Court: District Court, District of Columbia; February 4, 2019; Federal District Court
This multi-district litigation involves product liability claims against GlaxoSmithKline LLC (GSK) alleging that the drug Zofran (ondansetron) caused birth defects when used by pregnant women. GSK has sought summary judgment on the grounds of federal preemption, arguing that state-law claims regarding inadequate warning labels are overridden by federal law. The court denies this motion.
Zofran, an anti-emetic approved by the FDA in 1991 for specific uses but never for pregnancy-related nausea, has been prescribed off-label to pregnant women, reportedly due to GSK's alleged unlawful marketing. In 2010, the FDA requested safety data from GSK regarding Zofran's use in pregnancy, and although GSK submitted an analysis, no labeling changes were mandated. Subsequent requests to revise the label to indicate potential fetal risks were also denied by the FDA, which has consistently stated that available data does not confirm an association between Zofran and adverse fetal outcomes.
Plaintiffs assert that Zofran does cause birth defects and that GSK's failure to include a warning about this on the label constitutes tort liability under state law. They claim that GSK misrepresented data to the FDA, which led to incomplete assessments. GSK counters that it did not withhold information and that any state-law claims are preempted by federal law, which strictly regulates drug labeling.
The tension between federal regulation and state liability principles arises because federal law, as enforced by the FDA, generally restricts label changes to those approved by the agency, while state law allows for claims based on inadequate warnings. However, there is a federal provision, known as the "changes being effected" (CBE) process, that permits drug companies to unilaterally update labels based on new safety information, pending FDA approval. The Supreme Court has established that for federal preemption to apply, a drug company must present "clear evidence" that the FDA would have rejected the required state law warning.
The FDA rejected the plaintiffs' proposed warning, but they argue that GSK's incomplete disclosures to the FDA led to uninformed decisions by the agency, thus negating a finding of federal preemption. The legal framework for preemption claims is unclear, with ongoing debates about standards of proof and the roles of judges and juries. The U.S. Supreme Court has agreed to review an FDA preemption case to provide clarity. Meanwhile, the Court must operate within the current legal framework, which may later be found incorrect. The key factual issue regarding preemption—whether the FDA would have rejected the label with the evidence purportedly withheld by the plaintiffs—must be determined by a jury. Consequently, GSK's motion for summary judgment based on preemption is denied.
Under federal law, drug companies must obtain FDA approval before marketing a new drug, requiring them to submit a comprehensive New Drug Application (NDA) that includes safety and efficacy data and proposed labeling. FDA regulations mandate that NDAs fully disclose all relevant safety information and that companies update their NDAs with new safety information as it becomes available. This obligation extends beyond the initial NDA submission, emphasizing the sponsor's responsibility to keep the FDA informed of significant new risks associated with the drug.
The FDA approval process for drugs is characterized as "onerous and lengthy," requiring a New Drug Application (NDA) to demonstrate that a drug is safe, effective, and appropriately labeled (21 U.S.C. § 355(c)(1)(A)). The FDA not only approves the drug but also the exact wording of its label, with manufacturers generally prohibited from altering the label without FDA approval (Wyeth, 555 U.S. at 565). After drug approval, the FDA can mandate label changes based on new safety or efficacy information (21 U.S.C. § 355(o)(4)). Manufacturers bear the ongoing responsibility for ensuring their labels are accurate and sufficient (Wyeth, 555 U.S. at 570-71).
To amend drug labels, manufacturers have two options: they can submit a "Prior Approval Supplement" (PAS) for FDA approval before implementing changes (21 C.F.R. § 314.70(b)), or they can use the "changes being effected" (CBE) process to unilaterally strengthen warnings when new significant information arises (21 C.F.R. § 201.57(c)(6); § 314.70(b)(2)). The term "newly acquired information" encompasses not only new data but also additional analyses or reports that reveal previously unrecognized risks (21 C.F.R. § 314.3).
In contrast to consumer products, where manufacturers often provide extensive warnings to limit tort liability, the FDA adopts a more measured approach for pharmaceuticals. The FDA avoids approving warnings based solely on speculative risks, as this could mislead consumers and undermine the value of important risk information.
The FDA will reject a Prior Approval Supplement (PAS) application or a Changes Being Effected (CBE) amendment if there is inadequate evidence linking drug use to adverse events. The FDA's standard for requiring warning labels differs from state tort law requirements, which mandate that drug manufacturers must label their products to ensure reasonable safety if aware of dangers. Manufacturers of ethical drugs have a duty to provide timely warnings to the medical community about known side effects.
