identical to the Brethine label. We therefore examine plaintiffs' allegations against the backdrop of the distinctive legal framework governing labeling for brand-name and generic pharmaceuticals.
On review of a demurrer, we accept as true all properly pleaded facts. (
*157
Shirk v. Vista Unified School Dist.
(2007)
A. Federal Regulation of Drug Labeling
The Food, Drug, and Cosmetic Act (FDCA;
In 1984, Congress enacted the Hatch-Waxman Act. (
So under the federal scheme, "brand-name and generic drug manufacturers have different federal drug labeling duties." (
PLIVA
,
supra
, 564 U.S. at p. 613,
*342
FDA regulations require the brand-name drug manufacturer to update the warning label "as soon as there is reasonable evidence of an association of a serious hazard with a drug; a causal relationship need not have been proved." (
Because federal regulations preclude generic manufacturers from unilaterally altering the warning labels on their drugs (
PLIVA
,
supra
, 564 U.S. at p. 617,
B. Terbutaline, Brethine, and Novartis
Terbutaline is a beta-adrenergic agonist, acting upon the beta2 receptors in smooth muscle tissue and causing muscles to relax. The drug was originally developed by Draco, a Swedish company, and released for use as a bronchodilator to treat asthma. In 1974, the FDA approved terbutaline as a treatment for asthma in the United States. Astra AB (and later, AstraZeneca LP) licensed the right to manufacture and market terbutaline in its oral form to Ciba-Geigy (a predecessor to Novartis) under the brand name Brethine. Novartis owned the NDA for Brethine until 2001.
In 1976, a Swedish physician with ties to Draco published the results of a small study indicating that terbutaline was safe and effective as a tocolytic-a drug to suppress premature labor in pregnant women-on the theory that the drug could relax uterine smooth muscle tissue. Terbutaline subsequently gained wide acceptance as a tocolytic, but neither Novartis nor any other manufacturer sought FDA approval for this off-label use. 1 Later studies cast doubt on the safety and efficacy of terbutaline as a tocolytic.
*159 In 1978, a study published in the British Journal of Obstetrics and Gynaecology questioned the validity and conclusions of the original Swedish report. According to *343 plaintiffs' complaint, the British study warned that the benefits of this class of drugs on preterm labor was " 'not yet established,' " that the evidence was " 'too scanty to make conclusions about side effects possible,' " and that other data suggested " 'that labor inhibitors are potentially dangerous.' " A year later, a study published in the American Journal of Obstetrics and Gynecology reported adverse effects in both the pregnant mother and in the fetus following terbutaline exposure.
A team of American clinical investigators in 1982 sought to replicate the results of the 1976 Swedish study. They could not. In fact, the investigators were unable to find any benefit among the pregnant mothers who had been prescribed terbutaline as compared to those who received a placebo. A 1984 study published in the Journal of Reproductive Medicine similarly failed to confirm any benefits.
**25 In 1985, Dr. Theodore Slotkin and a team of Duke University Medical Center researchers found that a single dose of terbutaline given to pregnant rats interfered with an enzyme necessary for neuronal development in the fetal brain. Dr. Slotkin's study showed that terbutaline can cross the placenta and fetal brain barrier in sufficient quantities to affect brain development. Other studies in the 1980s revealed that children born to mothers who had received a different beta-adrenergic agonist had poorer academic achievement and were more likely to have impairments in vision and language development than children born to mothers who did not receive such treatment.
In 1989 and 1990, Dr. Slotkin published studies showing that terbutaline may interfere with the fetus's neurobehavioral development, presumably through its effects on receptors in the fetal cerebellum. Shortly thereafter, in 1992, scientists from the University of Texas undertook a comprehensive and critical evaluation of the literature relating to terbutaline and concluded, in a study published in the American Journal of Obstetrics and Gynecology, that the drug had not been shown to arrest preterm labor and that chronic exposure may adversely affect the fetus. A 1995 meta-analysis by researchers from the University of Pennsylvania likewise concluded that the relevant literature did not support the claimed benefit from maintenance tocolytic therapy. The American College of Obstetricians and Gynecologists (ACOG) subsequently issued a "Technical Bulletin on Preterm Labor" to its more than 40,000 members, which noted the asserted benefit from maintenance tocolytic therapy lacked any evidentiary basis and warned that the potential risks of such therapy, to both the mother and the fetus, were well documented. ACOG's bulletin stated that the risk associated with beta-mimetic agents (such as terbutaline ) appeared greater than that associated with other tocolytic *160 agents. In 1997, the FDA's Associate Commissioner for Health Affairs issued a "Dear Colleague" letter, which endorsed ACOG's assessment of the benefits and dangers of long-term tocolytic therapy.
In 2001, the German Central Institute of Mental Health issued a report concluding that children whose mothers had received beta-agonist tocolysis had a significantly higher rate of psychiatric disorders and psychopathology, and that such children scored lower on psychometric tests of cognitive development. Dr. Slotkin's Duke team released another study in October 2001, which revealed that beta2 receptors in the fetal brain, unlike those in mature brains, do not desensitize when exposed to continuous doses of terbutaline. Instead, the fetal receptors intensify their sensitivity to terbutaline and thus increase their *344 response to the drug as the dosage increases (and the brain develops).
Over the years, researchers developed-and companies brought to market-newer and more effective bronchodilators and other asthma treatments. Novartis continued to manufacture and distribute Brethine with the intention that it be used as a tocolytic. By 2001, nearly half of all prescriptions for terbutaline were for tocolysis, even though the drug was never approved by the FDA for that purpose. In December 2001, Novartis transferred the NDA for Brethine to NeoSan Pharmaceuticals Inc., a wholly owned subsidiary of aaiPharma.
C. The Facts Underlying This Lawsuit
On September 5, 2007, plaintiffs' mother, J.H., was hospitalized because of concerns about premature labor. She was prescribed terbutaline, to be administered every six hours, and was discharged on September 25, 2007. While in the hospital, J.H. received a generic version that was manufactured by Lehigh Valley Technologies, Inc.; after discharge, she received a generic version that was manufactured by Global Pharmaceuticals. J.H. continued taking terbutaline as directed until plaintiffs were born on October 9, 2007. Plaintiffs appeared to be normal until their pediatrician, during a routine checkup in December 2010, reported that the twins may have developmental delays. Despite specialized treatment for both children, a pediatric neurologist diagnosed them with autism in August 2012.
**26
Plaintiffs' first amended complaint alleges that terbutaline passed through the placenta and the blood-brain barrier. As a result, plaintiffs contend, terbutaline caused them to suffer severe and permanent neurologic injuries, including an inability to speak and significant limitations and abnormalities in their motor skills. Plaintiffs further allege that Novartis knew or should have known that terbutaline was of questionable efficacy as a tocolytic agent, that
*161
terbutaline carried serious risks of side effects for newborns whose mothers received the drug during pregnancy, and that federal law required Novartis to report this information to the FDA and to update the warning label-something Novartis could have done unilaterally. (See
To support and place in factual context their negligence cause of action, plaintiffs made a variety of specific allegations regarding Novartis. They alleged that Novartis had a duty to update the label to warn of the drug's effects on fetal development, that Novartis knew or should have known of these effects, that J.H.'s physicians prescribed her terbutaline because of their erroneous belief that terbutaline was safe to use as a tocolytic, that plaintiffs suffered neurological damage as a result of their exposure to terbutaline in utero, and that plaintiffs' injuries were foreseeable. Meanwhile, plaintiffs' negligent misrepresentation cause of action alleged that Novartis falsely represented that terbutaline was safe to use as a tocolytic, that Novartis had no reasonable basis for believing terbutaline was safe to use as a tocolytic, that Novartis intended for pregnant mothers and their physicians to rely on their false representations concerning the drug's safety as a tocolytic agent, that J.H.
