MOLNAR et al v. MERCK & CO., INC.

Filing 4541

OPINION filed. Signed by Chief Judge Freda L. Wolfson on 3/23/2022. (jem)

*FOR PUBLICATON*
UNITED STATES DISTRICT COURT
DISTRICT OF NEW JERSEY
_______________________________________
IN RE FOSAMAX (ALENDRONATE
SODIUM) PRODUCTS LIABILITY
LITIGATION

MDL No. 2243
Civil Action No. 3:08-08 (FLW)

THIS OPINION RELATES TO: ALL
ACTIONS

OPINION

WOLFSON, Chief Judge:
In this failure-to-warn case, more than 500 individuals (“Plaintiffs”) who took
Fosamax, a drug manufactured by Defendant Merck Sharp & Dohme (“Defendant” or
“Merck”) to prevent and treat osteoporosis in postmenopausal women, brought suit
claiming that they suffered atypical femoral fractures between 1999 and 2010. More
than eight years ago, following a bellwether trial, the late Hon. Joel A. Pisano,
U.S.D.J., granted summary judgment in favor of Merck, ruling that federal law
preempted Plaintiff’s state law failure-to-warn claims. 1 In re Fosamax (Alendronate
Sodium) Prod. Liab. Litig., 951 F. Supp. 3d 695, 701, 703-04 (D.N.J. 2013)
[hereinafter Glynn]. On appeal, the Third Circuit vacated and remanded this matter,
concluding that preemption presented “a question of fact for the jury,” not a question
of law for the judge. In re Fosamax (Alendronate Sodium) Prod. Liab. Litig., 852 F.
3d 268, 271, 293 (3d Cir. 2017) [hereinafter Fosamax], vacated and remanded, 139
S.Ct. 1668. And, in answering that question, the Third Circuit held that the jury
After Judge Pisano retired from the Court, the Multidistrict Litigation Panel
reassigned this MDL to me.
1

1

must apply a heightened standard of proof, sustaining the preemption defense only
if Merck proved it by “clear and convincing evidence.” Id. at 285-86. Merck, however,
petitioned for a writ of certiorari, which was granted by the United States Supreme
Court. In Merck Sharpe & Dohme Corp. v. Albrecht, 139 S.Ct. 1668, 1676, 1679-80
(2019), the Supreme Court vacated and remanded the Third Circuit’s decision,
holding that the preemption inquiry is “a legal one for the judge, not a jury.” Upon
remand, the Third Circuit returned the case to this Court to decide “in the first
instance whether the plaintiffs’ state law claims are preempted by federal law under
the standards described by the Supreme Court.” Order at 1, No. 14-1900 (3d Cir. Nov.
25, 2019). The Third Circuit further instructed this Court “to determine the effect of
the [Food and Drug Administration’s (“FDA” or “Agency”)] Complete Response Letter
[(“CRL”)] and other communications with Merck on the issue of whether such agency
actions are sufficient to give rise to preemption.” Id.
On remand, Merck reiterates its position that federal law preempts Plaintiffs’
state law failure-to-warn claims. In particular, Defendant relies on the FDA’s 2019
communication, in the form of a CRL, rejecting a warning concerning atypical femoral
fractures that Merck proposed. Plaintiffs, on the other hand, argue that the CRL is
not “clear evidence” that the FDA would have rejected any and all warnings. Having
reviewed the submission of the parties, the Court finds that based on clear and
convincing evidence, Defendant fully informed the FDA of the justifications for its
proposed warning, which was adequate under state law and encompassed the injury
Plaintiffs allege here, and the FDA, in turn, informed Defendant that it would not

2

approve changing the Fosamax label to include that warning in the CRL. Because the
FDA’s rejection was predicated on insufficient evidence of a causal link between
Fosamax and atypical femoral fractures, it is clear that the Agency would not have
approved a differently worded warning about such a risk. Plaintiffs’ state law failureto-warn claims are therefore preempted, and Defendant’s Motion for Summary
Judgment is GRANTED.
FACUAL BACKGROUND AND PROCEDURAL HISTORY
The factual background and procedural history of this case, which are largely
not in dispute, are primarily adopted from the Supreme Court and Third Circuit’s
decisions in this matter, as well as Judge Pisano’s dual decisions in Glynn and In re
Fosamax (Alendronate Sodium) Prod. Liab. Litig., 2014 WL 1266994, at *17 (D.N.J.
Mar. 22, 2014) [hereinafter OTSC Opinion].
A.

Fosamax

Merck manufactures Fosamax, a drug that treats and prevents osteoporosis in
postmenopausal women. Merck, 139 S.Ct. at 1668. Fosamax belongs to a class of
drugs called “bisphosphonates,” which operate on the “remodeling process,” where
the body breaks down bones and builds them back up. In postmenopausal women,
this process can “fall out of sync,” id. at 1673, such that the body removes old bone
cells faster than it replaces them. When resorption exceeds formation, the result is
osteoporosis, or low bone mass that increases the risk of fractures. Fosamax “slows
the breakdown of old bone cells and thereby helps postmenopausal women
avoid [such] fractures.” Id. However, by reducing resorption, the drug may cause

3

some microscopic stress fractures to develop into a specific type of stress fracture
known as atypical femoral fractures, or complete breaks that “cause great pain and
require surgical intervention to repair.” Id. at 1674.
A low energy, or also known as atypical, fracture is defined as one that is
caused by the equivalent of a fall from standing height or less, which involves
minimal force. A stress fracture is defined as a partial or complete fracture occurring
with either normal or increased activity, but without an identifiable external
traumatic event. Stress fractures, in this context, are included in the larger group of
low-energy fractures. In postmenopausal osteoporotic women, the proximal femur is
one of the most commonly affected sites for fractures, as are the pelvis, distal tibia
and metatarsals. See Def. Br., Ex. 1 at A2751-52.
B.

The Regulatory Framework for Drug Labeling

Congress has charged the FDA with ensuring that every prescription drug on
the market is “safe for use under the conditions prescribed, recommended, or
suggested” in its “labeling.” 21 U.S.C. § 355(d). As that directive suggests, labeling is
the “centerpiece” of the FDA’s risk management strategy for approved drugs, and the
primary means by which the FDA communicates its conclusions about drug safety to
the public. 71 Fed. Reg. 3922, 3944. Prospective drug manufacturers, such as Merck,
must work with the FDA to develop an appropriate label when they submit a new
drug for approval. 21 U.S.C. §§ 355(a), (b), (d)(7); 21 C.F.R. § 314.125(b)(6). The FDA

4

closely regulates the safety information on drug labels, down to the exact text of
warnings. 2 21 U.S.C. § 355(b)(1)(F); 21 C.F.R. § 201.57(a).
Drug labels include two sections relevant to this case: a “Precautions” section
and an “Adverse Reactions” section. The Precautions section narrowly describes
“clinically significant adverse reactions,” including any that are “serious even if
infrequent.” 21 C.F.R. § 201.57(c)(6)(i). The Adverse Reactions section more broadly
describes “the overall . . . profile of the drug based on the entire safety database,”
including a list of all “undesirable effect[s], reasonably associated with use.” Id. §
201.57(c)(7).
The threshold for placing a warning regarding an adverse event in the
Precautions section is “reasonable evidence of a causal association.” 21 C.F.R. §
201.57(c)(6)(iii) (providing that the Precautions section “must be revised to include a
warning about a clinically significant hazard as soon as there is [such evidence] . . . a
causal relationship need not have been definitely established”); Fed. Reg. 49,603,
49,604. The FDA designed this standard so as not to dilute “more important
warnings” or “deter appropriate use.” 73 Fed. Reg. at 49,605, 40,606. In other words,
the Precautions section is reserved for a “discrete set” of serious risks that would

2
In this context, the label “refers more broadly to the written material that is sent to
the physician who prescribes the drug and . . . that comes with the prescription bottle when
the drug is handed to the patient at the pharmacy.” Merck, 139 S.Ct. at 1672; 21 U.S.C. §
321(m). The label contains detailed information about the drug’s medical uses and health
risks. 21 U.S.C. § 355(b)(1)(F); 21 C.F.R. § 201.57(a). The FDA regulates the content, format,
and order of the safety information on the drug label. 21 C.F.R. § 201.57(c). Drug labels must
include, inter alia, warnings and precautions about potential safety hazards and adverse
reactions for which there is sufficient evidence of, as determined by the FDA, a causal
relationship between the drug and the occurrence of the adverse event. See infra.

5

affect a doctor’s prescribing decisions or be “potentially fatal.” 71 Fed. Reg. 3922-01,
3946; FDA, Guidance for Industry: Warnings and Precautions, Contraindications,
and Boxed Warning Sections of Labeling for Human Prescription Drug and Biological
Products – Content and Format, at 3 (Oct. 2011). On the other hand, the threshold
for warning of an adverse event in the Adverse Reactions section is comparatively
lower: “some basis to believe there is a causal relationship between the drug and the
occurrence of the adverse event.” 21 C.F.R. § 201.57(c)(7).
New information about a drug may require changing its label. 21 U.S.C. §§
314.80(c), 314.81(b)(2)(i). A drug manufacturer may change its label in one of two
ways. More commonly, it may seek advance permission from the FDA through a Prior
Approval Supplement Application (“PAS”). 21 C.F.R. § 314.70(b). Alternatively, it
may change a label immediately and unilaterally through a Changes Being Effected
Application (“CBE”) to reflect “newly acquired information” about “evidence of a
casual association between the drug and a risk of harm.” Merck, 139 S.Ct. at 1673
(quotations omitted); 21 U.S.C. § 314.70(c)(6)(iii)(A); 21 C.F.R. § 314.3(b) (defining
“[n]ewly acquired information” to mean, inter alia, risks not previously known or
previously underestimated). Whatever method a manufacturer chooses, it must meet
the causal thresholds described above, and significantly, the FDA retains authority
to reject even a CBE amendment if there is insufficient evidence of a link between
the drug and the adverse event. 73 Fed. Reg. 2848, 2851; 21 C.F.R. §
314.70(c)(6)(iii)(A) (providing that the FDA will approve a label change only if “the

6

evidence of a causal association satisfies the standard for inclusion in the labeling”);
id. §§ 314.125(b)(6), (b)(8).
Because of the availability of the CBE process, “a drug manufacturer will not
ordinarily be able to show that there is an actual conflict between state and federal
law such that it was impossible to comply with both.” Merck, 139 S.Ct. at 1679. At
the same time, the FDA will not approve a warning simply out of an abundance of
caution whenever a manufacturer posits an association between a drug and an
adverse event. As the FDA has long recognized, “[e]xaggeration of risk, or inclusion
of speculative or hypothetical risks, could discourage appropriate use of a beneficial
drug.” 73 Fed. Reg. 2848, 2851. Because “labeling that includes theoretical hazards
not well-grounded in scientific evidence can cause meaningful risk information to lose
its significance,” the FDA prohibits “a change to labeling [, either through the PAS or
CBE process,] to add a [Precautions] warning in the absence of [at least] reasonable
evidence of an association.” Id. This represents a more conservative approach than
state tort law, which generally incentivizes a manufacturer to warn about every
conceivable hazard to limit liability. See Wyeth v. Levine, 555 U.S. 555, 557 (2009).
Finally, the FDA has an independent obligation to ensure that drug labels
reflect new risks. 21 U.S.C. § 355(o)(4)(A) (providing that, if the agency “becomes
aware of new information, including any new safety information,” which “should be
included in the labeling of the drug,” it “shall promptly notify the [manufacturer]”).
Indeed, Congress has “reaffirmed the manufacturer’s . . . ultimate responsibility for
its label,” including when it “granted the FDA th[e] authority” to mandate label

7

changes in 2007. 3 Wyeth, 555 U.S. at 571; 21 U.S.C. § 355(o)(4)(I). If new safety
information arises regarding a particular risk, the manufacturer, similarly,
maintains “a duty to provide a warning that adequately describe[s] that risk,” Wyeth,
555 U.S. at 571, and “bears responsibility for the content of its label at all times,”
Merck, 139 S.Ct. at 1677 (explaining that this has “remained a central premise of
federal drug regulation”), regardless of whether the FDA takes parallel action.
C.