Specific labeling provisions apply to drugs for pregnant women. Until June 30, 2015, the FDA categorized drugs based on pregnancy safety into five categories (A, B, C, D, or X). For a drug classified as Category B, if animal studies indicate no risk to the fetus but lack adequate studies in pregnant women, the label must state this, along with a caution about using the drug during pregnancy only if necessary. This classification system was removed following a 2014 FDA final rule on pregnancy and lactation labeling.
Zofran (ondansetron hydrochloride) is a prescription drug approved by the FDA on January 4, 1991, for nausea and vomiting related to chemotherapy, radiation therapy, and postoperative conditions. It has been approved in various formulations over the years. While nausea and vomiting during pregnancy (NVP) affects a significant number of women, Zofran was never approved for this use, nor did the manufacturer seek such approval. Nevertheless, physicians can legally prescribe drugs off-label, although pharmaceutical companies cannot promote these uses.
Many physicians have prescribed Zofran to pregnant women, especially for hyperemesis gravidarum (HG). Initially classified as a pregnancy category B drug by the FDA in 1991, the label indicated that while animal studies showed no harm to fetuses, there were no adequate human studies, advising its use during pregnancy only if necessary. There has never been a contraindication against using Zofran for pregnant women on its label.
In December 2010, the FDA requested that GlaxoSmithKline (GSK) analyze existing literature on Zofran's use in pregnancy and lactation due to its common use during that time. GSK responded in April 2011, stating that the use of Zofran in human pregnancy was not established and not recommended, concluding that no label changes were warranted. The FDA did not respond or modify the label.
In January 2013, James P. Reichmann submitted a citizen petition to the FDA, requesting a reclassification of Zofran's pregnancy risk category and warning obstetricians about potential adverse outcomes associated with its use. The FDA denied this petition on October 27, 2015, stating there was insufficient evidence to conclude that Zofran increased the risk of birth defects or adverse fetal outcomes. The FDA considered both pre- and post-approval data, including GSK's submissions, and found no evidence supporting increased risks associated with Zofran use in pregnancy.
The FDA affirmed that ondansetron, marketed as Zofran, is appropriately categorized as pregnancy category B, indicating no established increased risk to the fetus. It rejected a request to inform doctors that Zofran is unsafe for use during pregnancy, citing a lack of supporting data for such a claim. In 2015, Novartis, which acquired Zofran from GSK, proposed updates to the pregnancy labeling to recommend against its use during pregnancy, citing possible risks to the fetus. Proposed revisions included warning language about potential harm to the fetus and emphasizing that animal studies may not predict human responses. Novartis submitted a substantial clinical overview summarizing its rationale for the proposed changes, including adverse event data, but acknowledged that evidence did not consistently indicate major birth defects from ondansetron exposure in early pregnancy. The FDA rejected these revisions in November 2015, removing specific warnings and the proposed section on reproductive potential, due to insufficient human data on congenital malformations. In December 2015, Novartis made another attempt to revise the labeling, suggesting that reports of congenital malformations warranted warnings about the lack of established safety for ondansetron during pregnancy and noting that the drug had not been studied in pregnant women for nausea and vomiting prevention.
In April 2015, the FDA rejected the proposal to label ondansetron as "not recommended" for use during pregnancy, clarifying that "limitations on use" statements should not restrict off-label use unless there is a known risk that outweighs therapeutic benefits. The FDA determined that there was insufficient evidence to conclude that ondansetron, marketed as Zofran, posed a significant risk of major congenital malformations in humans. By September 2016, the FDA and Novartis agreed on a revised label, which included statements indicating that available human data did not support a causal relationship between ondansetron and congenital malformations. The FDA noted no consistent safety concerns justifying a warning against its use during pregnancy and highlighted that evidence did not demonstrate Zofran's ineffectiveness for treating nausea and vomiting in this context. The FDA also stated that cardiac malformations, being common in the general population, could occur by chance among Zofran users without implying a drug-related risk. The 2016 label reflected that existing data did not reliably establish an association between Zofran and adverse fetal outcomes, and limitations in methodologies of studies reporting such associations were acknowledged. Plaintiffs allege that GSK failed to provide the FDA with material safety information regarding Zofran before and after its approval in 1991, claiming that this led to the drug being inaccurately classified as a pregnancy category B drug and to the FDA's refusal to implement warning labels for pregnant women.