*345 and her physicians relied on Novartis's representations, that plaintiffs suffered neurological damage as a result of their exposure to terbutaline in utero, and that plaintiffs' injuries were foreseeable.
Novartis's core assertion in its demurrer was that it had no duty to plaintiffs. To justify its position, the company offered two overarching rationales: It did not manufacture the terbutaline ingested by their mother; and it had transferred the Brethine NDA to another company in December 2001, nearly six years before plaintiffs' mother was prescribed terbutaline. In addition, Novartis argued that plaintiffs had failed to identify with specificity any misrepresentation by Novartis or allege that plaintiffs had relied on any such misrepresentation. In opposition to the demurrer, plaintiffs responded that Novartis had a duty to warn about the drug's effects on fetal development during the period it owned the NDA and manufactured Brethine ; that the six-year gap between Novartis's divestiture of the NDA and plaintiffs' in utero exposure is relevant to causation (and not the existence of the duty); and that the first amended complaint adequately pleaded the misrepresentations with specificity, given that the specific misrepresentations are more likely to be within Novartis's knowledge, and adequately pleaded reliance on those misrepresentations.
*162 The trial court sustained the demurrer without leave to amend. It concluded that Novartis owed plaintiffs no duty as a matter of law relating to claims arising from terbutaline exposure in 2007. Agreeing with Novartis, the court also found that the fraud-based claims suffered from a lack of specificity and that this defect could not be remedied by allegations about Novartis's conduct prior to the 2001 NDA divestiture.
The Court of Appeal reversed and directed that the order sustaining the demurrer be modified to grant plaintiffs leave to amend their causes of action for negligence and negligent misrepresentation. The appellate court reasoned that if plaintiffs could allege that Novartis failed to warn about fetal risks it knew or should have known were associated with terbutaline when used as a maintenance tocolytic prior to its divestiture of the brand-name drug in 2001, that the warning would have remained on the label in 2007 had **27 Novartis added a suitable warning to the label before divestiture in 2001, and that their mother's physician would not have prescribed terbutaline as a maintenance tocolytic had the drug been properly labeled, then their claims for negligence and negligent misrepresentation can survive demurrer.
We granted Novartis's petition for review to decide the existence and scope of warning label liability for brand-name drug manufacturers under California law.
II. DISCUSSION
The sole issue before us is whether the demurrer should have been sustained with respect to the negligence and negligent misrepresentation claims on the ground that Novartis owed no duty of care to plaintiffs. In reviewing an order sustaining a demurrer, we examine the operative complaint de novo to determine whether it alleges facts sufficient to state a cause of action under any legal theory. (
Lee v. Hanley
(2015)
*346 The gist of plaintiffs' warning label liability claim is that Novartis negligently failed to warn about the drug's risk to fetal brain development. They contend that the deficient label foreseeably and proximately caused harm not only to the children of women who were prescribed Brethine, but also to the children of women who were prescribed its generic bioequivalent, which was legally required to-and did-bear the same deficient label. Among other things, plaintiffs rely on section 311 of the Restatement Second *163 of Torts (section 311), which addresses negligent misrepresentation involving physical harm. Under section 311(1), "[o]ne who negligently gives false information to another is subject to liability for physical harm caused by action taken by the other in reasonable reliance upon such information, where such harm results [¶] ... [¶] to such third persons as the actor should expect to be put in peril by the action taken."
Section 311's theory of liability is intended to be "somewhat broader" than that for mere pecuniary loss. ( Rest.2d Torts, § 311, com. a.) It "finds particular application where it is a part of the actor's business or profession to give information upon which the safety of the recipient or a third person depends." (
Id
., § 311, com. b; see also Prosser,
Misrepresentation and Third Persons
(1966)
Duty is indeed the cornerstone of every negligence claim. In California, the general rule governing duty is codified in Civil Code section 1714, subdivision (a) : "Everyone is responsible ... for an injury occasioned
**28
to another by his or her want of ordinary care or skill in the management of his or her property or person ...." Thus, "each person has a duty to use ordinary care and 'is liable for injuries caused by his failure to exercise reasonable care in the circumstances ....' " (
Parsons v. Crown Disposal Co.
(1997)
In the context of prescription drugs, a manufacturer's duty is to warn physicians about the risks known or reasonably known to the manufacturer. (
Carlin v. Superior Court
(1996)
In this case, plaintiffs allege that the terbutaline label failed to warn about the risks to fetal brain development and falsely represented that the drug was safe for use by pregnant women. They further claim that Novartis's control over the Brethine label rendered it responsible for any deficiencies in the terbutaline label, given that generic drug manufacturers are legally obligated to use the label crafted by the brand-name drug manufacturer. Novartis contends that it owed no duty to plaintiffs to update or maintain an accurate label because (1) it did not manufacture the terbutaline that caused plaintiffs' injuries; and (2) it had divested ownership of Brethine, the name-brand drug, several years before plaintiffs' mother was prescribed terbutaline.
To determine whether to create an exception to a brand-name drug manufacturer's duty to warn, we balance the constellation of factors set out in
*165
Rowland
,
supra
, 69 Cal.2d at page 113,
A. Whether Plaintiffs Exposed to the Generic Bioequivalent Drug Can Assert Warning Label Liability Against Novartis, the Brand-name Drug Manufacturer
The first case to recognize warning label liability was
Conte v. Wyeth, Inc.
(2008)
Only a handful of courts have followed
Conte
. (See, e.g.,
Dolin v. SmithKline Beecham Corp.
(N.D.Ill. 2014)
*166 1. Foreseeability and related factors
In determining whether to create an exception to the general statutory duty of care, the "major" (
Cabral
,
supra
, 51 Cal.4th at p. 771, fn. 2,
A brand-name pharmaceutical manufacturer has a duty under federal law to draft, update, and maintain the warning label so that it provides adequate warning of the drug's potentially dangerous effects. (
*349
The duty for a manufacturer of generic drugs, on the other hand, is to ensure that its warning label is identical to the label of the brand-name drug. (
PLIVA
,
supra
, 564 U.S. at p. 613,
What a brand-name manufacturer thus knows to a legal certainty is that any deficiencies in the label for its drug will be perpetuated in the label for its generic bioequivalent. A brand-name manufacturer will
**30
also be aware that although the warnings communicated in its drug label are designed for physicians-and are intended to influence a physician's decision whether to prescribe the drug (see
Stevens v. Parke, Davis & Co.