The Fosamax Label History

When the FDA approved Fosamax in 1995, the label did not warn of a risk of
the adverse event Plaintiffs allege here, i.e., atypical femoral fractures. Fosamax,
852 F. 3d at 271, 274-75. However, Merck was “aware of at least a theoretical risk” of
such particular fractures as early as 1992, during clinical trials, and brought it to the
FDA’s attention at that time. Merck, 139 S. Ct. at 1674 (informing the FDA that
“antiresorptive agents may inhibit microdamage repair by preventing . . . bone
resorption at the sites of microdamage”). More evidence came to light after 1995,
when “Merck began receiving adverse event reports from the medical community

3
FDA regulations also require a New Drug Application (“NDA”) to disclose all
“pertinent” safety information. 21 C.F.R. § 314.50 (requiring “reports of all investigations of
the drug product sponsored by the applicant, and all other information about the drug
pertinent to an evaluation of the NDA that is received or otherwise obtained by the applicant
from any source”); id. § 314.50(d)(5)(vi)(a) (requiring “an integrated summary of all available
information about the safety of the drug product, including pertinent animal data[ and]
demonstrated or potential adverse effects of the drug”); id. § 312.50 (stating that “[s]ponsors
are responsible for . . . providing [investigators] with the information they need to conduct an
investigation properly . . . and ensuring that [the] FDA and all participating investigators
are promptly informed of significant new adverse effects or risks with respect to the drug”).
The FDA approval process is “onerous and lengthy.” Mut. Pharm. Co. v. Bartlett, 570 U.S.
472, 476 (2013).

8

indicating that long-term Fosamax users were suffering atypical femoral fractures.” 4
Merck, 139 S.Ct. at 1674. Based on its own analysis of these increasing reports, in
2005, Merck preliminarily concluded that there was a statistically significant
increase in the incidence of atypical femoral fractures among Fosamax users. Pl. Br.,
Ex. 8, at A1272-73.
Merck also “began to see numerous scholarly articles and case studies
documenting possible connections between long-term Fosamax use and atypical
femoral fractures.” Merck, 139 S.Ct. at 1674. However, none of these studies
concluded that Fosamax actually caused atypical femoral fractures, or even that they
were definitively associated with Fosamax use. Fosamax, 852 F.3d at 275 (citing
A1258) (stating that Fosamax may potentially increase the risk of such fractures); id.
(citing A1237) (stating that Fosamax may be associated with such fractures; id.
(citing A1243) (stating that certain findings raised the possibility that Fosamax may
lead to such fractures). Still, Merck forwarded them to the FDA. Fosamax, 862 F.3d
at 275.
In March 2008, Merck submitted to the FDA a 165-page periodic safety update,
the twenty-ninth of its kind, with thirty pages dedicated to “recent publications”
“implicat[ing] a link between prolonged bisphosphonate therapy and atypical lowenergy non-vertebral fractures,” and “relat[ing] these findings to severely suppressed
bone turnover that may develop during long-term” use of Fosamax. Def. Rep. Br., Ex.
4
For example, in 2002, Merck received a report from a doctor who said that his hospital
called atypical femoral fractures the “Fosamax Fracture” because “100% of patients in his
practice who have experienced femoral fractures (without being hit by a taxicab), were taking
Fosamax . . . for over 5 years.” Merck, 139 S.Ct. at 1674 (quotations and citations omitted).

9

14, at A2597. Later that month, Merck sent the FDA a letter from the New England
Journal of Medicine describing “a potential link between [bisphosphonate] use and
low-energy fractures of the femur.” Id., Ex. 13. The FDA, in turn, informed Merck in
June 2008, that it was “aware of reports regarding the occurrence of subtrochanteric
hip fractures in patients using bisphosphonates” and was “concerned about this
developing safety signal.” Pl. Br., Ex. 10, at A1145. The Agency asked Merck for
additional data and investigations by July 2008, and Merck complied.
In September 2008, while its data was pending review, Merck submitted to the
FDA a PAS, i.e., application to enlarge the warning label, to amend the Adverse
Reactions section of the Fosamax label with a warning about “low-energy femoral
shaft fractures,” id., Ex. 38, at A1349, and to cross-reference a longer discussion in
the Precautions section. Merck, 139 S.Ct. at 1674. Specifically, Merck proposed
adding the following language to the Precautions section:
Low-Energy Femoral Shaft Fracture
Low-energy fractures of the subtrochanteric and proximal femoral shaft
have been reported in a small number of bisphosphonate-treated
patients. Some were stress fractures (also known as insufficiency
fractures) occurring in the absence of trauma. Some patients
experienced prodromal pain in the affected area, often associated with
imaging features of stress fracture, weeks to months before a complete
fracture occurred. The number of reports of this condition is very low,
and stress fractures with similar clinical features also have occurred in
patients not treated with bisphosphonates. Patients with suspected
stress fractures should be evaluated, including evaluation for known
causes and risk factors (e.g., vitamin D deficiency, malabsorption,
glucocorticoid use, previous stress fracture, lower extremity arthritis or
fracture, extreme or increased exercise, diabetes mellitus, chronic
alcohol abuse), and receive appropriate orthopedic care. Interruption of
bisphosphonate therapy in patients with stress fractures should be

10

considered, pending evaluation of the patient, based on individual
benefit/risk assessment.
Pl. Br., Ex. 38 at A1371. 5
As part of its PAS to the FDA, Merck submitted a lengthy analysis of femoral
fractures in Fosamax users, cited to nine articles on such cases, and summarized the
findings in a clinical overview. Merck opined that, although “[i]t is not possible with
the present data to establish whether” Fosamax “increases the risk of . . . low-energy
subtrochanteric and/or proximal shaft fractures,” because they tended to arise
alongside Fosamax use, it is “important to include an appropriate statement” about
them in the drug’s precautions section. Id., Ex. 38, at A1349-51.
In April 2009, Merck discussed its pending PAS with FDA official, Dr. Scott
Monroe. According to Merck’s notes, Dr. Monroe expressed that, while the FDA could
agree to additional language in the Adverse Reactions section, it likely would not
approve similar language in the Precautions section. Pl. Br., Ex. 33, at A1970-71. Dr.
Monroe advised that the FDA would likely “approach the issue of a precaution from
the [perspective] of all bisphosphonates [from various drug manufacturers]” and was
“working with the Office of Safety and Epidemiology [“OSE”] to do so.” Id. But,
because “the conflicting nature of the literature does not provide a clear path forward,

Based on a review of scientific studies in the record, including the FDA’s September
2010 Task Force Report, as mentioned supra, I note that “low-energy femoral shaft fractures”
are the same as “atypical femoral fractures.” See Pl. Br., Ex. 2, at A1152. In layman’s terms,
“atypical femoral fractures” are a “rare type of complete, low-energy fracture [that] affects
the thigh bone.” Merck, 139 S.Ct. 1674. The low-energy component, critical to both terms,
generally means that the fracture was caused by a slip, trip, or fall from standing height or
less. See Pl. Br., Ex. 2. At A1152. Thus, low-energy fractures are typically caused by
mechanical forces that would not ordinarily result in fracture, while high-energy fractures,
on the other hand, are generally associated with a more focused and substantial trauma.
5

11

. . . more time will be need[ed] for FDA to formulate a formal opinion on the issue of
a precaution.” Id. In Dr. Monroe’s view, Merck’s “elevation” of the warning to a
Precaution was “prolonging” approval of any amendment to the label. Id.
Later that month, an FDA official emailed Merck that the FDA was not
prepared to include language about low-energy femoral fractures in the Precautions
section, and “could . . . only” “approve[]” such a warning “in the adverse events section
of the label.” Def. Br., Ex. 3, at A1498. The official asked Merck to “hold off on the
[Precautions] language” so that drug evaluators could “work with [the Office of
Surveillance and Epidemiology] and Merck to decide on . . . atypical fracture
language, if it is warranted.” Id. The next month, in May 2009, officially responding
to Merck’s PAS, Dr. Monroe drafted a CRL which stated that the FDA approved a
warning in the Adverse Reactions section, subject to some rewording, but rejected one
in the Precautions section. Then, the FDA explained:
We have completed the review of your [PAS], as amended, and have
determined that we cannot approve these applications in their present
form. We have described below our reasons for this action and our
recommendation to address this issue.
1. While the Division agrees that atypical and subtrochanteric fractures
should be added to the ADVERSE REACTIONS, Post-Marketing
Experience subsections of the [Fosamax] labels, your justification for
the proposed PRECAUTIONS section language is inadequate.
Identification of “stress fractures” may not be clearly related to the
atypical subtrochanteric fractures that have been reported in the
literature. Discussion of the risk factors for stress fractures is not
warranted and is not adequately supported by the available literature
and post-marketing adverse event reporting.
Def. Br., Ex. 2, at A1500-01.

12

On the same day that the FDA sent the CRL, Merck “asked the [Agency]” for
a “teleconference to discuss what [Precautions language] may be acceptable.” Pl. Br.,
Ex. 13. A few weeks later, undeterred, Merck again asked “for a meeting . . . to discuss
the issues that were raised in the [CRL] to Merck’s proposed text to the Precautions
section.” Id., Ex. 14. Merck also asked to “leave the previous PAS active to permit
further discussions with the agency.” Id., Ex. 15. The FDA “informed [Merck] that
the proposal was not in-line with Dr. Monroe’s request that all deficiencies need to be
addressed to start a new review cycle,” and any meeting must be formally requested.
Id. Merck maintained that “[atypical] fractures should [still] be described in the
Precautions section,” and suggested “broach[ing]” that topic in an unrelated
teleconference the following day, to which the FDA responded it might be “possible,”
albeit “not the objective of the meeting.” Id.
Pursuant to FDA regulations, within one year of the CRL, Merck had to
“resubmit” its application “addressing all deficiencies identified” in the CRL,
withdraw it, or request a hearing, after which “the agency will either approve” or
“refuse” the label change. 21 C.F.R. § 314.110(b). In July 2009, Merck elected to
withdraw, Def. Br., Ex. 4, at A2961, change the Adverse Reactions section through a
CBE amendment, as recommended by the FDA, id. at A2963-64, and leave the
Precautions section as-is. But, Merck did not do so without reiterating, once more, its
desire to add a Precautions warning. Id.
Unwavering, in March 2010, after reviewing the data submitted by Merck (and
other manufacturers), the FDA issued a Drug Safety Announcement reiterating that

13

there was not yet “a clear connection between bisphosphonate use and a risk of
atypical subtrochanteric femur fractures.” Def. Br., Ex. 5, at A1508-09. The FDA,
however, announced that it would work with an outside Task Force, which included
various experts in different agencies, to gather additional information. Id. In
September 2010, the Task Force found that “there is evidence of a relationship
between long-term [bisphosphonate] use and a specific type of subtrochanteric and
femoral shaft fracture,” although not enough to establish causation. Pl. Br., Ex. 2, at
A1167. The FDA responded with another Drug Safety Announcement stating that,
“[a]lthough it is not clear if bisphosphonates are the cause [of fractures], these
unusual femur fractures have been identified in patients taking [such] drugs.” Def.
Br., Ex. 9, at A1512. The FDA then “assembled and [reviewed] all long term data
available on the products, as well as all safety reports,” and promised to “keep the
public informed of additional findings.” Id.
In October 2010, more than a year after the FDA sent Merck its CRL, the FDA,
after completing its analysis, finally concluded that “atypical fractures may be related
to long-term . . . bisphosphonate use,” and announced that it would require all
bisphosphonate manufacturers to add information on that risk to the Precautions
sections of their labels. Pl. Br., Ex. 19, at A1118. In a media call accompanying the
announcement, the FDA’s Deputy Director of the Office of New Drugs stated that the
Task Force Report made the Agency “confident” that atypical femur fractures are
“potentially more closely related to” long-term use of bisphosphonates “than [we]
previously had evidence for.” Def. Br., Ex. 6, at A1396. The FDA wrote to Merck that

14

day to mandate a label change to Fosamax. Def. Br., Ex. 7, at A1516-17. Specifically,
the FDA provided language for a warning in the Precautions section:
Atypical Subtrochanteric and Diaphyseal Femoral Fractures:
Atypical, low-energy, or low trauma fractures of the femoral shaft have
been reported in bisphosphonate-treated patients. These fractures can
occur anywhere in the femoral shaft from just below the lesser
trochanter to above the supracondylar flare and are transverse or short
oblique in orientation without evidence of comminution. Causality has
not been established as these fractures also occur in osteoporotic
patients who have not been treated with bisphosphonates.
Atypical femur fractures most commonly occur with minimal or no
impact to the affected area. They may be bilateral and many patients
report prodromal pain in the affected area, usually presenting as dull,
aching thigh pain, weeks to months before a complete fracture occurs. A
number of reports note that patients were also receiving treatment with
glucocorticoids (e.g. prednisone) at the time of fracture.
Any patient with a history of bisphosphonate exposure who presents
with thigh or groin pain should be suspected of having an atypical
fracture and should be evaluated to rule out a femur fracture. Subjects
presenting with an atypical fracture should also be assessed for
symptoms and signs of fracture in the contralateral limb. Interruption
of bisphosphonate therapy should be considered, pending a risk/benefit
assessment, on an individual basis.
Id.
In response, Defendant proposed revised language that, once again, referred
to the risk of “stress fractures.” Pl. Br., Ex. 21, at A1556-57. But, the FDA rejected
that language, explaining that “the term ‘stress fracture’ was considered and was not
accepted” because, “for most practitioners, the term ‘stress fracture’ represents a
minor fracture and this would contradict the seriousness of the atypical femoral
fractures associated with bisphosphonate use.” Id. at A1540. In January 2011, Merck