Plaintiffs allege that GSK omitted critical evidence in its submissions to the FDA, focusing on four main categories: results from Japanese animal studies, Zofran's biological mechanism of action, adverse event data, and details from the Einarson birth defect study.
1. **Japanese Animal Studies**: The studies in question (100423, 100424, and 100441) were conducted by GSK’s affiliate, Glaxo Nippon, between 1988 and 1990. Dr. Bengt Danielsson, an expert for the plaintiffs, claims these studies indicate that Zofran has teratogenic effects, citing increased embryofetal death and malformations in treated animals compared to controls.
- **Study No. 100423**: Plaintiffs assert it reported increased embryofetal death in a Zofran-treated rat group. GSK counters that this was a preliminary study with no observed teratogenic effects.
- **Study No. 100424**: Plaintiffs claim it showed increased embryonic death and malformations, including ventricular septal defects. GSK argues that the study found no significant differences in the control groups and that Dr. Danielsson acknowledged no external malformations were observed.
- **Study No. 100441**: Plaintiffs allege it reported increased skeletal defects in Zofran-treated rabbits. GSK contends that observed defects were likely due to decreased maternal weight rather than Zofran exposure, with study investigators concluding no teratogenicity was evident.
GSK firmly disputes the plaintiffs' interpretations of these studies, asserting that their findings do not support the claims of teratogenic effects associated with Zofran.
GSK is accused of withholding three Japanese animal studies from the FDA, which allegedly contained animal reproduction data indicating adverse fetal effects. Zofran was initially approved on January 4, 1991, and in 1992, seven Japanese reproductive toxicology studies were published, two of which are central to this case. While GSK partially disclosed the existence of studies numbered 100423, 100424, and 100441 in a December 23, 1993, Annual Report, it did not provide the studies themselves, which were only available in Japanese. An FDA review in 1997 found no teratogenic effects associated with Zofran based on Study No. 100424. In 2014, the FDA requested comprehensive details about animal reproduction studies for updating Zofran's pregnancy label, to which GSK responded in March 2015, but plaintiffs argue this response omitted the three Japanese studies relevant to the request. The FDA's subsequent denial of a citizen petition related to Zofran referenced GSK's safety evaluation but did not mention the Japanese studies. GSK contends it was not required to provide additional information as the FDA had not sought clarification regarding the citizen petition. Plaintiffs also allege that GSK failed to accurately disclose Zofran's potential to cause embryonic arrhythmias linked to teratogenic effects, specifically alleging that Zofran inhibits hERG potassium channels, disrupting cardiac rhythm and potentially causing fetal heart defects.
Plaintiffs argue that GSK was aware of the hERG channel mechanism as early as 2002 but failed to inform the FDA adequately. GSK claims the hERG channel mechanism is merely a hypothesis without supporting evidence and asserts that the FDA reviewed the evidence and found insufficient data for a pregnancy warning. In 1994, a study noted that Zofran inhibits hERG potassium channels, potentially leading to QT prolongation, which GSK does not dispute but claims was not specifically about Zofran’s effects in humans. A 2000 study similarly linked Zofran to QT prolongation due to hERG channel inhibition, which GSK also acknowledges but states was not human-focused. An internal GSK document from 2002 indicated that drugs causing embryonic heart rate reduction could lead to embryonic death, suggesting GSK's awareness of related mechanisms. GSK maintains that the mechanism discussed does not pertain to Zofran and that no embryolethality or teratogenic effects were observed in animal studies. GSK also challenges the reliability of expert Dr. Danielsson's opinions. In a 2015 submission to the FDA, Novartis referenced Danielsson’s work on hERG channel blockade and developmental toxicity but ultimately dismissed the mechanism based on Zofran's reproduction studies, which did not indicate increased risks. The FDA rejected Novartis's request for a pregnancy warning in 2016, citing a lack of evidence for adverse fetal outcomes associated with Zofran.