,
supra
, 9 Cal.3d at pp. 64-65,
Under the second
Rowland
factor, we assess the degree of certainty that the plaintiff suffered injury. This factor, too, strengthens the case for finding a duty of care in these circumstances. Plaintiffs allege that they suffer from global neurological impairment, including autism and pervasive developmental delays. These are indisputably injuries and are compensable under the law. (See
Kesner
,
supra
, 1 Cal.5th at p. 1148,
The third Rowland factor implicates the closeness of the connection between the defendant's conduct and the plaintiff's injury. The label for a generic drug is (and must be) the same as the label for the brand-name drug, so any deficiency in the brand-name label will be reflected in the generic label. Plaintiffs allege that the deficient Brethine label led their mother's physician to prescribe terbutaline, which caused their neurological injuries. This scenario describes a close connection between Novartis's allegedly negligent conduct and plaintiffs' injuries.
Novartis, meanwhile, relies on
O'Neil v. Crane Co.
(2012)
Here, by contrast, federal regulations granted the brand-name drug manufacturer-and no other manufacturer-control over the active ingredients in the generic drug and the content of the warnings included in the generic's label. 2 In addition, the temporal connection between Novartis's allegedly negligent conduct, on the one hand, and plaintiffs' exposure to harm and subsequent injury, on the other, is much closer than was the case in O'Neil .
2. Considerations of public policy
Foreseeability alone, however, is not sufficient to justify a duty of care in every instance. (
Erlich v. Menezes
(1999)
Time and again we have recognized how " '[t]he overall policy of preventing future harm is ordinarily served, in tort law, by imposing the costs of negligent conduct upon those responsible.' " (
*351
Kesner
,
supra
, 1 Cal.5th at p. 1150,
If the policy of preventing harm has special relevance to any particular endeavor, surely prescription drug labeling is one. (
Sindell v. Abbott Laboratories
,
supra
, 26 Cal.3d at p. 611,
The brand-name drug manufacturer is the only entity with the unilateral ability to strengthen the warning label. So a duty of care on behalf of all those who consume the brand-name drug or its bioequivalent ensures that the brand-name manufacturer has sufficient incentive to prevent a known or reasonably knowable harm. In
O'Neil
, by contrast, we found "no reason" to believe that the defendant valve and pump manufacturers would have any control over the safety of other companies' replacement parts or companion products (or even the Navy's purchasing choices or specifications). (
O'Neil
,
supra
, 53 Cal.4th at p. 365,
Against the public interest in preventing harm, we must balance the defendant's burden and the consequences to the community of imposing a duty of care. The burden that matters, though, is not the cost of compensating individuals for injuries that the defendant has actually and foreseeably caused. As we recently explained in
Kesner
, "shielding tortfeasors from the full magnitude of their liability for past wrongs is not a proper consideration in determining the existence of a duty. Rather, our duty analysis is forward-looking, and the most relevant burden is the cost to the defendants of upholding, not violating, the duty of ordinary care." (
Kesner
,
supra
, 1 Cal.5th at p. 1152,
Strictly speaking, then, the burden on brand-name drug manufacturers of satisfying a common law duty of care to those who are prescribed the generic version of the drug is zero. Brand-name manufacturers
already
have a continuing duty to warn of potential risks "as soon as there is reasonable evidence of an association of a serious hazard with a drug; a causal relationship need not have been proved." (
Novartis complains that unless the ordinary duty of care is narrowed, the brand-name drug manufacturer would end up an insurer for the entire market. This would occur, Novartis contends, even though the brand-name manufacturer may hold only a small fraction of the combined sales of the brand-name drug and its generic bioequivalent. We disagree. A brand-name drug manufacturer would not be liable where, for example, the injury arose from a defect in the manufacturing process of the generic drug (see, e.g.,
Fisher v. Pelstring
(D.S.C. 2012)
Novartis cautions that warning label liability could perversely incentivize a brand-name manufacturer to withdraw its drug from the market, rather than expose itself to the risk of suit by those who-in reliance on the brand-name manufacturer's label-were prescribed the generic bioequivalent and suffered injury. Yet Novartis fails to explain why brand-name manufacturers would find it economically advantageous to withdraw drugs from the market rather than simply modify the warning labels to include the newly discovered risks. Nor does it offer any evidence that brand-name manufacturers have accelerated their withdrawal from the market in the nine years since
Conte
was decided. Moreover, a brand-name drug manufacturer cannot avoid its duty to
*172
update and maintain its warning label simply by unilaterally exiting the market. Under FDA regulations, a brand-name drug manufacturer's duty to update and maintain the warning label continues, even if the brand-name drug has been withdrawn from the market, until the FDA (having assured itself that the drug is safe, effective, and correctly labeled) withdraws approval of the NDA. (
Novartis complains next that it is unfair to subject a brand-name drug manufacturer to liability for harm caused by a competitor's product-a product from which the brand-name manufacturer derives no revenues or profit. But the plaintiffs' claim here is not that terbutaline is defectively designed or inherently dangerous. It is that terbutaline's warning label failed to mention the risk to fetal brain development, and that Novartis was responsible for the deficient label. So the
**34
alleged fault here lies with Novartis, not with its generic competitors. The brand-name drug manufacturer's
*354
burden to adequately label its drug as a means of ensuring adequate warnings for the generic bioequivalent is more than offset by the substantial benefits federal law confers on the brand-name manufacturer: a monopoly over the market for the life of the patent, which can be extended for the time consumed by FDA review of the NDA (see
We are equally unpersuaded by Novartis's contention that warning label liability would stifle innovation by substantially raising drug costs and chilling the development and marketing of new drugs. The logic buttressing this argument is far from self-evident. Warnings about a product's efficacy or danger may indeed risk diminishing its value to the manufacturer. Less obvious is the manufacturer's response to this predicament. One might just as easily assert that a drug company, after adding a new warning, will be incentivized to develop new and safer alternatives to the drug so that it can recapture the market for treatment of that disease. (See
Carlin
,
supra
, 13 Cal.4th at p. 1117,
Indeed, the pharmaceutical industry raised a similar objection in
Carlin
to the imposition of strict liability for the failure to warn about the known or reasonably scientifically knowable dangers of a drug. We found "no clear or sufficient basis for concluding that research and development will inevitably decrease" as a consequence of imposing liability for failure to warn of known or knowable risks (
Carlin
,
supra
, 13 Cal.4th at p. 1117,
The same is true here. When it comes to choosing whether the cost of an injury involving prescription medication should be borne by an innocent plaintiff or a negligent defendant, our case law has routinely held that the latter should bear the cost. (
*355
Sindell v. Abbott Laboratories
,
supra
, 26 Cal.3d at pp. 610-611,
Although we declared in
O'Neil
that "little moral blame can attach to a failure to warn about dangerous aspects of
other
manufacturers' products and replacement parts" (
O'Neil
,
supra
, 53 Cal.4th at p. 365,
Finally, Novartis offers no reason why a brand-name drug manufacturer would be unable to insure against the risk of warning label liability. Presumably, a brand-name manufacturer already insures against the risk of liability arising from a deficient warning label when a drug is introduced and the manufacturer has a monopoly over that market. It is far from clear why the brand-name drug manufacturer's exposure would become fatally uncertain merely because the brand-name manufacturer is sharing the market with generic manufacturers. (Cf.