15

Merck argued, in response, that federal law preempted Plaintiffs’ claims—
specifically, the May 2009 CRL rejecting Merck’s proposed label change. 6
Following a bellwether trial, Judge Pisano agreed with Merck, and granted
summary judgment in all cases. OTSC Opinion, 2014 WL 1266994, at *17 (D.N.J.
Mar. 22, 2014); Glynn, 951 F. Supp. 3d at 701, 703-04. In particular, Judge Pisano
found, “the fact that the FDA never required [Merck] to submit new language or
change the label [after rejecting its proposed warning] demonstrates that the FDA
did not think that the label should have been changed at that time,” and there was
“clear evidence that the FDA would have rejected a stronger Precautions warning
because the FDA did reject a stronger Precautions warning.” OTSC Opinion, 2014
WL 1266994, at *16 (emphasis in original). Indeed, Judge Pisano explained that
pursuant to the Supreme Court’s decision in Wyeth, a state law failure-to-warn claim
is preempted if there is “clear evidence” that the FDA “would not have approved” any
and all warnings.
Plaintiffs appealed the decision to the Third Circuit, which vacated Judge
Pisano’s decision. Fosamax, 852 F.3d 268. While recognizing that Wyeth controls the
analysis, the Court of Appeals reasoned that “[t]he term ‘clear evidence’ . . . does not
refer directly to the type of facts that a manufacturer must show, or to the
circumstances in which preemption will be appropriate.” Id. at 285. “Rather, it
specifies how difficult it will be for the manufacturer to convince the factfinder that
6
In 2011, the Judicial Panel on Multidistrict Litigation consolidated these cases, which
once exceeded 1,000 cases, for pre-trial administration in a multi-district litigation (“MDL”)
in the District of New Jersey. In re: Fosamax (Alendronate Sodium) Prods. Liab. Litig. (No.
II), 787 F. Supp. 2d 1355 (J.P.M.L. 2011).

17

the FDA would have rejected a proposed label change.” Id. And, the court determined
that the factfinder must be a jury not a judge. In that regard, the circuit court devised
a novel standard: “for a defendant to establish a preemption defense under Wyeth,
the [jury] must conclude that it is highly probable that the FDA would not have
approved a change to the drug’s label.” Id. at 286.
Accepting Merck’s petition for certiorari, the Supreme Court vacated the Third
Circuit’s opinion and judgment, holding that preemption must be decided by “a judge,
not the jury,” who, in turn, “must simply ask himself or herself whether the relevant
federal and state laws irreconcilably conflict.” Merck, 139 S.Ct. at 1676, 1679-80
(quotations and citation omitted). The Court also “elaborate[d] Wyeth’s” clear
evidence standard “along the way.” Id. It explained that “[c]lear evidence” exists
where a drug manufacturer “show[s] that it fully informed the FDA of the
justifications for the warning required by state law and that the FDA, in turn,
informed the drug manufacturer that the FDA would not approve changing the drug’s
label to include that warning.” Id. at 1678. This will not “ordinarily” be the case. Id.
at 1679. Moreover, “whatever the means the FDA uses to exercise its authority, those
means must lie within the scope of the authority Congress has lawfully delegated,”
an “obvious point” which the Court reiterated even though “[t]he question of
disapproval method is not now before [the Court].” 7 Id. at 1679-80.

7
Justice Thomas wrote separately to explain his “understanding of the relevant preemption principles and how they apply to this case.” Merck, 139 S.Ct. at 1681 (Thomas, J.,
concurring). Justice Thomas would not find preemption here because, in his view, nothing
prevented Merck from using the CBE process to unilaterally add a warning to the
Precautions section, even though the FDA retains the authority to reject a CBE amendment
if it lacks causation. Id. at 1683. Further, according to Justice Thomas, even if Merck believed

18

The Supreme Court remanded to the Third Circuit with instructions “to
consider fully the standards we have described.” Id. at 1680-81. Rather than deciding
the issue, the Third Circuit remanded to this Court “to determine in the first instance
whether the plaintiffs’ state law claims are preempted by federal law.” Order at 1,
No. 14-1900 (3d Cir. Nov. 25, 2019). The Third Circuit also instructed this Court “to
determine the effect of the FDA’s [CRL] and other communications with Merck on
the issue of whether such agency actions are sufficient to give rise to preemption.” Id.
E.

The Parties’ Arguments on Remand

The issue on remand is the same as it was eight years ago: whether the CRL
“prohibited [Merck] from adding any and all warnings to the drug label that would
satisfy state law.” Merck, 139 S.Ct. at 1678. Plaintiffs answer in the negative, and
advance several arguments, some of which merely restate their prior positions. First,
they reiterate that Merck did not fully inform the FDA of the risks of Fosamax use.
Pl. Br., at 30-35. Second, relying on Justice Thomas’ concurrence, they argue for the

that the FDA would have ultimately rejected a CBE amendment, that “hypothetical” would
not constitute “[l]aw with pre-emptive effect,” because “the possibility of impossibility is not
enough.” Id. Justice Thomas also rejected the preemptive effect of a CRL to the extent that
such a letter is not a final agency action. Id. at 1682. In response, Justice Alito wrote
separately to ensure that the Court’s “discussion of the law and the facts” are not “misleading
on remand.” Id. at 1684 (Alito, J., concurring in the judgment). Chief Justice Roberts and
Justice Kavanaugh joined his opinion. Justice Alito explained that “a statutory provision
enacted after the events in [Wyeth] [ ] may have an important bearing” on this case, namely
21 U.S.C. § 355(o)(4)(A), which requires the FDA to initiate a label change under certain
circumstances, but does not require it “to communicate to the relevant drug manufacturer
that a label change is unwarranted; instead, the FDA could simply consider the new
information and decide not to act.” Id. Justice Alito then detailed the back and forth between
Merck and the FDA to counter the majority’s “one-sided account,” stating “for years the FDA
was: aware of this issue, communicating with drug manufacturers, studying all relevant
information, and instructing healthcare professionals and patients alike to continue to use
Fosamax as directed.” Id. at 1685-86.

19

first time during this litigation that the CRL does not carry preemptive effect because
it is not a final agency action, id. at 12-15, 27; however, they primarily dispute the
meaning and scope of the CRL. They begin by arguing that Merck did not propose a
warning that would have been adequate under state law in the first place. According
to Plaintiffs, Merck’s PAS emphasized “garden variety” stress fractures, which are
scientifically different from the more serious atypical femoral fractures. Id. at 16-19,
24 n.4. Because of this focus, Plaintiffs posit that the FDA could not have rejected a
warning about atypical femoral fractures at all, but only one about commonplace
stress fractures. So construed, Plaintiffs advance that the CRL does not constitute
“clear evidence” that the FDA would have prohibited any and all warnings to
Fosamax, despite the Agency’s other communications from the same time period. Id.
at 24-30.
Merck maintains that it has always fully informed the FDA of the risks of
Fosamax, particularly the risk of developing atypical femoral fractures. Def. Br., at
17-20; Def. Rep. Br., at 1-4. Merck also calls Plaintiffs’ position that the CRL lacks
preemptive power “idiosyncratic” and “unsupported” by law. Def. Br., at 28-30; Def.
Rep. Br., at 11-12. Further, as to the meaning and scope of the CRL, Merck argues
that its proposed warning was “perfectly” adequate under state law, see Def. Br., at
5-6, 14, 20-24, and the FDA rejected it for insufficient causal evidence linking
bisphosphonate use to atypical femoral fractures, not because of the garden variety
“stress fracture” language on which Plaintiffs improperly focus. Def. Br., at 23-27;
Def. Rep. Br., at 10-11. As explained by Merck, to the extent that the basis for the

20

CRL was the FDA’s skepticism of the underlying science regarding causal connection,
there is necessarily clear evidence that the FDA would have rejected any and all
changes to the Fosamax label. See 21 C.F.R. § 201.57(c)(6)(iii) (requiring “reasonable
evidence of a causal association” to add a Precautions warning). Finally, even if the
terms of the CRL themselves are unclear, Merck maintains that the letter constitutes
clear evidence when construed in light of the FDA’s other communications from
around the same time. Def. Br., at 8-9, 21-23, 26; Def. Rep. Br., at 12-13, 15.
In short, Merck submits that the CRL conveyed that the FDA would not have
approved any warning about atypical femoral fractures because of its then-existing
perspective on the causal connection between such fractures and Fosamax use.
Plaintiffs, on the other hand, take the position that the FDA had conveyed a far more
limited message in the CRL: Merck’s particular warning, as worded, was
unacceptable, but the FDA might have approved different language had Merck
proposed it through a revised PAS or a CBE amendment.
STANDARD OF REVIEW 8
Summary judgment is appropriate “if the pleadings, depositions, answers to
interrogatories, and admissions on file, together with the affidavits, if any, show that

While the parties’ briefing does not discuss the legal standard the Court should apply
on remand, they agreed that Rule 56 was the proper framework by which Judge Pisano
resolved the dispositive issues presented by Defendant’s preemption defense in the first
instance. Fosamax, 852 F.3d at 281 (“Although both sides disputed the propriety of the showcause procedure and the substance of Merck’s preemption arguments, the parties and the
District Court all agreed that Federal Rule of Civil Procedure 56 ‘provides the exclusive
mechanism by which the Court can resolve the dispositive issues presented by Merck's
preemption defense before trial(s).’”). Accordingly, because the Third Circuit remanded that
very issue to me, I will apply that same standard, here.
8

21

there is no genuine issue as to any material fact and that the moving party is entitled
to a judgment as a matter of law.” Fed. R. Civ. P. 56(c). A factual dispute is genuine
only if there is “a sufficient evidentiary basis on which a reasonable [factfinder] could
find for the non-moving party,” and it is material only if it has the ability to “affect
the outcome of the suit under governing law.” Kaucher v. Cty. of Bucks, 455 F.3d 418,
423 (3d Cir. 2006); see also Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986).
Disputes over irrelevant or unnecessary facts will not preclude a grant of summary
judgment. Anderson, 477 U.S. at 248. “In considering a motion for summary
judgment, a district court may not make credibility determinations or engage in any
weighing of the evidence; instead, the non-moving party's evidence ‘is to be believed
and all justifiable inferences are to be drawn in his favor.’” Marino v. Indus. Crating
Co., 358 F.3d 241, 247 (3d Cir. 2004) (quoting Anderson, 477 U.S. at 255); see also
Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 587 (1986); Curley
v. Klem, 298 F.3d 271, 276-77 (3d Cir. 2002).
The party moving for summary judgment has the initial burden of showing the
basis for its motion. Celotex Corp. v. Catrett, 477 U.S. 317, 323 (1986). “If the moving
party will bear the burden of persuasion at trial, that party must support its motion
with credible evidence ... that would entitle it to a directed verdict if not controverted
at trial.” Id. at 331. On the other hand, if the burden of persuasion at trial would be
on the nonmoving party, the party moving for summary judgment may satisfy Rule
56’s burden of production by either (1) “submit[ting] affirmative evidence that
negates an essential element of the nonmoving party’s claim” or (2) demonstrating

22

“that the nonmoving party’s evidence is insufficient to establish an essential element
of the nonmoving party's claim.” Id. Once the movant adequately supports its motion
pursuant to Rule 56(c), the burden shifts to the nonmoving party to “go beyond the
pleadings and by her own affidavits, or by the depositions, answers to interrogatories,
and admissions on file, designate specific facts showing that there is a genuine issue
for trial.” Id. at 324; see also Matsushita, 475 U.S. at 586; Ridgewood Bd. of Ed. v.
Stokley, 172 F.3d 238, 252 (3d Cir. 1999). In deciding the merits of a party’s motion
for summary judgment, the court’s role is not to evaluate the evidence and decide the
truth of the matter, but to determine whether there is a genuine issue for trial.
Anderson, 477 U.S. at 249. Credibility determinations are the province of the
factfinder. Big Apple BMW, Inc. v. BMW of N. Am., Inc., 974 F.2d 1358, 1363 (3d Cir.
1992). There can be “no genuine issue as to any material fact,” however, if a party
fails “to make a showing sufficient to establish the existence of an element essential
to that party’s case, and on which that party will bear the burden of proof at trial.”
Celotex, 477 U.S. at 322-23. “[A] complete failure of proof concerning an essential
element of the nonmoving party’s case necessarily renders all other facts immaterial.”
Id. at 323; Katz v. Aetna Cas. & Sur. Co., 972 F.2d 53, 55 (3d Cir. 1992).
DISCUSSION
I. Scope of Remand
The Third Circuit’s mandate to this Court is clear: to determine whether
Plaintiffs’ state law failure-to-warn claims are preempted by federal law under the
standards described by the Supreme Court in Merck. Order at 1, No. 14-1900 (3d