Plaintiffs allege that GSK failed to properly disclose adverse event reports related to Zofran beginning in 2005, incorrectly coding these events and omitting them from the Zofran safety database, which affected data analysis submitted to the FDA. Under FDA regulations, manufacturers must fully disclose all adverse event data during the New Drug Application (NDA) process and afterward, including individual case report forms for patients who died or dropped out due to adverse events. GSK coded these events using the Medical Dictionary for Regulatory Activities (MedDRA), categorizing cardiac-related congenital adverse events under six distinct System Organ Classes (SOCs). Plaintiffs argue that a 2014 analysis by GSK, responding to an FDA request regarding Zofran use in pregnancy, was limited to only two SOCs—cardiac disorders and pregnancy-related conditions—resulting in an underreporting of congenital cardiac adverse events and potentially obscuring the risk of birth defects in the summary provided to the FDA. In response, GSK denies the allegations of miscoding, disputes the submission of the 2014 analysis to the FDA, asserts that it regularly provided extensive information on pregnancy-related events, and claims that plaintiffs cannot demonstrate that different coding would have revealed an increased risk. GSK also contends that the FDA considered and ultimately rejected the need for pregnancy warnings after finding adverse event reports insignificant compared to the background incidence rate of heart defects.
Plaintiffs in the Einarson Birth Defect Study allege that GSK misled the FDA and the medical community regarding a 2004 study, claiming it proved Zofran's safety during pregnancy. The study, authored by Adrienne Einarson and published in September 2004, reportedly showed a lack of teratogenic effects, which GSK highlighted while failing to disclose its involvement in editing the study. Plaintiffs assert that GSK omitted significant birth defect data and ignored critical feedback from its scientists about the study's flaws. GSK contends that the study is irrelevant to the preemption issue since the FDA considered its findings during the drug's labeling approval.
The legal standard for summary judgment involves determining whether there is a genuine need for trial, focusing on material facts and evidence that support each party's claims. Summary judgment is granted when no genuine dispute exists, requiring the nonmoving party to present specific facts rather than mere allegations. The analysis also addresses FDA preemption, noting that federal law can preempt state law when Congress explicitly states so, intends federal law to dominate, or when state law conflicts with federal statutes. The case involves "conflict" or "obstacle" preemption, where compliance with both federal and state laws is impossible or where state law obstructs federal objectives.
The preemption analysis begins with the Supreme Court case Wyeth v. Levine, which examined whether state law failure-to-warn claims against drug manufacturers are preempted by federal law when the FDA approved the drug's warning label. The Court determined that because the Changes Being Effected (CBE) regulations allowed manufacturers to independently strengthen their warning labels, it was not impossible for them to comply with both federal and state labeling requirements, absent clear evidence that the FDA would not have approved such changes.
This "clear evidence" standard was reiterated in PLIVA, Inc. v. Mensing, where the Court distinguished the cases. Unlike Wyeth, the generic manufacturer in PLIVA could not unilaterally change its label and needed FDA permission to comply with state law. Thus, the Court concluded that it was impossible for the manufacturer to meet both federal and state labeling requirements, leading to the preemption of state-law claims.
The distinction drawn by PLIVA limits Wyeth to scenarios where a manufacturer can independently modify its label per state tort duties. The line established between changes that can be made under the CBE regulation and those requiring prior FDA approval complicates the application of the Wyeth preemption standard, which has been criticized for being cryptic and challenging for lower courts to navigate.
The First and Third Circuits have encountered this standard, with the First Circuit in Gustavsen v. Alcon Laboratories, Inc. addressing a "major" change requiring FDA approval, thus preempting state law claims. In Celexa, the First Circuit determined that the plaintiffs lacked "newly acquired information" allowing for a label change through the CBE process, leading again to preemption of the claims. Overall, the application of the Wyeth and PLIVA standards remains complex and poses significant interpretative challenges.
A drug manufacturer can successfully argue a preemption defense if it demonstrates either that the Changes Being Effected (CBE) process was unavailable, preventing unilateral label changes, or if it provides "clear evidence" that the FDA would not have approved the proposed label changes claimed by the plaintiffs. The issue of whether preemption is an affirmative defense for which the manufacturer carries the burden of proof is affirmed by case law. Impossibility preemption is characterized as a demanding defense.
In the analysis of post-FDA approval preemption, a two-stage process is proposed, as seen in Utts v. Bristol-Myers Squibb Co. The plaintiff must first identify "newly acquired information" that would allow the manufacturer to amend the label under the CBE regulation without FDA approval. However, even with such information, the manufacturer can still argue preemption if it shows that the FDA would not have approved the suggested change.
This burden-shifting framework raises potential complications, as it is not universally accepted by appellate courts. Additionally, it raises questions about whether all plaintiffs in failure-to-warn cases must preemptively demonstrate the existence of "newly acquired information," or only when the manufacturer asserts preemption. This could lead to a cumbersome legal process where the burden shifts multiple times between parties. The First Circuit has indicated that a plaintiff must show that the complaint includes a labeling deficiency that the manufacturer could correct using the CBE process, thereby implying that the manufacturer’s ability to utilize the CBE must be established in the context of the specific changes in question.