*356
O'Neil
,
supra
, 53 Cal.4th at p. 365,
*175 3. Out-of-state authorities
Novartis (and its amici curiae) rely in substantial part on what they call the "overwhelming" majority of courts that have declined to recognize warning label liability owed to those who were prescribed a generic version of the drug in reliance on the brand-name drug label. Although the decisions of our sister states and the lower federal courts may be instructive to the extent we find their analysis persuasive, they are neither binding nor controlling on matters of state law. (
People v. Gonzales and Soliz
(2011)
The " 'leading case' " (
Strayhorn v. Wyeth Pharms., Inc.
(6th Cir. 2013)
Foster reasoned first that the negligent misrepresentation cause of action was in essence a claim of product liability, but "without meeting the requirements [Maryland] law imposes in products liability actions"-i.e., "that the defendant manufactured the product at issue." ( Foster , supra , 29 F.3d at p. 168.) The court next addressed the peculiarities of the regulated pharmaceutical market, under which "any representations [the defendant] makes when advertising Phenergan also apply to generic promethazine"; a warning "will simply not be made" if the brand-name manufacturer does not issue one; and a patient who is prescribed Phenergan"may actually receive *176 generic promethazine." ( Id . at p. 169.) In rejecting liability nonetheless, the Foster court assumed that although generic manufacturers "must include the same labeling information as the equivalent name brand drug, they are also permitted to add or strengthen warnings and delete misleading statements on labels, even without prior FDA approval. ... Manufacturers of generic drugs, like all other manufacturers, are responsible for the representations they make regarding their products." ( Id . at p. 170.) The *357 court also concluded that "to impose a duty in the circumstances of this case would be to stretch the concept of foreseeability too far" under Maryland law, which had recognized the tort of negligent misrepresentation only where " 'one party has the right to rely for information upon the other, and the other giving the information owes a duty to give it with care.' " ( Id . at p. 171.) In the court's view, no such relationship could ever exist because Brandy "was injured by a product that [the defendant] did not manufacture." ( Ibid . )
At the core of the
Foster
court's analysis is an erroneous assumption: that generic drug manufacturers may "add or strengthen warnings and delete misleading statements on labels, even without prior FDA approval," and that they can be sued for their failure to do so. (
Foster
,
supra
, 29 F.3d at p. 170.) In reality, generic drug manufacturers are legally obligated to conform their drug label to the brand-name manufacturer's label (
PLIVA
,
supra
, 564 U.S. at p. 613,
**37
*177
Even on its own terms, though,
Foster
's reasoning proves unhelpful in construing California law, and finds no support in it. First, California law does not conflate negligent misrepresentation and strict liability in the manner
Foster
believed was true of Maryland law. (
Conte
,
supra
, 168 Cal.App.4th at p. 108,
Second, California law places greater weight on the element of foreseeability in the duty analysis than does Maryland law. Indeed, this state
*178
treats foreseeability as "[t]he most important factor" (
Kesner
,
supra
, 1 Cal.5th at p. 1145,
Third, in one crucial respect,
Foster
is like the vast majority of the out-of-state cases on which Novartis relies: it arose in federal court under diversity jurisdiction. Federal courts sitting in diversity are "extremely cautious" about recognizing innovative theories under state law (
Combs v. Int'l Ins. Co.
(6th Cir. 2004)
We likewise discount decisions from those jurisdictions that differ from California by
**39
categorically excluding from liability certain defendants (see, e.g.,
*360
Huck v. Wyeth, Inc.
,
supra
, 850 N.W.2d at p. 371 (plur. opn. of Waterman, J.) ["the tort of negligent misrepresentation does not apply to sellers of products but rather is limited to those in the business or profession of supplying information for the guidance of others"] ) or certain injuries (see, e.g.,
Flynn v. American Home Products Corp.
(Minn.Ct.App. 2001)
The negligence causes of action are potentially viable because of the allegedly deficient representations in Novartis's
*361
warning label. Novartis is not being sued for dangers inherent in the generic terbutaline manufactured by some other entity. Nor do plaintiffs claim that any product manufactured by Novartis caused them harm. They claim instead that allegedly deficient representations and omissions in Novartis's warning label caused them harm. The fact that Novartis
also
manufactured a product is extrinsic to the analysis and does not insulate it from liability for its alleged misrepresentations. (See
Conte
,
supra
, 168 Cal.App.4th at pp. 109-110,
We have determined that Novartis owed a duty of care to those who were prescribed Brethine or its generic bioequivalent in reliance on Novartis's warning label. Novartis claims that the demurrer should nonetheless be sustained without leave to amend on the ground that it sold the Brethine NDA to aaiPharma in 2001 and no longer had control over its warning label in 2007, when plaintiffs' mother was prescribed terbutaline. So we now consider whether Novartis should be relieved of all liability for its allegedly negligent failure to update the label as a matter of law, despite the fact that aaiPharma continued using the label Novartis had crafted.
Plaintiffs fault Novartis. But they do not claim the company was responsible for any negligent acts or omissions after the transfer of ownership. After all, under FDA regulations, only the current NDA holder has the authority to unilaterally modify the drug's warning label. Plaintiffs claim instead that they were harmed by Novartis's failure to update the label prior to transferring the NDA to aaiPharma, in that aaiPharma continued to use the same warning label until at least 2007, when their mother was prescribed terbutaline. In effect, plaintiffs claim that the Brethine warning label was deficient at the time Novartis transferred the NDA-and it was reasonably foreseeable that it would remain deficient, given the incentives facing any successor manufacturer.
To address this aspect of plaintiffs' claim, we must determine whether to recognize an exception to a brand-name manufacturer's duty to warn. Is a brand-name drug manufacturer's duty to warn extinguished simply because the deficiency in the label caused the injured plaintiff to be exposed to the drug
after
the manufacturer had transferred the NDA to a successor? Foreseeablity of harm is the touchstone of our duty analysis. (
Kesner
,
supra
, 1 Cal.5th at p. 1148,
*362 1. Foreseeability and related factors
We explained above why it was foreseeable that Novartis's failure to update the Brethine warning label could affect fetuses exposed to the generic *182 version of the drug in utero. And there is no dispute that plaintiffs have alleged injury. Although Novartis was no longer responsible for updating the warning label at the time plaintiffs' mother was prescribed the drug, aaiPharma was using the same label it had inherited from Novartis. According to plaintiffs, neither NDA holder had sufficient financial incentive to update the label: Nearly half of all prescriptions for Brethine or its generic equivalent were to slow preterm labor. Under the circumstances, it was certainly foreseeable that aaiPharma would be no more conscientious about updating the warning label than Novartis allegedly had been.
Novartis contends the connection between its alleged negligence and plaintiffs' injury was extremely remote, as it had ceased producing the drug six years before the injury. But it is not clear why liability should turn on Novartis's role in the manufacturing process. What warning label liability stems from is Novartis's failure to warn about a drug's risks, not its production of a defective drug. The complaint alleges that Novartis and aaiPharma were concurrent tortfeasors whose liability stemmed from failure to warn, **41 because each negligently failed to update the warning label.