23

Cir. Nov. 25, 2019). In considering that question, the Third Circuit also instructed
this Court to determine whether the FDA’s CRL and other communications with
Defendant are sufficient to give rise to preemption. Id.
As a preliminary matter, I note that on appeal, the Third Circuit vacated in
full, Judge Pisano’s underlying decision granting summary judgment in favor of
Defendant, and the Supreme Court remanded with instructions to “consider fully”
its elaboration of Wyeth’s clear evidence standard. Although I will conduct a de novo
review of the legal issues and record, that does not necessarily mean, however, that
Judge Pisano’s factual findings will be ignored. In fact, my decision, here, will refer
to Judge Pisano’s opinion—at least as it relates to certain facts that are generally
not in dispute. Indeed, Judge Pisano held a full trial on the merits, heard expert
testimony, made numerous factual findings related to the narrow legal question on
appeal in Merck, decided the preemption inquiry, and unsurprisingly, the evidence
before me is virtually identical to the evidence presented then. To be certain, in
2013, Judge Pisano answered the preemption question posed to the Court on
remand, here, consistent with the standard set forth in Wyeth—a standard that the
Supreme Court did not overrule, but merely clarified and expounded upon in Merck.
See infra. Indeed, Merck decided the narrow question of whether a jury or judge
determines preemption—agreeing with Judge Pisano that it was a question for a
judge. That issue constitutes new law, which I take as the law of this case now.
Bankers Trust Co. v. Bethlehem Steel Corp., 761 F.2d 943, 950 (3d Cir. 1985). But,
what remains is exactly what Judge Pisano had to decide eight years ago: assess “in

24

the first instance” whether “the FDA would have rejected a change,” considering any
relevant factual disputes along the way. For these reasons, I will refer to Judge
Pisano’s factual findings where appropriate.
II. Preemption
“A fundamental principle of the Constitution is that Congress has the power to
preempt state law.” Crosby v. Nat’l Foreign Trade Council, 530 U.S. 363, 372 (2000).
“Preemption follows automatically by the operation of the Supremacy Clause,” Wyeth,
555 U.S. at 624 (Alito, J., concurring in the judgment), which “invalidates state laws
that interfere with, or are contrary to, federal law.” Hillsborough County, Florida v.
Automated Medical Laboratories, Inc., 471 U.S. 707, 712 (1985) (quotations omitted).
Federal law can preempt state law in three ways: (1) express preemption, (2) field
preemption, and (3) conflict preemption. 9 Farina v. Nokia Inc., 625 F.3d 97, 115 (3d
Cir. 2010). Both parties agree that the issue in this case is conflict preemption, which
exists “where it is impossible for a private party to comply with both state and federal
requirements.” Sprietsma v. Mercury Marine, a Div. of Brunswick Corp., 537 U.S. 51,
64 (2002) (quotations omitted); PLIVA, Inc. v. Mensing, 564 U.S. 604, 620 (2011) (“The
question . . . is whether the private party could independently do under federal law

Federal regulations with the force of law preempt state laws in the same manner as
federal statutes. See, e.g., Geier v. American Honda Motor Co., 529 U.S. 861 (2000); Fellner
v. Tri-Union Seafoods, L.L.C., 539 F.3d 237, 243 (3d Cir. 2008) (“Where Congress has
delegated the authority to regulate a particular field to an administrative agency, the
agency’s regulations issued pursuant to that authority have no less preemptive effect than
federal statutes, assuming those regulations are a valid exercise of the agency’s delegated
authority.”). There is no dispute here that preemption, if appropriate, applies to all forms of
state law, including civil actions based on state law, such as Plaintiffs’ failure-to-warn claims.
Holk v. Snapple Beverage Corp., 575 F.3d 329, 331 (3d Cir. 2009).
9

25

what state law requires of it.”); Klotz v. Celentano Stadtmauer & Walentowicz LLP,
991 F.3d 458, 463 (3d Cir. 2021).
Conflict, or impossibility, preemption “is a demanding defense” in the drug
labeling context. Wyeth, 555 U.S. at 573. Essentially, a defendant must show that it
could not have unilaterally changed its label in any way to add the warning required
by state law. Id. at 569-71; Sikkelee v. Precision Airmotive Corp., 822 F.3d 680, 703704 (3d Cir. 2016); Knight v. Boehringer Ingelheim Pharmaceuticals, Inc., 984 F.3d
329, 337 (4th Cir. 2021) (“A state law challenge to FDA-approved warnings, including
a tort action under state law, can [ ] proceed only when the defendant had the
unilateral ability to change that labeling; otherwise, the claim is preempted.”).
The “possibility of impossibility [is] not enough” to establish preemption,
PLIVA, 564 U.S. at 624 n.8 (quotations omitted); Rice v. Norman Williams Co., 458
U.S. 654, 659 (1982) (rejecting “hypothetical” impossibility), and there is a
“presumption against preemption,” Wyeth, 555 U.S. at 595 n.3, which applies with
special force in fields involving traditional state police powers. Medtronic, Inc., v.
Lohr, 518 U.S. 470, 485 (1996) (“In all pre-emption cases, and particularly in those in
which Congress has ‘legislated . . . in a field which the States have traditionally
occupied,’ . . . [courts] ‘start with the assumption that the historic police powers of the
States were not to be superseded by the Federal Act unless that was the clear and
manifest purpose of Congress.’”). On the other hand, “the possibility of possibility” is
not sufficient to defeat preemption. PLIVA, 564 U.S. at 624 n.8.

26

Rather, under Wyeth, if there is “clear evidence that the FDA would not have
approved a change” to a drug’s label, then it is impossible to comply with both federal
and state law, and a plaintiff’s failure-to-warn claims are preempted. 555 U.S. at 571.
To establish clear evidence, a drug manufacturer must “show that it fully informed
the FDA of the justifications for the warning required by state law and that the FDA,
in turn, informed the drug manufacturer that the FDA would not approve changing
the drug’s label to include that warning.” Merck, 139 S.Ct. at 1678.
A.

Merck Did Not Repudiate Wyeth 10

At the outset, Plaintiffs contend that Merck repudiates Wyeth’s “premise that
a manufacturer can show preemption by arguing that the FDA would have rejected
a warning that it did not actually reject.” Pl. Br., at 13-14 (emphasis in original). In
Plaintiffs’ view, impossibility preemption now “requires an affirmative showing that
the FDA took ‘action[]’ to ‘prohibit[] the drug manufacturer from adding any and all
warnings to the drug label that would satisfy state law.’” Pl. Br., at 14 (quoting Merck,
139 S.Ct. at 1676, 1678).
Plaintiffs’ position has some facial appeal, but it is ultimately specious. In
Wyeth, the phrase “would not have approved” implies that a drug manufacturer may
prove preemption without showing that it ever proposed or pursued a label change.
Plaintiffs argue, however, that Merck’s phrasing of the law should be read to mean

10
In Wyeth, the Supreme Court rejected a drug manufacturer’s preemption defense after
an antinausea drug caused a patient to develop gangrene. Notably, there was no prior agency
action in that case. The question was whether the FDA would have rejected a CBE
amendment had the manufacturer attempted to pursue one. However, Wyeth does not
instruct how this Court should interpret the meaning of an actual FDA decision on labeling,
such as the CRL here that rejected Merck’s proposed warning, which is the crux of this case.

27

that a manufacturer must have actually requested a label change that the FDA then
expressly rejected. 11 Specifically, Plaintiffs rely exclusively on the Supreme Court’s
finding that to establish preemption, a manufacturer “is required to show that it fully
informed the FDA of the justifications for the warning required by state law and that
the FDA, in turn, informed the drug manufacturer that the FDA would not approve
changing the drug’s label to include that warning.” Merck, 139 S.Ct. at 1678.
According to Plaintiffs, “anything less is insufficient.” Pl. Br., at 14.
The Seventh Circuit has declined to view Merck in that manner, and I find that
court’s reasoning persuasive. Dolin v. GlaxoSmithKline, 951 F.3d 882, 890-91 (7th
Cir. 2020). Indeed, the court, there, observed, in the context of a Rule 60(b) motion,
that Merck “explicitly grounded its analysis in the Court’s holdings in Wyeth . . . .
began by citing the Wyeth ‘clear evidence’ standard[,] and formulated the question for
decision in terms of the Wyeth framework,” and further, that Merck uses “the
language of ordinary evolution” rather than ‘reversal and overruling.” The Tenth
Circuit ruled similarly. Cerveny v. Aventis, Inc., 783 Fed. App’x. 804 n.8 (10th Cir.
2019) (dismissing, in the context of a Rule 28(j) letter, the contention that “only
labeling changes sought by the manufacturer can lead to preemption.”)
The Third Circuit also had an opportunity to reinterpret Wyeth in the manner
proposed by Plaintiffs, but chose not do so in light of the facts before it. In re Avandia

Of course, this is precisely the factual scenario of this case; that is, Defendant claims
that the CRL issued by the FDA expressly rejected Defendant’s proposed warning regarding
atypical femoral fractures. And, the primary dispute between the parties is whether the CRL
could be so interpreted as to have rejected Defendant’s proposed warning based on the causal
connection between the use of Fosamax and atypical femoral fractures.
11

28

Marketing, Sales and Prod. Liab. Litig., 945 F.3d 749, 759 (3d Cir. 2019) (stating that
it “need not speculate regarding the possibility that the FDA would have rejected the
proposed warning” because the FDA in fact “ordered” one) (emphasis in original). In
dozens of district court cases since, not one court has interpreted Merck to establish
a new standard for impossibility preemption requiring actual agency or manufacturer
action. See, e.g., In re Incretin-Based Therapies Prod. Liab. Litig., No. 13-2452, 2021
WL 880316 (S.D. Cal. Mar. 9, 2021) (“Plaintiffs also contend that [Merck] limited
preemption to cases where the manufacturer has proposed a label change. The Court,
however, does not read [Merck] so narrowly. Rather, the Court finds that [Merck]
simply reiterated the lesson in Wyeth that the availability of the CBE label change
process makes it such that a manufacturer will not ‘ordinarily’ be able to show an
irreconcilable conflict between state and federal law.”); Crockett v. Luitpold
Pharmaceuticals, Inc., No. 19-276, 2020 WL 433367, at *6 (E.D. Pa. Jan. 28, 2020)
(“The defense of impossibility preemption is premised on a contention that a federal
regulation would have prohibited the additional warnings that the plaintiff alleges
state law requires.”); Yamagata v. Reckitt Benckiser LLC, 445 F. Supp. 3d 28, 33 (N.D.
Cal. 2020) (“The preemption analysis in [Merck] turned on whether the FDA would
have approved a change to the drug label.”); McGrath v. Bayer HealthCare
Pharmaceuticals, Inc., 393 F. Supp. 3d 161, 171 (E.D.N.Y. 2019) (finding preemption
because the plaintiff “has not pleaded a plausible claim that the CBE regulation
would have permitted [the defendant] to change the [drug] label”); Silverstein v.
Boehringer Ingelheim Pharmaceuticals, Inc., No. 19-81188, 2020 WL 6110909, at *9

29

(S.D. Fla. Oct. 7, 2020) (“[Preemption] can be satisfied [under Merck] even if the
labeling change has not been presented to, and rejected by, the FDA.”).
As such, based on these authorities, the “universal” standard that a
manufacturer need not submit a PAS and CBE to the FDA to preserve its preemption
defense remains intact after Merck. See, e.g., Cerveny v. Aventis, Inc., 155 F. Supp. 3d
1203, 1213-16 (D. Utah Mar. 16, 2016) (“Courts have universally rejected the notion
that Wyeth requires a showing that the manufacturer attempted to apply the
warning suggested by the plaintiff but that the labeling change was ultimately
rejected by the FDA.”); In re Zofran (Ondansetron) Prod. Liab. Litig., No. 15-2657,
2021 WL 2209871, at *32 (D. Mass. June 1, 2021) (“Multiple courts have found
[conflict] preemption where the manufacturer had not requested the precise warning
sought by the plaintiffs when the FDA had nonetheless made it clear that it would
not accept that label change.”); Ridings v. Maurice, 444 F. Supp. 3d 973, 998 (W.D.
Mo. 2020) (finding the second prong of Merck to be satisfied when all of the
information justifying the proposed warning had been given to the FDA and the FDA
did not revise the label to add the warning); Seufert v. Merck Sharp & Dohme Corp.,
187 F. Supp. 3d 1163, 1170 (S.D. Cal. May 11, 2016) (“[M]anufacturer submission of
a proposed labeling change is relevant, but not dispositive, in determining whether a
defendant can establish conflict preemption.”).
In the end, it is, of course, the Supreme Court’s “prerogative alone to overrule
one of its precedents.” State Oil v. Khan, 522 U.S. 3, 20 (1997). But it is difficult to
reconcile the Court doing so when no party disputed Wyeth’s clear evidence standard

30

on appeal, 12 when the question before the Court was who should apply that standard,
not whether the standard should survive, and when the Court itself held that its
decision “flow[ed] from [its] precedents.” 139 S.Ct. at 1678 (emphasis added).
Accordingly, like all other courts having considered the issue, I find that Merck does
not overrule Wyeth.
B.