The court determined that the complaint did not plausibly rely on newly acquired information, leading to the conclusion that FDA preemption is treated as an affirmative defense, similar to other defenses, without a burden-shifting framework. The defendant retains the burden of proof, whether moving to dismiss based on insufficient allegations or seeking summary judgment on insufficient evidence. A manufacturer may demonstrate the unavailability of the Changes Being Effected (CBE) process or that the FDA would not have approved a proposed label change by addressing weaknesses in the complaint or evidence.
The issue of whether preemption is a question of law for the judge or a question of fact for the jury remains unresolved. Citing United States v. Rhode Island Insurers' Insolvency Fund, the court noted that federal preemption typically presents a legal question but acknowledged the complexity of the current case involving scientific and regulatory factors, which includes disputed factual issues. The court highlighted the need for evidence and factual findings, referencing the Wyeth opinion's requirement for "clear evidence." The Third Circuit's ruling in Fosamax concluded that determining "clear evidence" is a question of fact involving assessing probabilities, conflicting evidence, and motives—tasks traditionally assigned to a jury.
Motive and state of mind inquiries regarding FDA officials' considerations are typically fact questions for the jury, not the court. The Ninth Circuit reinforced this view by reversing a summary judgment order that dismissed "new safety information" as irrelevant, asserting that uncertainty about the FDA's consideration of such information created a material fact dispute precluding summary judgment on preemption claims. Other circuits, however, have ruled differently, with some granting summary judgment to defendants based on undisputed facts, arguing that no reasonable jury could find that the FDA would have approved proposed warnings under the CBE regulation. A California court explicitly stated that federal preemption is a legal question for the court, not a jury, noting that decisions on proposed label changes and "clear evidence" of FDA approval are legal determinations. This court emphasized the importance of resolving federal supremacy issues early in litigation, which is feasible only if treated as a question of law. Despite the complexity and significance of the regulatory issues, the general principle remains that factual disputes should be resolved by a jury, casting doubt on the authority to deviate from this norm.
The Court concludes, in the absence of appellate guidance, that the Third Circuit's analysis in Fosamax is likely correct, indicating that whether there was "newly acquired information" and if the FDA would have rejected a proposed label change are factual questions for a jury. These issues can be resolved through a motion to dismiss or summary judgment under appropriate circumstances. The Court raises a concern about the fairness of having jurors decide complex scientific and regulatory questions, suggesting that this issue should be addressed by Congress or a higher court.
Regarding the "clear evidence" standard, the Third Circuit noted its undefined nature despite increasing application in preemption cases. The Fosamax court interpreted "clear evidence" to mean proof by clear and convincing evidence, contrasting with other circuits that use the facts from Wyeth as a benchmark for determining whether evidence is compelling enough to establish FDA rejection of a label change. The Court expresses concerns that this "clear evidence" standard is vague and may complicate district court applications.
For the purposes of the current summary judgment motion, the Court assumes that "clear evidence" does not impose a higher burden than the preponderance of the evidence standard. However, it emphasizes that a materiality standard should apply to claims of false statements and omissions to the FDA. A false statement is deemed material if it can influence the decision-making body. Importantly, as preemption is an affirmative defense, GSK must demonstrate that the omitted information was not material to succeed in its defense; plaintiffs are not required to prove materiality to defeat preemption.
The legal standard for evaluating claims related to FDA disclosures may be either subjective, reflecting the actual thoughts and motives of FDA officials at the time, or objective, based on what a reasonable FDA official would do in similar circumstances. The Third Circuit in Fosamax suggested a subjective standard by emphasizing the understanding of FDA officials' mental states, while other courts have leaned towards an objective standard, focusing on the reasonable actions of FDA officials based on available information. Plaintiffs argue that GSK made false statements and omissions in its disclosures to the FDA, typically assessed under an objective standard of materiality, as per the Restatement (Second) of Torts. The document argues that an objective standard is more practical given the complexity of the FDA and the historical context of the decisions involved. Consequently, the court will proceed with the assumption that the standard is objective, evaluating the impact of the alleged misstatements on a reasonable FDA regulator.