We agree that Novartis's failure to update the warning label could foreseeably cause harm to plaintiffs. Under the circumstances arising from the federal regulatory regime for prescription drugs, a successor manufacturer's negligent failure to update the warning label is foreseeable. According to federal regulatory rules, a successor brand-name drug manufacturer has no choice but to use the former manufacturer's warning label-or a warning label at least as strong as the one used by the previous brand-name manufacturer-unless directed otherwise by the FDA. (
Wyeth
,
supra
, 555 U.S. at p. 568,
Novartis highlights the six years that elapsed between its surrender of the NDA for the drug at issue in this case and the decision to prescribe terbutaline to plaintiffs' mother. Yet the gap between the transfer of this particular NDA and the time at which plaintiffs' mother was prescribed terbutaline does not bear on the question of duty, "which must be addressed at a higher level of generality." (
Kesner
,
supra
, 1 Cal.5th at p. 1158,
It is true enough that a successor drug manufacturer has an obligation, under state as well as federal law, to ensure adequacy of the warning label. But the scenario at issue here implicates whether a successor drug manufacturer is sufficiently likely-as a matter of law-to modify the warning label when the brand-name manufacturer, which labored under an identical obligation, negligently failed to do so. In such circumstances, it is at least plausible that a successor manufacturer may choose to undertake only a cursory investigation of the medical literature, on the assumption that the prior manufacturer must **42 have done a more thorough inquiry during the period that it was responsible for maintaining the warning label. This option will seem especially attractive when the prior manufacturer has greater resources or expertise than its successor. A successor manufacturer may also undertake an adequate inquiry but make no changes to the label in close cases, partially or *184 entirely trusting the judgment of the prior manufacturer. Or a successor manufacturer, like the prior manufacturer, may fear an adequate warning would damage the market share for the drug and balance its lost revenue and potential exposure in the same way as the prior manufacturer. Indeed, plaintiffs claim that neither NDA holder wanted to jeopardize Brethine's use as a tocolytic, which accounted for almost half of the drug's market share. Any or all of these factors could explain why a drug's warning label may prove "stickier" than what is optimal to protect public safety at a reasonable cost, and why a successor drug manufacturer would not be categorically distinguishable from the prior manufacturer in its likelihood of being conscientious about its obligations to disclose relevant risks.
Under the "general" rule, " ' "every person has a right to presume that every other person will perform his duty and obey the law." ' " (
Webb v. Special Electric Co., Inc.
,
supra
, 63 Cal.4th at p. 191,
*364
But we have never allowed a defendant to excuse its own negligence as a matter of law simply by asserting that
someone else
should have picked up the slack and discharged the duty at issue. (See
Stewart v. Cox
(1961)
According to Novartis, the policy of preventing future harm would not be advanced by subjecting the brand-name drug manufacturer to liability after it has already divested itself of the drug and no longer has control over the warning label. But in examining the prevention of future harm, we undertake the duty analysis "look[ing] to the time when the duty was assertedly owed." (
Kesner
,
supra
, 1 Cal.5th at p. 1150,
The more substantial danger is that neither manufacturer will have sufficient incentive to update the label. Unless there is warning label liability, it will be economically rational in some circumstances for a brand-name manufacturer to offload the drug to a successor rather than update the warning label. And if the brand-name manufacturer fails to update the label to disclose a known or knowable risk, economic interests and simple inertia may lead the successor manufacturer to the same strategy. (See ante , at 226 Cal.Rptr.3d at pp. 363-364, 407 P.3d at pp. 42-43.) The better rule is to provide appropriate incentives for the brand-name manufacturer to update the warning label at the earliest possible time, given that the successor manufacturer cannot remove any aspect of the warning without FDA approval.
To determine how best to incentivize a drug manufacturer to provide prompt
*366
warnings, we turn to the very factors on which Novartis trains its attention: the extent of the duty's burden on the defendant and the
**44
consequences to the community. Novartis complains first that plaintiffs' proposed rule would lead to immeasurable and perpetual liability for brand-name drug manufacturers. This appears to be an overstatement. Only during the time it holds the NDA does the brand-name drug manufacturer have a duty of care. Although a breach of that duty can have enduring effects-effects that do not magically disappear merely because the brand-name manufacturer no longer holds the NDA-a plaintiff would still need to prove that the injury was foreseeable at the time the brand-name manufacturer held the NDA, that the brand-name manufacturer's deficient label proximately caused the injury, and that the prescribing physician relied on the brand-name manufacturer's misrepresentations or omissions. The passage of time would naturally tend to undermine a plaintiff's ability to prove that an injury was foreseeable at that
*187
earlier stage,
8
that the physician actually relied on the defendant's warning label, or that the defendant's negligence proximately caused injury. (See
Beacon
,
supra
, 59 Cal.4th at p. 583,
Yet the question before us involves neither causation nor these other elements of negligence. What role the six-year gap between Novartis's transfer of the NDA to aaiPharma and plaintiffs' exposure to terbutaline might play in plaintiffs' ability to prove these remaining elements is beyond the scope of this proceeding. We granted review to decide only the threshold question of a brand-name drug manufacturer's duty of care and therefore have no occasion to address other arguments Novartis might advance to defeat liability. (See
Kesner
,
supra
, 1 Cal.5th at p. 1157.)
9
But we reject
*367
Novartis's contention that a finding of duty will result in perpetual liability for brand-name drug manufacturers as well as the burden of a trial to address every claim of harm. Time's effect on causation, while ordinarily a question of fact, becomes a question of law " 'where the facts are such that the only reasonable conclusion is an absence of causation.' " (
State Dept. of State Hospitals v. Superior Court
(2015)
Moreover, the greater the gap between transfer of the NDA and the plaintiffs' exposure to the drug, the greater the likelihood that the NDA would have been transferred to yet another manufacturer, which would multiply the number of potential defendants available to share responsibility for damages. Because a defendant's liability for noneconomic damages is not joint but several ( Civ. Code, § 1431.2, subd. (a) ), a negligent brand-name manufacturer would be liable for noneconomic damages only in an amount that was directly proportional to its percentage of fault. ( Ibid .)
Indeed, a brand-name manufacturer could entirely avoid the prospect of extended exposure by including an indemnification provision when it transferred ownership of the NDA. (See, e.g.,
Conte
,
supra
, 168 Cal.App.4th at p. 95, fn. 1,
A somewhat analogous situation lay at the heart of a case the court addressed recently. In
*189
Centinela Freeman Emergency Medical Associates v. Health Net of California, Inc.