Prong One of Impossibility Preemption

I now turn to the substance of the parties’ dispute. To establish impossibility
preemption, a drug manufacturer must first show that it “fully informed the FDA of
the justifications for the warning required by state law.” Merck, 139 S.Ct. at 1678. I
find that Defendant has met this standard; indeed, Judge Pisano found as much, the
Third Circuit agreed, and the Supreme Court never questioned that finding on
appeal. I reach the same conclusion based on my independent evaluation of the
record.
After a full trial on the merits, including extensive expert testimony, Judge
Pisano found no evidence that “Defendant failed to provide all the information it had
. . . to the FDA.” Glynn, 951 F. Supp. 2d at 703, 705. After a post-trial opportunity for
Plaintiffs to present further proof, Judge Pisano again rejected their claim as
“speculation.” OTSC Opinion, 2014 WL 1266994, at *14, *17. The Third Circuit
characterized the record in more certain terms: “Merck kept the FDA informed of the
scores of case studies, reports, and articles . . . published documenting possible
connections between long-term bisphosphonate use and atypical femoral fractures,”
12
Tellingly, Plaintiffs themselves argued on appeal that Wyeth “was correctly decided.”
U.S. Merits Brief, at *25-28.

31

and “[i]t is undisputed that the FDA was aware of the possible link between Fosamax
and atypical fractures well before September 2010.” Fosamax, 852 F.3d at 275, 296.
The Supreme Court did not consider—let alone challenge—these factual findings on
appeal. Merck, 139 S.Ct. at 1680.
Plaintiffs disagree, pointing first to the way in which the Supreme Court’s
summary of the facts characterizes what the FDA knew and when. 13 See, e.g., Merck,
139 S.Ct. at 1673-76. But that is insufficient to support the inference that the Court
actually found that Merck did not fully inform the FDA of the risks of Fosamax. For
one, the Court is “an appellate tribunal, ill-equipped for the task of factfinding,” and
prong one of impossibility preemption is a fact-intensive inquiry involving a record
exceeding one-thousand pages. Ohio v. Wyandotte Chems. Corp., 401 U.S. 493, 498
(1971). More to the point, the Court does not “lightly overturn the concurrent findings
of the two lower courts” on factual matters, in the background section of an opinion,
especially not without any explanation, and when such findings were never on review

Generally speaking, the Supreme Court used a harsher tone when describing Merck’s
actions throughout the labeling process. Merck, 139 S.Ct. at 1673-76. The Court stated that
at the time the FDA first approved the Fosamax label in 1995, Merck scientists were aware
of the risk of atypical femoral fractures, but “perhaps because [Merck’s] concerns were only
theoretical, the FDA approved Fosamax’s label without requiring any mention of this risk.”
Id. at 1674 (emphasis added). Then, in 2008, after additional scientific evidence arose
connecting Fosamax to atypical femoral fractures, the Court explained that Merck applied to
the FDA for preapproval to change the drug’s label, attempting to add language to both the
Adverse Reactions and the Precautions sections of the label. Id. The Court emphasized that
although the FDA denied Merck’s request, it also invited the manufacturer to “resubmit” its
application to “fully address all the deficiencies” identified by the FDA’s review. Id.
According to the Court, however, Merck “instead withdrew its application,” choosing to make
the changes to the Adverse Reactions section through the CBE process. Id. Moreover, with
respect to the Fosamax label’s eventual warning about atypical femoral fractures, the Court
commented that Merck was “initially resistant” to the change, because it failed to reference
“stress fractures.” Id. at 1674-75.
13

32

in the first place. Glossip v. Gross, 576 U.S. 863, 882 (2015); Exxon Co., U.S.A. v.
Sofec, Inc., 517 U.S. 830, 841 (1996) (explaining that the Court “‘cannot undertake to
review concurrent findings of fact by two courts below in the absence of a very obvious
and exceptional showing of error’”) (quoting Graver Tank & Mfg. Co. v. Linde Air
Products Co., 336 U.S. 271, 275 (1949)); cf. Easley v. Cromartie, 532 U.S. 234, 242
(2001) (doing so only when “there is no intermediate court,” “we are the only court of
review,” “the trial here at issue was not lengthy and the key evidence consisted
primarily of documents and expert testimony,” and “[c]redibility evaluations played
a minor role”). Saliently, Justice Alito concurred in part to ensure that the majority’s
“discussion . . . of the facts is not misleading.” Merck, 139 S.Ct. at 1684. On this point,
Justice Alito wrote, “[r]esolution of the legal question that the Court decides does not
require much discussion of the facts, but . . . . the Court provides a one-sided account
. . . [that] omits any mention of the extensive communication between Merck and the
FDA during the relevant period.” Id. at 1685. In the end, Plaintiffs’ position cannot
be reconciled with the Court’s explicit decision to remand with instructions to apply
its standards anew. Cf. Farrar v. Hobby, 506 U.S. 103 (1992) (deciding the merits
rather than remanding); McCkeskey v. Zant, 499 U.S. 467 (1991) (same).
Plaintiffs then argue that the standard Judge Pisano applied in Glynn is
somehow less demanding than Merck’s requirement that Defendant “fully” inform the
FDA. Pl. Br., at 32-34 (“Judge Pisano did not apply the Supreme Court’s standard [in
Merck].”). But Plaintiffs never explain Judge Pisano’s supposedly erroneous
standard, the relevant difference between that standard and the Merck standard, or

33

why such a difference would be legally significant. Independently, I do not see any
meaningful difference between what Merck demands and what Judge Pisano
determined. Under Merck, the basic inquiry, which Judge Pisano applied, is whether
the FDA had “all the information it deemed necessary to decide whether to approve
or reject the proposed warning at the time it issued the [CRL].” In re Avandia, 945
F.3d at 759 (emphasis removed). Indeed, Merck itself phrases the inquiry in a
substantially similar way: “the litigants may dispute whether the drug manufacturer
submitted all material information to the FDA.” 139 S.Ct. at 1680.
In any event, revisiting this question as a matter of first impression, as
instructed by the Third Circuit, I reach the same result as Judge Pisano. Between its
formal safety updates, periodic emails, and PAS, Defendant clearly and fully
informed the FDA of the panoply of risks associated with long-term Fosamax use and
the justifications for its proposed label change. Having culled through the extensive
record, I summarize below what Defendant submitted to the FDA.

Defendant

repeatedly and voluntarily sent relevant articles to the FDA between 1992 and 2010.
See Fosamax, 852 F.3d at 275 (citing A1774, A1258, A1237, A1243); Def. Rep. Br., Ex.
13, at A1928-33; Fosamax, 862 F.3d at 275 (further describing communications).
Indeed, Defendant’s 165-page March 2008 safety update, which surveyed medical
studies, journal publications, and internal data compiled between July 16, 2007 and
January 15, 2008, included numerous pages on atypical femoral fractures. Def. Rep.
Br., Ex. 14. That safety update provided (1) an overview of three published safety
studies identified in the medical literature describing new information regarding the

34

connection between prolonged alendronate 14 use and low-energy or atypical femoral
fractures, (2) a discussion of eight publications on long-term therapy with
bisphosphonates, including the link between prolonged bisphosphonate therapy and
atypical low-energy femoral fractures, and (3) a summary of post-marketing data on
atypical low-energy fractures associated with prolonged bisphosphonate therapy in
response to the FDA’s request for such an update. Id. at A2594-2613.
With respect to the three safety studies and various publications, Merck
cautioned that although they contain important clinical information, some of the
studies and publications found no “obvious defects in mineralization or bone quality
after use of the drug.” Id. at A2595. However, Merck did highlight one particular
study that “raised the possibility of a link between prolonged bisphosphonate therapy
and atypical low-energy non-vertebral fractures, predominantly with femoral
diaphyseal location.”

Id.

The authors of that study attributed this pattern of

fractures to severely suppressed bone turnover that may develop during long-term
alendronate therapy. Id. Similarly, several of the publications referenced in the safety
update also hypothesized about a link between prolonged bisphosphonate therapy
and the atypical low-energy fractures suffered by Plaintiffs. Id. at A2597-98. Finally,
the safety update provided a trove of data, compiled from Merck’s Worldwide Adverse
Experience System database using search terms like “bone disorder,” “stress
fracture,” “femur fracture,” and “bone formation decreased.” Id. at A2598-2613.
Using these terms, Merck generated 175 post-marketing reports, providing insight

14

Alendronate is a type of bisphosphonate.

35

into patients treated with alendronate sodium from October 1, 2005 through
December 31, 2007. Id. at A2599. Specifically, the data from the post-marketing
reports included, inter alia, the age and gender of the patient, location of the fracture,
and the duration of alendronate therapy. Id. at A2609-10. While Merck commented
that a review of the post-marketing reports did not provide “clear evidence of a causal
link” between alendronate therapy and atypical low-energy femoral fractures, it
committed to further monitor future reports for these types of fractures. Id. at A2613.
In June 2008, Defendant “promptly complied” with the FDA’s request for
further investigations that Merck had conducted and reports Merck had received.
Fosamax, 862 F.3d at 275. And, what is more, Defendant’s September 2008 PAS not
only cited nine articles reporting cases of low-energy femoral fractures in Fosamax
users, but included a clinical overview in which Defendant itself asserted a
statistically significant association. Cf. Wyeth, 555 U.S. at 272-73 (noting that the
manufacturer never “supplied the FDA with an evaluation or analysis concerning the
specific dangers” at issue); In re Taxotere (Docetaxel) Prod. Liab. Litig., No. 16-17039,
2020 WL 7480623, at *11 (E.D. La. Dec. 18, 2020) (finding that the FDA was not “fully
informed” because its limited knowledge of the risk and repeated requests to the
manufacturer for information indicated that the manufacturer was not “making an
‘earnest attempt’ to keep the FDA informed”) (citations omitted). Despite this profuse
evidence of information sent to the FDA, Plaintiffs, on remand, insist that more
evidence was needed, and that Merck misled the FDA with the information it sent.

36

Having reviewed, myself, those documents, I find no basis in the record to reach that
conclusion.
In that regard, Plaintiffs’ evidence that the FDA was somehow left in the dark
about the use of Fosamax and the potential risk of atypical femoral fractures is
unpersuasive. Plaintiffs begin by offering six specific studies between 1995 and 2010
which purport to show a connection between long-term bisphosphonate use and
atypical femoral fractures. The flagrant flaw with Plaintiffs’ proffer, however, is that
Defendant cited all these same studies in its communications with the FDA. Plaintiffs
then take issue with minute details of the data Defendant submitted to the FDA,
which they insist shows that Defendant “provid[ed] misleading information . . . [,]
describ[ed] atypical femoral fractures inaccurately and conflat[ed] them with stress
fractures.” Pl. Br., at 31-32. Specifically, Plaintiffs maintain that Defendant (1) did
not “provide the FDA with any possible pathogenesis for [atypical femoral fractures],”
id., Ex. 3, at A884; (2) stated in its clinical overview that “fractures with similar
clinical features had previously been reported in patients not taking Fosamax,” id. at
A881; (3) “identified risk factors that simply were not associated with [atypical
femoral fractures],” id.; and (4) failed to provide “additional information” after
receiving the CRL in May 2009, “should have provided [the clarification which came
from the September 2010 Task Force Report] much earlier,” and “rebuffed the FDA’s
repeated pleas for further engagement” prior to the Task Force Report. Pl. Br., at 3334. Based on the record before me, I disagree, and I address each of these,
individually.