Additionally, the text addresses whether claims of fraud-on-the-FDA should be analyzed differently from the Wyeth standard. This question arises from the Supreme Court's ruling in Buckman Co. v. Plaintiffs' Legal Comm., where the Court held that state-law fraud-on-the-FDA claims were preempted by federal law. In Buckman, plaintiffs alleged that fraudulent misrepresentations made to the FDA led to their injuries, but the Supreme Court determined that such claims conflicted with federal regulations and were therefore preempted.
The conflict arises from the FDA's authority to combat fraud, which is designed to maintain a balance of statutory objectives. Allowing state tort claims related to fraud against the FDA could disrupt this balance. The Wyeth court differentiated this case from Buckman, noting that Buckman dealt with state-law fraud-on-the-agency claims, while the current claims are state-law failure-to-warn product-liability claims primarily based on alleged misrepresentations violating FDA disclosure requirements. Consequently, Buckman does not apply here.
Regarding summary judgment, preemption is an affirmative defense that GSK must prove, and it presents a factual question. The court will not determine the standard of proof at this stage but notes that the omitted information must be material and assessed objectively based on its potential impact on a reasonable regulator. GSK must show either that it could not have changed the label through the Changes Being Effected (CBE) process or that the FDA would have rejected any proposed changes that state law would have required. Plaintiffs argue that GSK could have unilaterally changed its label after the initial approval and claim that the FDA's rejection of enhanced warning labels was based on incomplete evidence due to GSK withholding information.
Plaintiffs argue that had the FDA received specific information regarding the use of GSK's drug during pregnancy, it would have mandated stronger warnings. GSK must demonstrate either that the CBE (Changes Being Effected) process was unavailable to them or that the FDA would have rejected a proposed label change. GSK needs to prove that the FDA was adequately informed of the relevant scientific data, and that any alleged omissions were not material, meaning the disputed information was either already known to the FDA or would not have influenced its decision.
GSK asserts that it provided appropriate safety information at the relevant times and claims there was no "newly-acquired" data that warranted a label change under the CBE process. GSK also argues that such changes regarding "use in specific populations" could only be made through a PAS (Prior Approval Supplement) request.
In support of their claims, plaintiffs submitted an expert report from Dr. Brian E. Harvey, who critiques GSK's disclosures and the significance of any omissions. The parties agree that GSK referenced three Japanese animal studies in its 1993 Annual Report but GSK contends this was adequate under federal regulations, asserting that it was the FDA's choice to accept citations rather than full reports. GSK claims these studies did not constitute "newly acquired information" under CBE regulations.
Conversely, plaintiffs, supported by Dr. Harvey's report, argue that GSK's classification of Pregnancy Category B was based on earlier studies, while the results of the Japanese studies were not fully disclosed to the FDA, potentially violating regulatory standards that require comprehensive reporting of all pertinent studies. Dr. Harvey emphasizes that any adverse effects observed in these studies should have been submitted to the FDA for thorough evaluation, underscoring the obligation of drug sponsors to disclose all relevant reproductive toxicology studies without selective omission.
A brief description of study results must be provided if the study is completed or interim results are known, as per regulatory requirements. There is uncertainty surrounding GSK's compliance with this regulation. Dr. Harvey's report references 21 C.F.R. 312.50 but only broadly states that "regulatory standards" necessitate the submission of full reports, indicating that GSK cannot select only favorable studies. GSK has not submitted any expert affidavit or cited specific regulations to support its disclosures, aside from arguing that the full reports do not qualify as "newly acquired information" under 21 C.F.R. 314.3. Consequently, the Court cannot ascertain the completeness and accuracy of GSK's disclosures to the FDA.
Assuming GSK failed to disclose necessary information, the inquiry continues as GSK may argue that the omitted information was not material, meaning it would not have impacted the FDA's decision. Plaintiffs primarily rely on Dr. Harvey's expert report, which asserts GSK's duty to disclose results from three Japanese studies before FDA approval. Dr. Harvey contends that had GSK disclosed these results, the FDA would have considered all seven reproductive toxicology studies, ultimately leading to a classification of Zofran as Pregnancy Category C instead of B due to evidence of fetal harm.
GSK had multiple opportunities to provide comprehensive information about reproductive toxicology studies, which should have been disclosed in the 1997 NDA application and by the end of 1991 for the supplement. If GSK had complied, Zofran would have been classified as a Pregnancy Category C drug, and appropriate labeling would have included warnings about potential teratogenic effects. Dr. Harvey also indicates that had GSK properly reported the Japanese animal study results, this information would have been critical for the FDA's review regarding a proposed label change, likely altering the outcome of Novartis' label change attempt.