(2016)
The defendant health care service plans made an argument that echoed what Novartis argues here: that they had lawfully transferred
**46
their legal responsibilities to another entity and had therefore terminated any duty of care. We unanimously rejected the argument that the delegation of financial responsibility to an IPA necessarily relieved the health care service plans of any obligation to pay for its enrollees' covered emergency care. (
Centinela Freeman
,
supra
, 1 Cal.5th at p. 1001-1002,
Centinela Freeman
tends to undermine Novartis's absolutist position that a lawful transfer of its duty to another entity necessarily terminated its liability for its own negligence. Under our ruling in
Centinela Freeman
, a health care service plan remains responsible for the costs of its enrollees' emergency medical care, despite a lawful delegation of that financial responsibility, if the plan knows or should know the IPA would be unable to fulfill that financial responsibility. (
Centinela Freeman
,
supra
, 1 Cal.5th at pp. 1001-1002,
Novartis renews its claim that warning label liability would severely chill both the innovation and marketing of new drugs if imposed after the brand-name manufacturer exits the market. Yet once again, it offers neither evidence nor a persuasive rationale to support its contention-and no reason for us to prefer some unknown increment of drug development over the urgent need to compensate a victim whose injury was foreseeably and proximately caused by a brand-name manufacturer's negligence. (See
Carlin
,
supra
, 13 Cal.4th at p. 1117,
We explained earlier why significant moral blame attaches to the failure to warn about a drug's risks when the brand-name drug manufacturer
**47
knew or should have known about those risks. The fact that the brand-name manufacturer has since exited the market does not alter the calculus. Under plaintiffs' theory, the actionable conduct occurred while the manufacturer still had control over the warning label. Had Novartis updated the warning label before surrendering the NDA, the federal regulations make it very likely that the warning would have remained on the label in 2007. (See
Wyeth
,
supra
, 555 U.S. at p. 568,
The concurring and dissenting opinion makes much of the fact that no other jurisdiction has yet recognized a brand-name drug manufacturer's duty to maintain a warning label in these circumstances. The legal landscape was just as bare when the Court of Appeal recognized a brand-name drug manufacturer's duty to consumers of the generic bioequivalent drug (see
Conte
,
supra
, 168 Cal.App.4th at p. 101,
III. CONCLUSION
We do not doubt the wisdom of crowds in some settings. But the value of an idea conveyed by or through a crowd depends not on how loudly it is proclaimed or how often it is repeated, but on its underlying merit relative to *192 the specific issue at hand. Despite the impressive case authority Novartis has collected on its behalf, none of it purports to interpret California law. Yet it is California law that we must construe and apply in this case.
In doing so, we find that brand-name drug manufacturers have a duty to use ordinary care in warning about the safety risks of their drugs, regardless of whether the injured party (in reliance on the brand-name manufacturer's warning) was dispensed the brand-name or generic version of the drug. We also conclude that a brand-name manufacturer's sale of the rights to a drug does not, as a matter of law, terminate its liability for injuries foreseeably and proximately **48 caused by deficiencies present in the warning label prior to the sale. We therefore affirm the Court of Appeal.
We Concur:
Chin, J.
*371 Liu, J.
Mauro, J. *
Corrigan, J.
I accept the majority's holding that a brand-name drug manufacturer's duty to warn extends to consumers of a generic bioequivalent, but only because federal regulations currently require that generic drugs carry the same warning label as appears on the brand-name product. (See
*193 The majority's second holding, however, would extend indefinitely a drug manufacturer's duty to warn the customers of its successor, even after sale of the product line. No special feature of FDA law or practice warrants this rule. Plaintiffs' theory of "predecessor liability" represents a substantial and unprecedented expansion of tort duties. The majority cites no case holding a predecessor manufacturer liable for failing to warn about injuries caused by its successor's product. Indeed, it appears that predecessor liability for failure to warn has never before been recognized by any court, in any jurisdiction. To the extent the theory has been raised, courts across the country have universally rejected it.
For example, in the silicone breast implant litigation, some plaintiffs whose implants were manufactured by McGhan Medical Corporation (McGhan) sought to sue the Minnesota Mining and Manufacturing Company (3M) on the theory that 3M was the original designer. (See, e.g.,
In re Minnesota Breast Implant Litigation
(D.Minn. 1998)
The majority insists pharmaceutical drugs are different from other products. However, both courts that have considered the issue have refused to extend tort liability to drug manufacturers after transfer of the product's New Drug Application (NDA) to a successor company. In
In re Darvocet, Darvon and Propoxyphene Products Liability Litigation
(E.D.Ky. Mar. 7, 2012, MDL Docket No. 2226)
O'Neil
,
supra
, 53 Cal.4th at pages 363-366,
The majority attempts to distinguish
O'Neil
as applied to a brand-name manufacturer's liability for generic bioequivalents. (See maj. opn.
ante
, at 226 Cal.Rptr.3d at pp. 349-350, 407 P.3d at pp. 30-31.) The distinction does not hold when applied to predecessor liability. In the generic drug context, public policy supports imposing a duty of care on brand-name manufacturers because the brand-name manufacturer is the
only
party with the practical and legal ability to
*373
warn about product risks. Generic manufacturers cannot write their own labels. Their products are legally required to carry the same warnings as appear on their brand-name counterparts. (
When a drug manufacturer acquires a new product line, it assumes the responsibility to update the warning label if and when reasonable evidence demonstrates a link to a serious health hazard. (
Moreover, the accumulation of scientific studies will often make the correlation with a health risk more clear over time. The majority's analysis elides this important feature of pharmaceutical practice. Adverse effects from a drug typically appear first in anecdotal case reports. It can take several years for epidemiological studies, the gold standard for establishing causation, to be conducted and published. Indeed, a 2013 FDA study found that the "most critical safety-related label changes, boxed warnings and contraindications, occurred a median 10 and 13 years after drug approval (and the range spanned from 2 to 63 years after approval)." (Lester et al., Evaluation of FDA Safety-related Drug Label Changes in 2010 (2013) 22 Pharmacoepidemiology & Drug Safety 302, 304.) A connection to adverse effects that appears reasonably clear when a successor produces a drug may well have been more tenuous, perhaps not even rising to the *374 FDA's "reasonable evidence" standard, when it was the predecessor's product. Scientific evidence may not demonstrate the link to a health risk until after divestiture. Yet, at that point there is little to nothing a predecessor manufacturer can do to warn about the harm.
This case demonstrates the point. The majority concedes that "approximately half" of the studies plaintiffs cite to show terbutaline's impact on fetal *196 brain development postdated Novartis's sale of the product line. (Maj. opn. ante , at 226 Cal.Rptr.3d at p. 366, fn. 8, 407 P.3d at p. 44, fn. 8.) But this summary does not tell the full tale. A few rat studies in the 1980s showed that terbutaline could cross the placenta and affect fetal brain development, and effects in human children were beginning to be documented. But most of the early studies were focused on the drug's effectiveness, or lack thereof, at preventing preterm labor. The first long-term study cited in plaintiffs' complaint that demonstrates a potential impact on human development was published in 2001, the same year Novartis sold the Brethine NDA to aaiPharma. It is undisputed that after 2001 Novartis had no ability to change the drug's warning label. Moreover, the scientific link between terbutaline and autism remains questionable. In 2011, four years after plaintiffs' mother was given terbutaline and nearly 10 years after Novartis's divestiture, the FDA reviewed the scientific literature investigating this link and concluded the studies did not constitute " 'positive evidence' " of a risk to fetal health. (Food & Drug Admin., letter to James P. Reichmann responding to citizen petition, Feb. 17, 2011, Docket No. FDA-2008-P-0358, p. 12.)