37

Pathogenesis. Plaintiffs first argue that Defendant did not provide the FDA
with any possible pathogenesis, the manner of development of a disease, for atypical
femoral fractures. The record belies this assertion. Defendant repeatedly indicated
how Fosamax might cause the very injury Plaintiffs suffered. See, e.g., Def. Br., Ex.
1, at A2757 (mentioning “[s]everely suppressed bone turnover”); id. at A2754
(describing “bone biopsy results” which “indicated low bone turnover”); Def. Rep. Br.,
Ex. 14, at A2597 (explaining, in its safety update, that the attached studies “related
[atypical femoral fractures] to severely suppressed bone turnover that may develop
during long-term” Fosamax use). In fact, in clinical trials three decades ago,
Defendant informed the FDA that “antiresorptive agents may inhibit microdamage
repair by preventing . . . bone resorption at the sites of microdamage,” Fosamax, 852
F.3d at 275 (citing A1774), which was borne out to be the correct pathogenesis
according to Plaintiffs’ own experts. Id., Ex. 3, at A880 (“[D]ecreased bone toughness
can lead to stress fracture. Fosamax and other [bisphosphonates] can reduce the
body’s ability to repair a stress fracture once it has begun, prior to complete fracture.
This might explain why a large number of bisphosphonate-induced stress fractures
go on to completion.”).

Indeed, on appeal, the Third Circuit acknowledged that

“Merck kept the FDA informed” of the “scores of case studies, reports, and articles …
published documenting possible connections between long-term bisphosphonate use
and atypical femoral fractures.” Fosamax, 852 F.3d at 275. Plus, the FDA itself has
since agreed that Merck “provided [the Agency] with the relevant scientific data about

38

Fosamax’s risks.” FDA Brief as Amicus Curiae, at *14. 15 Thus, based on the record,
this argument lacks merit.
Clinical Features. Plaintiffs next argue that Defendant’s clinical overview
indicated that some clinical features associated with Fosamax use presented in
patients not taking Fosamax. To be clear, the PAS stated, in this regard, only that
“stress fractures with similar clinical features also have occurred in patients not
treated with bisphosphonates.” Based on this solitary statement, Plaintiffs suggest
that by failing to communicate the “unique features” of atypical femoral fractures to
the FDA, in particular, their “fracture pattern,” Defendant created a misleading
impression that such fractures are “much more common in the absence of
[bisphosphonates]” than they actually are. Pl. Opp. Br., at 32.
Upon closer examination of the PAS’s clinical overview, however, the Court
does not find Defendant’s submission misleading, deceptive, or ambiguous in any
way. While the clinical overview identified that the fractures at issue here occur in a
similar population of elderly individuals as other osteoporotic low-energy fractures,
it also explained that “these [atypical femoral] fractures are less common than other
osteoporotic low-energy fractures,” and only represent “about 6% of fractures of the
femur.” In other words, Defendant informed the FDA that atypical femoral fractures
are rare—even in elderly individuals who are taking Fosamax, and in that regard,
there is no evidence in the record of Defendant “hiding-the-ball,” as suggested by
Plaintiffs.

Moreover, while Plaintiffs take issue with the fact that the clinical

The Court notes that the FDA filed this brief as amicus curiae in support of Defendant
in Merck.

14

39

overview states that atypical femoral fractures have been reported in patients not
taking bisphosphonates, they fail to acknowledge the significant fact that the FDAmandated warning itself observed that atypical femoral fractures occur in
“osteoporotic patients who have not been treated with bisphosphonates.” Indeed, all
along, the FDA questioned whether taking bisphosphonates for a prolonged period of
time would actually lead to more atypical femoral fractures because other
osteoporotic patients, who were not on such therapy, also suffer from the same
fractures. This is the very causation-related concern that led the FDA to reject
Merck’s PAS in the first place. See infra. In the end, even though the FDA approved
an amendment regarding atypical femoral fractures, the warning includes the
observation that osteoporotic patients, generally, have suffered such fractures. This
fact, alone, dooms Plaintiffs’ claim that Defendant misled the FDA by pointing out
the same.
False Risk Factors. Plaintiffs contend further that Defendant emphasized
“false risk factors” in materials sent to the FDA, the implication being that Merck
“attempted to confound the true nature of the association between Fosamax and
[atypical femoral fractures].” Pl. Br., at 32. Specifically, Plaintiffs argue that when
the Task Force examined the actual data, some of the risk factors identified in the
clinical overview and Defendant’s proposed warning, namely, “abnormally decreased
bone mineral density associated with osteoporosis, long-term immobilization/disuse,
and use of glucocorticoids, the presence of joint deformity, leg-length discrepancies,
muscle weakness, and spasm with resulting alteration in force distribution across the

40

joints,” “simply were not associated with [atypical femoral fractures].” Id., Ex. 3, at
A882-83.
But, Plaintiffs misconstrue the language of the PAS to support their position.
Rather, the “Spontaneous Reports” section of the clinical overview examined 132
reports where alendronate therapy was given for treatment of several conditions,
looking specifically for evidence and information related to fractures. In part, it also
discussed fracture risk factors, noting that 70 of the 132 reports provided sufficient
information on the patient’s medical history, concurrent conditions, and concomitant
medications. In that regard, however, the clinical overview did not express any
conclusions, nor did it make any pronouncements. Instead, it provided a laundry list
of pre-existing conditions, comorbidities, and other attributes, along with the
percentage of the 70 patients whose medical history reported those conditions.
Specifically, the clinical overview stated that musculoskeletal disorders, including
osteoarthritis and rheumatoid arthritis, were reported in 38 of the 70 patients; the
“presence of joint deformities, muscle imbalance, leg-length discrepancies, and
change in activity” were “common” for this subgroup of patients; and 28 of the 70
patients had a history of fracture. Indeed, the clinical overview also highlighted that
only 10 of the 70 patients sustained atypical fractures following joint replacement or
surgery, 17 patients had endocrine or metabolic disorders like diabetes mellitus and
obesity, 10 patients reported malignant disease, and 3 patients were smokers. Thus,
the purpose of the “Spontaneous Reports” section of the clinical overview was not to
provide a definitive list of risk factors, but rather to provide a complete picture of the

41

clinical landscape to physicians prescribing Fosamax. Def. Br., Ex. 1, A2754-A2756.
It was not meant, as Plaintiffs have advanced, to obfuscate the seriousness of
potential injuries or to mislead the FDA. The record reflects that Defendant clearly
appreciated the seriousness, and sought to alert the FDA, of these fractures on
numerous occasions. Id. at A2756 (“[C]onsidering the clinical importance of these
fractures . . . it is important to include an appropriate statement about them.”). More
compelling, the Task Force Report also concluded that certain fracture risk factors,
unrelated to bisphosphonate use, exist; the Report specified that comorbid conditions
are “Minor Features” of atypical femoral fractures, making them relevant rather than
irrelevant. See infra.
Failure to Provide Additional Information. Finally, Plaintiffs assert that
Defendant deprived the FDA of relevant information between 2008 and 2009, such
as information that the Task Force eventually reported, leaving the agency
“uncertain about the nature of atypical femoral fractures” and “[d]elayed by
[Defendant’s] inaction.” Pl. Opp. Br., at 34. This argument also lacks merit. For one,
Plaintiffs do not point to any specific instance in which Defendant failed to provide
any timely and relevant information, data, case studies, or evidence to the FDA, or
rebuffed a request for further engagement. While Plaintiffs make much of
Defendant’s decision to withdraw its PAS instead of applying for a formal meeting,
they ignore the fact that Defendant did so at the FDA’s direction, Def. Br., Ex. 3, at
A1498, that it was entitled to do so by statute, 21 U.S.C. § 314.70(c)(6)(iii)(A), and
that it subsequently stated in its CBE amendment to the Adverse Reactions section

42

that it still wished to discuss a Precautions warning. Id., Ex. 4, A2963-64 (“Merck
believes that further discussion with regard to text for the Precautions section of the
label . . . would be beneficial.”). Likewise, Plaintiffs’ contention that, Defendant
should have provided the additional information contained in the Task Force Report
before the Task Force independently reviewed it, fails. The Task Force relied on 24
new case studies and 63 new articles after the FDA issued its CRL, according to
Plaintiffs’ own experts. Pl. Br., Ex. 3, at A879 (“In 2008 [at the time of the PAS], 13
of [ ] 37 published case series and reports [cited by the Task Force] were available to
Merck. By May of 2009, 19 of [ ] 37 published case series were available to Merck.
Additionally, the Task Force cited a total of 177 published or available articles and
posters. Of those 177, 114 were available in 2008 [at the time of the PAS] or earlier
and 120 were available before May of 2009.”). Additionally, Defendant knew that the
FDA, outside experts, and other manufacturers were working “closely” during this
period to study atypical femoral fractures, which obviated the need to continue
forwarding piecemeal research, see, e.g., Pl. Br., Ex. 18, at A1508, particularly since
the FDA specifically informed Merck that the Agency will continue to independently
study and investigate the issues.
If any doubt remains as to whether Defendant fully informed the FDA of the
justification for its warning, the Agency itself agrees that Defendant “provided [it]
with the relevant scientific data about Fosamax’s risks.” FDA Brief as Amicus Curiae,
at *14. Because the FDA alone is the “arbiter of which data and information is or is
not ‘material’ to [its] decision to approve or reject a change to a drug’s label” under

43

Merck, the FDA’s view of the evidence matters. 16 In re Avandia, 945 F.3d at 759.
Accordingly, I conclude that Defendant has satisfied the first Merck prong.
C.

Prong 2 of Impossibility Preemption

As to the second prong of preemption, the crux of the parties’ dispute is
whether the FDA informed Defendant that it would not approve changing Fosamax’s
label to add the warning required by state law. Arguing in the negative, Plaintiffs
advance two reasons why: (1) the CRL does not carry preemptive effective because it
is not a final agency action, and (2) the FDA rejected Defendant’s proposed warning
for emphasizing “garden variety” stress fractures, not because it disagreed with the
underlying science linking Fosamax use to atypical femoral fractures; in that regard,
Plaintiffs claim that the Agency might have approved some other version of the
warning had Defendant proposed one. I will address each, in turn.
i.

The Preemptive Effect of the CRL

Plaintiffs argue—for the very first time in this long-pending MDL—that the
CRL is not preemptive because it is not a final agency action which consummates the
FDA’s decisionmaking process. Pl. Br., at 12-14.
The Supremacy Clause grants “supreme” status only to the “the Laws of the
United States.” U.S. CONST. ART. VI, CL. 2. “Nothing short of federal law can have that
effect.” Fellner, 539 F.3d at 243; Gibbons v. Ogden, 22 U.S. 1 (1824). Federal agency
actions can constitute “Laws” in the sense of the Supremacy Clause. Hillsborough

16
Because I conclude that Defendant fully informed the FDA of the justifications for its
warning, I need not address Defendant’s contention that Plaintiffs “waived any contrary
argument [on this issue] several times over” by not raising it on appeal. Def. Br., at 20.

44

County, 471 U.S. at 713 (“[S]tate laws can be preempted by federal regulations as
well as by federal statutes.”); New York v. Fed. Commc’n Comm’n, 486 U.S. 57, 63
(1988) (“The phrase ‘Laws of the United States’ [in the Supremacy Clause]
encompasses both federal statutes themselves and federal regulations that are
properly adopted in accordance with statutory authorization”). However, this applies
“only when [ ] [the agency] is acting within the scope of its congressionally delegated
authority, . . . for an agency literally has no power to act, let alone pre-empt the validly
enacted legislation of a sovereign State, unless and until Congress confers power upon
it.” New York v. FERC, 535 U.S. 1, 18 (2002) (quotations and alterations omitted);
Fidelity Fed. Savings and Loan Ass’n v. de la Cuesta, 458 U.S. 141, 153-54 (1982).
Relying on Justice Thomas’ concurrence, Plaintiffs argue that the CRL does
not carry preemptive effect because it is not a final agency action. Pl. Br., at 27-28.
According to Plaintiffs, the CRL does not mark “the consummation of the agency’s
decisionmaking process,” Bennett v. Spear, 520 U.S. 154, 178 (1997) (quotations
omitted), finally determine the parties’ “rights or obligations,” or impose “legal
consequences.” Port of Boston Marine Terminal Assn. v. Rederiaktiebolaget
Transatlantic, 400 U.S. 62, 71 (1970). This argument is misplaced for several reasons.
To begin, the majority in Merck explicitly cited 21 C.F.R. § 314.110(a), which
empowers the FDA to “formally reject” a drug manufacturer’s proposed warning
through a CRL, as an example of an FDA action that does constitute “Law” in the
sense of the Supremacy Clause. 139 S.Ct. at 1679. That should end the inquiry.