Novartis's proposed labeling changes were based solely on safety signals from adverse event reports. The FDA, assuming animal studies showed no teratogenicity, declined to implement stronger pregnancy warnings. If the FDA had received comprehensive evidence, it likely would have approved Novartis's proposed changes. GSK argues that the submission of full Japanese animal studies would not have influenced the FDA’s decision, claiming that it provided extensive data on thousands of animals exposed to Zofran during preclinical studies. GSK maintains that the risks identified in the Japanese studies did not differ in type or severity from existing data. FDA scientists concluded that the preclinical data did not indicate teratogenicity, a finding echoed by investigators of the Japanese studies, who reported no teratogenic effects. However, GSK has not submitted expert testimony to support its position. A significant factual dispute exists regarding whether GSK adequately disclosed the Japanese studies to the FDA and whether such disclosures would have been relevant to the FDA's decision-making. Consequently, the court cannot definitively rule out the possibility that these studies constituted "newly acquired information" that might have revealed new risks or that the FDA would have rejected the proposed label changes with a more thorough disclosure.
Additionally, there is contention over GSK's disclosure of adverse event data to the FDA and its potential impact on labeling decisions. GSK does not fully defend its coding of adverse event data but argues that the plaintiffs inaccurately claim that only coded lists were provided. GSK asserts that it supplied detailed information regarding adverse events, including those related to pregnancy, as required by regulations. GSK further contends that the plaintiffs cannot demonstrate that different coding would have indicated an increased risk of birth defects, that the adverse event data did not reveal new risks, and therefore, GSK had no obligation to amend the label using the CBE process. GSK notes that the FDA disregarded the significance of adverse events in discussions about label changes, highlighting the challenge of relying on case reports of congenital abnormalities due to the background incidence rates.
Clear evidence indicates that the FDA would not have approved a label change for Zofran based on the data provided by GSK. Plaintiffs reference Dr. Harvey's report, which asserts that the FDA considers adverse event reports among other evidence when evaluating a drug's safety. However, GSK failed to report all relevant adverse events related to Zofran’s use during pregnancy, did not adequately respond to the FDA's requests for information on animal reproduction studies, and did not provide sufficient details on the drug's mechanism of action. This lack of comprehensive data deprived the FDA of the necessary information to assess the drug's safety in pregnancy accurately.
The former Division Director, who oversaw Zofran regulation, stated that the omitted information was significant and would have been important to any reasonable FDA official. GSK has not presented expert testimony regarding its disclosure obligations, and while Dr. Harvey noted gaps in GSK's reporting, he did not specifically state instances of misreporting or assert that the FDA would have altered its stance based on different data coding. The burden to prove preemption lies with GSK, and there appears to be a material factual dispute regarding whether GSK appropriately coded and disclosed adverse event data. The court cannot definitively conclude that the adverse event data was not "newly acquired information" or that the FDA would have rejected proposed label changes had more comprehensive information been provided.
Additionally, plaintiffs argue that GSK's failure to disclose an accurate description of Zofran's biological mechanism of action constitutes a significant omission, suggesting GSK was aware of this data but chose not to share it with the FDA.
Plaintiffs assert that GSK's failure to disclose critical information led Novartis to undervalue Dr. Danielsson's evidence regarding the hERG channel mechanism in its 2015 FDA submission. They also argue that a 2016 FDA statement, which cited a lack of evidence as a reason for denying Novartis's request to add a pregnancy warning, supports their claim that the non-disclosure was material. GSK counters that the hERG mechanism presented by Dr. Danielsson does not qualify as "newly acquired information" since the FDA had previously reviewed this evidence and found it insufficient for a pregnancy warning. The record lacks substantial evidence on this matter. Dr. Harvey's report indicates that the FDA evaluates the totality of evidence when assessing drug safety, but GSK allegedly failed to provide adequate mechanism of action data, leaving the FDA without all relevant safety information for Zofran in pregnancy. Dr. Harvey suggests that had GSK used the CBE process to update Zofran's label, the FDA might have permitted changes based on existing evidence linking Zofran to birth defects.