I discuss these developments not to express a view on the merits of plaintiffs' suit, but simply to point out that the scientific investigation of an alleged harmful effect takes time. Anecdotal case reports, in vivo
**51
studies, or animal studies that initially suggest an association are sometimes discredited by later epidemiological studies, which are more authoritative but take longer to conduct.
3
Yet the majority's new duty rule makes it nearly imperative for manufacturers to issue warnings in advance of the science if they are selling a drug's NDA. In the normal course, a responsible drug manufacturer can monitor scientific developments and work with the FDA to determine when additional warnings are warranted. It loses this ability after transferring the NDA to another company. By holding that such a manufacturer owes a duty to warn its successor's customers even many years later, the majority creates an incentive for manufacturers to warn about every conceivable harm before transferring an NDA, lest their successors fail to include appropriate warnings when a risk is later validated.
*197
It is certainly possible to foresee that a successor manufacturer will shirk its legal obligation to warn. That a harm is foreseeable does not necessarily mean we should
*375
recognize a duty of care, however. On "clear judicial days ... a court can foresee forever." (
Thing v. La Chusa
(1989)
First, as noted, a predecessor manufacturer has no control over the successor's product warnings. Only the current NDA holder has the power to change a drug's warning label. (
Second, the majority's holding will likely encourage over-warning by drug manufacturers. Drug manufacturers are already under a duty to update their warning labels, and they already face the risk of suit from their own customers if they fail to comply with that duty. The knowledge that they will still be subject to liability years in the future, even after divesting a product line, might well cause companies to seek the FDA's permission to add warnings about potential adverse effects that have only the barest support in evolving scientific literature. We have noted
**52
before that overabundant product warnings breed consumer disregard. (See, e.g.,
O'Neil
,
supra
, 53 Cal.4th at p. 365,
Third, the majority's rule could conceivably have the perverse effect of diminishing successor corporations' incentive to update labels as scientific evidence develops. Current product manufacturers already have a disincentive to add warnings that may lower their profits. By holding that predecessor companies must potentially share liability for injuries caused by a successor's product, the court effectively reduces successor companies' exposure to tort liability. Aware that the cost of tort suits can be shared with their predecessors, some successor companies may decide to delay or perhaps even forgo additional warnings.
Fourth, and perhaps most troubling, creating a broad duty of care to consumers
*376
of a successor's product will expose pharmaceutical companies to liability in perpetuity. There is no logical stopping point for such a duty. The majority asserts that injuries will eventually become too remote for proximate causation to be established. It, however, declines to predict when that time might be reached and ventures no opinion about whether the
six-year
gap in this case is long enough. (But cf.
Lyman v. Pfizer, Inc.
,
supra
,
Absent some such limiting principle, a proximate cause inquiry cannot reliably prevent excessive liability because proximate cause is ordinarily a question of fact for the jury. (
Lacy v. Pacific Gas & Electric Co.
(1934)
Exposing drug manufacturers to broad liability with no predictable end point has the clear potential to destabilize the pharmaceutical industry and chill innovation. Although the majority contends no evidence has been presented to support this prediction (maj. opn.
ante
, at 226 Cal.Rptr.3d at p. 369, 407 P.3d at p. 46), we do not typically demand evidence on the
Rowland
factors. The
Rowland
analysis is inherently predictive, not evidence-based. (Cf.
Cabral
,
supra
, 51 Cal.4th at p. 772,
Fifth, there is no reason to think the majority's predecessor liability holding will be limited to the pharmaceutical industry, or even to immediate predecessors. Despite its suggestion that pharmaceutical drugs are somehow different, the majority opinion identifies no specific feature of drug regulation that makes an extension of duty especially desirable or necessary in this context. What is now to stop users of any product from suing a former manufacturer, arguing it was foreseeable the successor would fail to update the product's warnings? The path of least resistance for all successor companies is to continue the product warnings used by their predecessors. It will generally be against a successor's financial interest to add warnings, no matter what the product. The majority's rule thus opens the door to predecessor liability for all products. It is also unclear exactly how far back this liability would extend. Some amicus briefs advised that product line acquisitions are common in the pharmaceutical industry. If a product line has changed hands two or three times, do all of these manufacturers have a duty of care toward the eventual plaintiff? Apparently they do, given the majority's remarks on joint and several liability. (See maj. opn. ante , at 226 Cal.Rptr.3d at p. 367, 407 P.3d at p. 44.) Again, however, this reasoning suggests no logical stopping point.
*200 Sixth, it is not clear that an expanded duty of care is needed to prevent drug manufacturers from concealing risks when their product line is acquired. These transactions involve highly sophisticated parties, and acquiring companies can be expected to discover a drug's known risks when conducting due diligence. Plaintiffs here never alleged that Novartis hid risks from aaiPharma. Moreover, the appropriate remedy for this wrong is not an overbroad duty rule, but a fraud or breach of contract lawsuit from the acquiring company. A lawsuit related to the acquisition offers a more immediate and effective deterrent than the prospect of future tort claims by the acquirer's customers. Such an approach would make clear the duties of full disclosure and due diligence. It would also encourage successor companies to remain attentive to the evolving science relating to their acquisitions.
In discussing
Centinela Freeman Emergency Medical Associates v. Health Net of California, Inc.
(2016)
Seventh, with respect to moral blame, the majority focuses too narrowly on the facts of this specific case. The broader question is what moral blame attaches to a manufacturer's failure to warn its successor's customers about a product defect. A predecessor manufacturer's share of moral blame may well be lower than that of a successor company that fails to update warnings, especially if scientific knowledge has advanced over time to provide stronger evidence of the product's link to an adverse effect.
Eighth, and finally, despite the majority's blithe assurance that drug manufacturers can "entirely avoid" perpetual liability through insurance or indemnity agreements (maj. opn. ante , at 226 Cal.Rptr.3d at p. 367, 407 P.3d at p. 44), there is no precedent for coverage against claims arising from another company 's product. None of the cases cited in the majority opinion addressed this unusual scenario. While insurance might be available in theory, the policy would have to cover a potentially enormous future risk that the insured would have no ability to mitigate. At the very least, the coverage would be difficult to manage and extremely costly. Defining the covered events could also be difficult, given that the potential plaintiffs would have no relationship with the insured. Even if appropriate insurance does become available, the majority's holding will require that pharmaceutical companies maintain it on all drugs for several years after they have stopped selling the products and realizing a profit. The high cost of insuring against the majority's extension of the liability will almost certainly drive up the prices for prescription drugs. 5
*379
*202
For all these reasons, although I join the majority's decision to affirm
Conte v. Wyeth, Inc.
(2008)
Cantil-Sakauye, C. J.
Kruger, J.
Physicians may, in their professional judgment, prescribe a drug for a purpose other than that for which it has been approved by the FDA. (
Buckman Co. v. Plaintiffs' Leg. Com.