45

In any event, Plaintiffs’ position appears to confuse the question whether an
agency action is final—for example, for the purposes of providing judicial review
under the Administrative Procedures Act, Port of Boston, 400 U.S. at 71—with the
question of whether the agency action is “Law” with the power to preempt. These are
distinct inquiries and have different legal consequences. The preemption question
turns on whether Congress delegated to the agency the authority to act in such a
manner in the first instance, not on whether the agency’s action is necessarily a
“final” one. FERC, 535 U.S. at 19 (“This sort of case . . . defining the proper scope of
federal power . . . requires us to be certain that Congress has conferred the
authority.”). The yardstick is congressional intent, not the finality of its action. See,
e.g., English v. General Elec. Co., 496 U.S. 72, 78-79 (1990) (“[P]reemption
fundamentally is a question of congressional intent.”); Medtronic, 518 U.S. at 485
(“‘The purpose of Congress is the ultimate touchstone’ in every preemption case.”)
(quoting Retail Clerks v. Schermerhorn, 375 U.S. 96, 103 (1963)); Cipollone v. Liggett
Group, Inc., 505 U.S. 504, 516, 530 n.27 (1992) (holding that the scope of preemption
must rest “on a fair understanding of congressional purpose”); Malone v. White Motor
Corp., 435 U.S. 497, 504 (1978) (“It is uncontested that whether [the statute at issue
is preempted] depends on the intent of Congress.”); Louisiana Pub. Serv. Comm’n v.
FCC, 476 U.S. 355, 374 (1986) (stating that the best way to determine preemption “is
to examine the nature and scope of the authority granted by Congress to the agency”).
It follows that for preemption purposes, it is mostly irrelevant whether the
CRL is “of a merely tentative or interlocutory nature,” Bennett, 520 U.S. at 178, or

46

that it simply “informs sponsors of changes that must be made before an application
can be approved, with no implication as to the ultimate approvability of the
application.” 73 Fed. Reg. 39588. As Defendant points out, Def. Rep. Br., at 11-12, if
Plaintiffs’ position were to prevail, no CRL could ever carry preemptive effect because
all CRLs require some subsequent action on the part of the manufacturer, and
preserve some procedural mechanism to further engage with the FDA, even if futile.
21 C.F.R. § 314.110(b) (providing three options: “[r]esubmit the application . . . ,
addressing all deficiencies identified in the [CRL],” “[w]ithdraw the application . . .
without prejudice to a subsequent submission,” or “[a]sk the agency to provide . . . an
opportunity for a hearing,” after which “the agency will either approve” or “refuse . .
. the application”). And, more importantly, it would abrogate the very preemption
effect of the federal regulation, 21 C.F.R. 314.110(a), that the FDA promulgated
pursuant to congressional authority. For these reasons, I reject Plaintiffs’ claim that
the CRL does not have preemptive effect under the Supremacy Clause. I turn, next,
to the content of the CRL.
ii.

The CRL

The parties dispute how to construe the meaning, and impact, of the CRL,
which centers on four issues: (1) whether Defendant proposed an adequate warning;
(2) whether the contents of the CRL, alone, support the inference that the FDA
rejected Defendant’s warning based on the Agency’s belief that the underlying science
did not justify one; (3) if the CRL does not convey such an inference on its face,
whether the CRL, when construed in addition to the FDA’s other communications

47

from the same time period, support that inference; and (4) how the surrounding
regulatory regime informs the CRL. Plaintiffs posit that Defendant’s warning was
inadequate under state law, and the FDA rejected it merely because of the general
“stress fractures” language, which does not indicate whether a differently worded
warning would have been accepted by the FDA. Defendant, on the other hand,
maintains that it sought to warn of the very injury Plaintiffs suffered, and the CRL—
construed either on its own or in light of the FDA’s other communications—prohibited
Merck from adding any and all warnings to the Fosamax label because the Agency
seriously questioned, and therefore doubted, a causal connection between
bisphosphonates and atypical femoral fractures.
1. Adequacy of Defendant’s Proposed Warning
To show that the FDA rejected a warning that would have been adequate under
state law, Defendant must first establish that it actually proposed such a warning,
an implicit but critical step in the analysis. Plaintiffs insist that Defendant failed to
do so, because Defendant merely proposed “garden variety” stress fractures in its
language, rather than atypical femoral fractures, despite scientific evidence allegedly
differentiating between the two. Pl. Br., at 1, 5, 16. Plaintiffs point to the text of the
warning as support: “every sentence after the first sentence described . . . ‘stress
fractures’” not “atypical” fractures, id. at 17, the warning referenced “similar clinical
features” in fractures in “patients not treated with bisphosphonates,” and Defendant
suggested evaluating patients for other “known causes and risk factors,” in addition
to bisphosphonate use.

48

Defendant responds that it “tried to warn of the precise low-energy fractures
that Plaintiffs allegedly suffered.” Def. Rep. Br., at 5. In its proposed warning,
Defendant highlights that it emphasized the essential features of atypical femoral
fractures even if it did not use the term “atypical.” Id. Defendant also points to “the
warning that the FDA mandated following the Task Force Report,” which conveys
similar information as Defendant’s proposed one, id. at 8, and which Plaintiffs
concede is adequate. Pl. Br., at 10. Finally, Defendant notes communications with the
FDA characterizing the warning as pertaining to “atypical . . . fractures,” Def. Br.,
Ex. 2, at A1500; Pl. Br., Ex. 10, at A1145, and expert testimony that it “approach[ed]
the FDA with respect to [such] fractures.” Def. Rep. Br., at 8; Def. Br., Ex. 3, at A1498;
id., Ex. 15, at 660.
As an initial matter, Plaintiffs raised this argument before Judge Pisano to no
avail. Glynn, 951 F. Supp. 2d at 701 (rejecting position that “the FDA rejected the
PAS because [Defendant] used the phrase ‘stress fracture’ instead of ‘atypical’
fracture, and the FDA would have approved an appropriately worded warning”). After
hearing expert testimony from both parties on the relevant terminology, Judge
Pisano found Defendant’s warning to contain “the same language” that Plaintiffs
contend state law requires. Id. at 703-04.
I reach the same result upon a fresh review of the record. To reiterate, Merck
proposed adding the following language to the Precautions section:
Low-Energy Femoral Shaft Fracture
Low-energy fractures of the subtrochanteric and proximal femoral shaft
have been reported in a small number of bisphosphonate-treated

49

patients. Some were stress fractures (also known as insufficiency
fractures) occurring in the absence of trauma. Some patients
experienced prodromal pain in the affected area, often associated with
imaging features of stress fracture, weeks to months before a complete
fracture occurred. The number of reports of this condition is very low,
and stress fractures with similar clinical features also have occurred in
patients not treated with bisphosphonates. Patients with suspected
stress fractures should be evaluated, including evaluation for known
causes and risk factors (e.g., vitamin D deficiency, malabsorption,
glucocorticoid use, previous stress fracture, lower extremity arthritis or
fracture, extreme or increased exercise, diabetes mellitus, chronic
alcohol abuse), and receive appropriate orthopedic care. Interruption of
bisphosphonate therapy in patients with stress fractures should be
considered, pending evaluation of the patient, based on individual
benefit/risk assessment.
Pl. Br., Ex. 38 at A1371. To begin, I refer to the science regarding bone resorption
and formation. All bones, whether healthy or osteoporotic, can develop microscopic
cracks—called stress fractures—from everyday activity. These “ordinarily heal on
their own through the bone remodeling process.” Merck, 139 S.Ct. at 1673. When that
process is disrupted, from a bisphosphonate for example, the body may not naturally
repair itself, creating stress fractures as a result. Relevant here, stress fractures may
then progress to atypical femoral fractures, or complete breaks of the femur, which
cause pain and require surgery rather than rest. Stated differently, atypical femoral
fractures are stress fractures, but more severe than other types of stress fractures,
such as those that heal on their own. Shane et al., Atypical Subtrochanteric and
Diaphyseal Femoral Fractures: Second Report of a Task Force of the American Society
for Bone and Mineral Research, 29 J. Bone & Min. Res. 1, 12 (2014) (concluding same).
Plaintiffs all but concede this point: atypical femoral fractures “start as . . . stress
fractures.” Pl. Br., Ex. 4, at 12.

50

It is also important to consider the Task Force Report, which defined key
characteristics of, and risk factors for, atypical femoral fractures. The Task Force
listed “Major Features,” which are necessary to diagnose a patient with an atypical
femoral fracture, and Minor Features, which may be associated with such a fracture
but are not required characteristics. As to the Major Features, the fracture is (1)
“located anywhere along the femur from the distal to the lesser trochanter to just
proximal to the supracondylar flare”; (2) “associated with no trauma or minimal
trauma”; (3) transverse or short oblique in configuration; (4) noncomminuted,
meaning that there are not multiple breaks; and (5) complete in that it extends
through both cortices and may be associated with a medial spike. The Minor Features
are: (1) localized periosteal reaction of the lateral cortex; (2) generalized increase in
cortical thickness of the diaphysis; (3) prodromal symptoms such as dull or aching
pain in the groin or thigh; (4) bilateral fractures and symptoms; (5) delayed healing;
(6) comorbid conditions (e.g., vitamin D deficiency, rheumatoid arthritis,
hypophosphatasia); and (7) use of pharmaceutical agents (e.g., bisphosphonates,
glucocorticoids, and proton pump inhibitors). Shane et al., Atypical Subtrochanteric
and Diaphyseal Femoral Fractures: Report of a Task Force of the American Society for
Bone and Mineral Research, 25 J. Bone & Min. Res. 2267, 2268-69 (2010).
Having set forth the foundational science, I turn to the proposed warning.
First, Plaintiffs argue that the title “Low-Energy Femoral Shaft Fracture” references
“a broad category” of fractures including “[atypical femoral fractures] and less serious
fractures,” Pl. Br., at 17 n.3, and thus, does not constitute an adequate warning. I

51

disagree. The title itself describes aspects of an atypical fracture, that is, it occurs
from minimal trauma (i.e., low-energy) and in a discrete part of the thigh bone (i.e.,
the femoral shaft), which, according to the Task Force, are the two Major Features of
atypical femoral fractures. This is consistent with the Patient Packet Insert for
Fosamax, which alerts patients that some users “have experienced fracture in a
specific part of the thigh bone.” Def. Br., Ex. 1, at A2742 (emphasis added). The first
sentence of the warning then describes that the type of fracture at issue, or the subject
of the warning, occurs in the “subtrochanteric and proximal” region of the “femoral
shaft,” which is another Major Feature identified in the Report, and a distinguishing
characteristic according to Plaintiffs’ own brief. Pl. Br., Add. 8 (containing an x-ray
image of an atypical femoral fracture displaying these features); id., Ex. 2, at A114849 (explaining that atypical femoral fractures are distinguishable, in part, because
they occur “perpendicular to the femoral shaft” and in “the proximal (upper) third . .
. or the subtrochanteric region”).
Next, the warning advises that “[s]ome” low-energy femoral shaft fractures
“[are] stress fractures.” Plaintiffs interpret this sentence as conflating garden variety
stress fractures with atypical femoral fractures, despite a distinction between them.
Pl. Br., at 32 (“Merck improperly conflated the underlying fracture mechanism that
leads to [atypical femoral fractures] with the ultimate outcome.”). I do not see any
basis in the science for such a strict dichotomy. As discussed supra, atypical femoral
fractures are stress fractures, just severe ones and located in a particular part of the
body, exhibiting a difference in degree but not necessarily in kind. Merck, 139 S.Ct.