Regarding the Einarson birth defect study, plaintiffs claim GSK did not disclose its involvement or internal criticisms of the study's validity and that a significant birth defect event was omitted from the published study but recorded in GSK's safety database. They argue these omissions are relevant to the FDA's assessment of Zofran's teratogenicity. GSK contends that the Einarson study is not "newly acquired information" as the FDA had reviewed it and acknowledged GSK's involvement. GSK also claims the omitted adverse event is not material since it is not associated with any plaintiff's allegations. Dr. Harvey's report further notes that GSK had information on adverse events from the Einarson study that were not included in the published abstracts, yet GSK promoted the study to healthcare providers as showing "no birth defects."
GSK had knowledge of unreported birth defects from the Einarson study that were not disclosed in the study's abstracts or final publication. Despite this, GSK included the Einarson abstracts in its voluntary handouts to healthcare providers regarding Zofran's safety in pregnancy, violating its own guidelines that require Medical Information documents to present all known information, both favorable and unfavorable. Dr. Harvey's report indicates that GSK exhibited an inconsistent reporting pattern concerning birth defects from the Einarson study and stated that these defects should have been reported to the FDA. However, Dr. Harvey did not assert that GSK violated any disclosure requirements to the FDA, and GSK has not provided a counter-affidavit, leading to a disputed factual issue regarding GSK's disclosure practices.
Regarding regulatory processes, GSK argued that the Changes Being Effected (CBE) process could not be used to amend the "use in specific populations" section of its label, asserting that changes to this section could only be made through a Prior Approval Supplement (PAS) request. GSK's position is based on the FDA's omission of this section from the CBE regulation revisions in 2008, suggesting that CBE changes are limited to explicitly enumerated sections. Plaintiffs counter that the CBE regulation existed before the "use in specific populations" section was added in 2006 and was not amended to restrict changes to that section. They note that the only prohibition on unilateral CBE changes pertains to the "Highlights of Prescribing Information," which is separate from the proposed changes regarding pregnancy categorization.
The CBE (Changes Being Effected) process was available for Zofran throughout its market life, allowing the manufacturer to unilaterally enhance labeling based on new safety information, including for pregnancy. The plaintiffs' interpretation of the CBE regulation appears correct, as the FDA did not intend to restrict changes to labeling about "use in specific populations." It is deemed unlikely that the FDA would prohibit a drug company from voluntarily updating labels to protect vulnerable populations, such as pregnant women, upon discovering new risks post-approval. Consequently, GSK's motion for summary judgment, asserting that the CBE process could not be utilized to modify the "Use in Specific Populations" label, has been denied.
The record reveals disputed material facts regarding GSK's claim of federal preemption. GSK has not demonstrated that it was unable to use the CBE process to issue stronger warnings about Zofran in pregnancy or that there is "clear evidence" that the FDA would have rejected such warnings. Therefore, GSK's request for summary judgment based on federal preemption is denied. Additionally, Glaxo, Inc. originally sponsored Zofran's new drug application. Citizens can petition the FDA for drug label changes, and GSK was required to update the Zofran label to comply with the new Pregnancy and Lactation Labeling Rule established in December 2014. The FDA also reviewed post-marketing data and scientific literature in its assessments. Despite acknowledging a study indicating that ondansetron could influence cardiac repolarization, GSK's Director of Safety Pharmacology expressed that the evidence linking ondansetron to QT prolongation was not convincing and its clinical relevance remained uncertain.
Plaintiffs reference GSK's 2011 Lamictal label to support their claim that GSK was aware of the drug's mechanism of action. The definition of an "adverse event" is provided as any medical issue related to drug use in humans, according to 21 C.F.R. 312.32(a). The court's approach in McGee v. Boehringer Ingelheim Pharmaceuticals highlights the importance of the statute of limitations as an affirmative defense; if a plaintiff’s complaint suggests the claim was filed outside the limitations period, a defendant may seek dismissal, but the plaintiff can counter this at summary judgment by showing a genuine issue of material fact regarding timeliness, possibly relying on the discovery rule. The text notes that Zofran had five NDAs approved from 1991 to 1999, and plaintiffs argue that GSK could have updated the label during the NDA application process. Dr. Harvey, who served as Director of the Division of Gastroenterology Products from 2005 to 2007, was involved in label negotiations for Zofran. GSK has filed a motion to exclude Dr. Harvey's conclusions under Fed. R. Civ. P. 702 and Daubert standards, but the memorandum assumes his opinions may be admissible. The court indicates that even if omitted information is deemed material, it does not guarantee a different outcome. In summary judgment, the court must favor the non-moving party, assuming that any omitted information could potentially impact the outcome.