(2001)
The FDA has been considering for some time a rule that would effectively abrogate PLIVA and enable generic drug manufacturers to update a drug's warning label unilaterally, even if the brand-name manufacturer had not yet done so. (See FDA, Supplemental Applications Proposing Labeling Changes for Approved Drugs and Biological Products, 78 Fed.Reg. 67985 (Nov. 13, 2013) ; see Dept. of Health and Human Services, Regulatory Agenda, 82 Fed.Reg. 40277, 40279 (Aug. 24, 2017).) If adopted, the new rule "would create parity between NDA holders and ANDA holders with respect to submission of CBE-0 supplements for safety-related labeling changes based on newly acquired information" (78 Fed.Reg., supra , at p. 67989 ) and may conceivably justify reweighing of the Rowland factors and some reconsideration of the brand-name manufacturer's duty in this category of cases.
There is a sad coda to
Foster
. The Fourth Circuit's ruling relieved the brand-name manufacturer of any duty to warn of known or knowable risks of the drug when a plaintiff had been given the generic equivalent-and (contrary to
Foster
's key assumption) the generic manufacturer had no ability to deviate from the brand-name manufacturer's label. As a result, it took until 2000-six years after
Foster
was decided-for the FDA to modify the warning to recommend that promethazine not be given to children younger than two years old. (Starke et al.,
Boxed Warning Added to Promethazine Labeling for Pediatric Use
(2005)
Novartis suggests that we recently conflated strict liability and negligence in
Webb v. Special Electric Co., Inc.
(2016)
O'Neil
did not erase the distinction between strict liability and negligence, either. Far from conflating the two theories, we said simply that "[t]he same policy considerations that militate against imposing strict liability
in this situation
apply with equal force in the context of negligence." (
O'Neil
,
supra
, 53 Cal.4th at p. 366,
We therefore do not find persuasive those out-of-state cases discounting the role of foreseeability (see, e.g.,
Huck v. Wyeth, Inc.
,
supra
, 850 N.W.2d at p. 376 (plur. opn. of Waterman, J.) [" 'foreseeability should not enter into the duty calculus' "] ) or requiring the existence of a specific type of relationship between the plaintiff and the defendant (see, e.g.,
Moretti v. Wyeth, Inc.
(9th Cir. 2014)
Nor can the concurring and dissenting opinion derive any support from the scattering of federal district court cases involving a challenge to the adequacy of a medical device label. Federal law preempts state tort actions based on deficient warnings for medical devices. (
Riegel v. Medtronic, Inc.
(2008)
This case does not present the question, and we do not decide, whether a brand-name manufacturer would remain liable for deficiencies in its warning label when the FDA has formally withdrawn its approval of the NDA and has determined "that the drug was voluntarily withdrawn from sale for reasons other than effectiveness or safety." (Lasker,
supra
, 82 Def. Counsel J. at p. 306 ; see
Indeed, approximately half of the studies cited in the first amended complaint to demonstrate a link between terbutaline exposure in pregnancy and fetal brain development postdated Novartis's sale to aaiPharma. To avoid the distortion caused by hindsight bias, trial courts should be careful to protect the jury from needlessly being exposed to or considering scientific studies connecting a drug to some harm where those studies postdate transfer of the NDA.
Recognizing a brand-name drug manufacturer's duty of care in these circumstances does not prevent the manufacturer from arguing in a given case that it did not breach its duty given the scientific knowledge at the time; that its label could not have proximately caused the harm given the passage of time between the transfer of the NDA and the plaintiff's exposure to the drug, as well as the successor's exclusive responsibility for promoting the assertedly dangerous off-label use of the drug (see
Lyman v. Pfizer, Inc.
(D.Vt. July 20, 2012, No. 2:09-CV-262),
The concurring and dissenting opinion finds "perhaps most troubling" the court's unwillingness to "predict" when the gap between transfer of the NDA and exposure to the drug will be so remote as to preclude a finding of proximate cause. (Conc. & dis. opn., post , at 226 Cal.Rptr.3d.at p. 375, 407 P.3d at p. 52.) But neither party has briefed the issue of proximate cause, nor is proximate cause fairly encompassed within the issue presented-indeed, the issue presented involves exclusively the tort law element of duty . Novartis remains free to contest the existence of proximate cause-as well as any of the other elements of negligence and negligent misrepresentation.
Associate Justice of the Court of Appeal, Third Appellate District, assigned by the Chief Justice pursuant to article VI, section 6 of the California Constitution.
However, the pertinent regulations are now under review and subject to imminent change. In November 2013, the Food and Drug Administration (FDA) proposed rule changes that would allow generic drug makers to revise their product warning labels, and depart from the labeling of the brand-name drug, using the "changes being effected" process. (Supplemental Applications Proposing Labeling Changes for Approved Drugs and Biological Products, 78 Fed.Reg. 67985 (Nov. 13, 2013) (Supplemental Applications).) In view of the Supreme Court's preemption decisions, the FDA explained it sought to "create parity" between brand-name and generic manufacturers. ( Id . at p. 67989.) Moreover, the change was needed to ensure postmarket safety surveillance, because the FDA had found that safety-related labeling changes are typically required many years after a drug's initial approval, when generic versions are widely prescribed and the original brand-name manufacturer may have left the market. ( Id . at p. 67988.) In 2015, the FDA reopened the comment period and scheduled a public meeting on the proposed rule change. (80 Fed.Reg. 8577 (Feb. 18, 2015).) If this regulatory change is implemented, our decision to allow suits against brand-name manufacturers for injuries caused by generic equivalents will need to be reevaluated because the rationale supporting it will be largely, if not entirely, undermined. (See maj. opn. ante , at 226 Cal.Rptr.3d at pp. 350-351, fn. 2, 407 P.3d at pp. 30-32, fn. 2.)
Rowland v. Christian
(1968)
One example of tort liability leapfrogging scientific knowledge occurred in the Bendectin litigation. Several cases alleging the anti-nausea drug Bendectin caused birth defects went to trial in the 1980s, leading the manufacturer to withdraw the drug from the market. (Sanders,
From Science to Evidence: The Testimony on Causation in the Bendectin Cases
(1993)
Moreover, the court's holding in
Centinela Freeman
included a scienter requirement, specifying that health care plans may be liable for negligent delegation of financial responsibility if they "knew or should have known" the transferee would not be able to pay. (
Centinela Freeman
,
supra
, 1 Cal.5th at pp. 1001-1002,
Nor are indemnity agreements a satisfactory answer. After today's holding, drug companies subject to suit in California will undoubtedly include indemnity provisions in all NDA transfer contracts. Such provisions could effectively put the burden of liability back where it rightfully belongs, i.e., on the actual manufacturer of the product used by the plaintiff. But they would not necessarily relieve the predecessor of all costs related to future claims. Indemnification rights may be capped or may exclude costs in defending the underlying lawsuits. There may also be costs if the predecessor must sue to enforce the indemnity agreement.
Related
Cite This Page — Counsel Stack
407 P.3d 18, 226 Cal. Rptr. 3d 336, 4 Cal. 5th 145, Counsel Stack Legal Research, https://law.counselstack.com/opinion/th-v-novartis-pharmaceuticals-corporation-cal-2017.