52

at 1674 (stating that atypical femoral fractures “progress” from microscopic stress
fractures); Glynn, 951 F. Supp. 3d at 704 (quoting one of Plaintiffs’ experts who
testified that Fosamax “can lead . . . to subsequent stress fracture formation”). In the
Task Force’s own words, “[t]he radiologic presentation of atypical femoral fractures
bears striking similarities to that of stress fractures.” Shane et al., at 2270. In
addition, in Plaintiffs’ expert’s words, “decreased bone toughness can lead to stress
fracture. Fosamax and other [bisphosphonates] can reduce the body’s ability to repair
a stress fracture once it has begun, prior to complete fracture. This might explain
why a large number of bisphosphonate-induced stress fractures go on to completion.”
Pl. Br., Ex. 3, at A880.
On this point, Plaintiffs inexplicably overlook Defendant’s PAS, which
explains that Defendant uses the term “stress fracture” in its warning to mean an
“insufficiency fracture” that occurs with no “identifiable external traumatic event.”
Def. Br., Ex. 1, at A2751-52. While the term “stress fracture” often, in generic terms,
“connotes a fracture resulting from excessive loading of a normal bone,” as is common
in athletes, an “insufficiency [ ] fracture” is a specific type of stress fracture “caused
by normal loading of poor-quality bone,” as allegedly happened to Plaintiffs after
taking Fosamax and about which Defendant sought to warn. DORLAND’S ILLUSTRATED
MEDICAL DICTIONARY 710-11 (29th ed. 2000) (defining an “insufficiency f[racture]” as
“a stress fracture that occurs during normal stress on a bone of abnormally decreased
density”). Defendant’s PAS goes on to state that 91% of the fractures discussed
therein, and referenced in the warning, resulted in surgical intervention, while the

53

other 9% involved patients who sustained only “incomplete stress fractures,” Def. Br.,
Ex. 1, at A2755, which further distinguishes the warned-of injury from the garden
variety type.
Next, Plaintiffs cite an internal email between Merck colleagues from
December 19, 2010, in which several Merck employees shared redline revisions to
rationales for their proposed changes to the Fosamax label. Pl. Br., Ex. 27, at A1573.
Within those rationales, Defendant states that “most of the stress fractures general
physicians have seen are associated with repetitive stress injury related to exercise
(e.g., running) in younger adults, and that this type of stress fracture generally heals
well with rest.” Id. Plaintiffs offer this as “belated” evidence that Defendant knew its
“proposed focus on stress fractures [in 2008] would confuse general physicians.” Pl.
Br., at 18. Having reviewed the email, such an inference cannot be drawn. Setting
aside the questionable relevance of internal correspondence regarding the FDAmandated label change over two years later, read in context of the email, Defendant’s
statement sought to clear up any confusion by suggesting that physicians rule out
common causes before diagnosing a rarer atypical femoral fracture. Indeed, as the
warning stated, because “[t]he number of reports of this condition is very low,”
patients should “be evaluated . . . for known causes and risk factors.” This mirrors
the Task Force’s own determination two years later that atypical femoral fractures
occur with “relative rarity” and may be “associated” with “comorbid conditions.” Pl.
Br., Ex. 2, at A1147.

54

Furthermore, Plaintiffs make much of the fact that atypical femoral fractures
tend to “cause great pain,” Pl. Br., at 5 (quoting Merck, 139 S.Ct. at 1674); id. at 6
(describing them as “debilitating”); id. at 17 (same, but “gruesome”), and they
contrast this fracture with garden variety stress fractures, which usually heal
themselves with rest and presumably do not cause much pain. Id. at 18. Plaintiffs
argue that the lack of language regarding severe pain in Defendant’s warning is
evidence that Defendant was describing garden variety stress fractures.

But

Plaintiffs’ position is belied by: (1) the Task Force’s finding that such pain is a “Minor
Feature,” not a required characteristic; (2) Defendant’s warning indeed provides that
“[s]ome patients experience[] prodromal pain in the affected area” in any event, which
suffices to capture any potential pain-related difference between atypical femoral
fractures and garden variety stress fractures; and, most importantly, (3) the FDAmandated label includes an almost identical statement, which Plaintiffs concede is
adequate under state law. Def. Br., Ex. 7, at A1516-17. Further consistent with this
purported feature of atypical femoral fractures, the Patient Package Insert instructs
patients to “[c]all your doctor if you develop new or unusual pain in the hip or thigh.”
Def. Br., Ex. 1, at A2742 (emphasis added).
Relatedly, Plaintiffs emphasize the difference between “the nature” of garden
variety stress fractures, which are “barely perceptible,” and atypical femoral
fractures, where “the thigh bone (the largest and strongest bone in the body) looks
like a pencil snapped in two.” Pl. Br., at 18; compare id., Add. 8 (containing an x-ray
image of an atypical femoral fracture), with id., Add. 7 (containing an x-ray image of

55

a microscopic stress fracture). Again, however, Plaintiffs’ argument that Defendant
glossed over this unique feature of atypical femoral fractures in their warning lacks
merit. Significantly, Defendant’s warning explicitly describes “a complete fracture,”
a phrase that appears in the FDA-mandated label as well. The warning also cautions
that such injuries can occur “weeks to months” after “prodromal pain . . . associated
with imaging features of stress fracture.” The term “prodromal” denotes a transitory
phase between the appearance of an initial symptom—i.e., a stress fracture—and the
full development of a condition—i.e., “a complete fracture” of the thigh bone. To that
extent, the warning captures the progression from microscopic fracture to total shaft
fracture that defines the relationship between bisphosphonates and atypical femoral
fractures, and the impact Fosamax may have on this type of fracture overtime. See
Pl. Br., at 32; id., Ex. 3, at A880 (“Fosamax and other [bisphosphonates] can reduce
the body’s ability to repair a stress fracture once it has begun, prior to complete
fracture.”).
Additionally, Plaintiffs focus on another portion of Defendant’s warning:
“stress fractures with similar clinical features also have occurred in patients not
treated with bisphosphonates,” which “threaten[s] to mislead physicians [and the
FDA] about the nature of the relevant risk.” Pl. Br., at 18. As stated supra, I do not
find this misleading, because the statement clarifies and underscores the rarity of
atypical femoral fractures. In any case, according to the Task Force Report, the
“nature of the relevant risk” can include comorbid conditions, which are a Minor
Feature. Likewise, Plaintiffs offer nothing to reconcile their position with the fact

56

that the FDA-mandated warning contains precisely the same statement—“these
fractures also occur in osteoporotic patients who have not been treated with
bisphosphonates.” Def. Br., Ex. 7, at A1516-17. If the FDA warning is adequate, as
Plaintiffs acknowledge, so must be the warning proposed by Defendant in this regard.
Finally, and more compellingly, regardless of any inadequacies in the text of
Defendant’s warning, the FDA clearly understood the type of fracture at issue. In a
June 2008 email titled “Fosamax Information Request – Atypical Fractures,” the
Agency asked Defendant for more data concerning “the occurrence of atypical
fractures.” Pl. Br., Ex. 10, at A1145. Then, in an email from April 2009, the FDA
described Defendant’s PAS as the “currently pending [Supplemental Label Revision]
for atypical fracture,” and stated that it would likely approve “atypical fracture
language” in the “postmarketing adverse events section of the label” only, Def. Br.,
Ex. 3, at A1498, which led to expert testimony at trial concluding that Defendant
“approach[ed] the FDA with respect to atypical femur fractures” in 2008. Def. Br., Ex.
15, at T660:5-8. What is more, the FDA even called the fractures at issue “atypical”
in its CRL, Def. Br., Ex. 2, at A1500; id., Ex. 1, at A2751-52 (defining how the PAS
uses the term stress fracture for the FDA, and distinguishing garden variety stress
fractures), and stated in its October 2010 Safety Announcement that it had been
studying “atypical” fractures all along.
Plaintiffs’ only response to this evidence 17 can be distilled down to a single
point: Defendant did not use the word “atypical” in its proposed warning. Not only is
As discussed more fully, infra, this type of evidence, including email communications,
may be considered by the Court in examining the CRL.

17

57

that a terse and superficial interpretation of the text, but as Judge Pisano observed,
and I agree, “atypical” was hardly settled scientific jargon at the time. Glynn, 951 F.
Supp. 3d at 704 (quoting one of Plaintiffs’ trial experts who was “central” to their
preemption argument, and who said that word was not “contrived” until about 2010).
While this non-material characterization makes Defendant’s warning different from
the FDA-mandated warning, it does not make the warning any less adequate under
state law, nor does it create the inference that Defendant misunderstood or
miscommunicated the underlying science. To the contrary, Defendant’s warning
describes how/when atypical femoral fractures occur (low-energy events in the
absence of trauma), where they occur (to the subtrochanteric and proximal femoral
shaft), their nature (complete fractures), their progression (they develop out of garden
variety stress fractures), and their severity (they can be associated with prodromal
or unusual pain). Indeed, as explained supra, in this context, “atypical” is virtually
synonymous with the term “low-energy” to describe the femoral fractures at issue.
Accordingly, the warning had all the hallmarks of atypical femoral fracture such that
not having employed the word “atypical” would not somehow change the nature of
the proposed warning as plainly expressed by its language.
2. The Language of the CRL
Next, Plaintiffs argue that the CRL, by its terms, rejected Defendant’s proposal
based on language used, not on the fact that the FDA was unconvinced of a causal
relationship between atypical femoral fractures and bisphosphonate. Pl. Br., at 19.
Plaintiffs rely on the text of the letter to make this point, which, in Plaintiffs’ view,

58

does not expressly reference any disagreement with the evidence linking atypical
fractures to bisphosphonates. They also emphasize that Defendant’s litigation
position differs from its own scientists’ “contemporaneous reading of the [CRL].” Id.,
Ex. 29, at A1506; id., Ex. 30, at A1504; id., Ex. 17, at T265:12-18. Specifically,
Plaintiffs point out that the day Defendant received the CRL, its Director of Clinical
Research, Arthur Santora, interpreted it to convey that the “FDA wouldn’t let [Merck]
mention stress fractures.” Id., Ex. 29 at A1506. That same day, Plaintiff highlights
that Defendant’s U.S. Regulatory Liaison, James Adams, informed his colleagues
that the FDA “believes that ‘stress fractures’ may not be clearly related to atypical
subtrochanteric fractures.” Id., Ex. 30 at A1504. According to Plaintiffs, however,
Adams later testified that the CRL does not mention any belief that “there was
insufficient evidence to establish a causal association between Fosamax and atypical
femur fractures.” Id., Ex. 17.
Defendant insists that FDA rejected its proposed warning in the CRL because
“the data was not yet sufficient to allow for [such a warning],” rather than because
the Agency disagreed with Defendant’s wording. Def. Br., at 27. Like Plaintiffs,
Defendant points to the text of the CRL, which states that the “justification” for the
warning was “inadequate.” Defendant reads this as “a commentary on the absence of
a sufficiently clear link between Fosamax and the atypical fractures at issue.” Id.
And, because the CRL rejected Defendant’s warning for “reasons,” plural, the FDA
could not have opposed the “stress fracture” language, alone. Def. Rep. Br., at 9.

59

The CRL was a response to Defendant’s PAS, which, as discussed supra, sought
to include a proposed warning that advised patients of the risk of developing atypical
femoral fractures by taking Fosamax. In that regard, the CRL begins by describing
Defendant’s proposal as “adding language to the PRECAUTIONS section and the
ADVERSE REACTIONS, Post-Marketing Experience subsection of the Package
Inserts (Pls) to describe low energy fractures at the subtrochanteric region of the
femoral shaft. In addition these supplements propose adding language describing
this type of fracture in the Patient Package Insert (PPIs).” Def. Br., Ex. 2, at A1500
(emphasis added).

The FDA rejected Defendant’s proposal for amending the

Precautions section, explaining:
While the Division agrees that atypical and subtrochanteric fractures
should be added to the ADVERSE REACTIONS, Post-Marketing
Experience subsections of the FOSAMAX Tablets and Oral Solution and
FOSAMAX Plus D Tablets labels, your justification for the proposed
PRECAUTIONS section language is inadequate. Identification of
“stress fractures” may not be clearly related to the atypical
subtrochanteric fractures that have been reported in the literature.
Discussion of the risk factors for stress fractures is not warranted and
is not adequately supported by the available literature and postmarketing adverse event reporting.
Id. at A1500-01 (emphasis added).
I appreciate that, as worded, the language of the CRL gives rise to competing
inferences with respect to why the FDA rejected Defendant’s warning. On the one
hand, the CRL describes the “justification” for the warning as “inadequate.” Logically,
the CRL was presumably referencing the data Defendant submitted with its PAS,
linking low-energy femur fractures to bisphosphonates. On the other hand, the CRL
discusses Defendant’s use of the term “stress fracture,” stating that such fractures

60

“may not be clearly related to the atypical . . . fractures that have been reported in
the literature” and it is “not warranted” to discuss risk factors for them. Def. Br., Ex.
2, at A1500-01. In light of these competing readings, I must look beyond the CRL’s
terms alone to ascertain its meaning and scope.
3. The FDA’s Communications from the Same Time Period
If the CRL were the sum total of the evidence of FDA action in this case,
Plaintiffs might be on firmer footing with regards to their preemption arguments.
But, Defendant points to various communications from the FDA during the same time
period to “understand what the FDA action [i.e., CRL] meant.” Def. Rep. Br., at 10.
For instance, in April 2009, a month before the CRL, agency officials wrote that “the
conflicting nature of the literature does not provide a clear path forward” on the
question whether to add a warning to the Precautions section, Def. Br., at 26; Pl. Br.,
Ex. 33, at A1970-71, and “more time [would] be need[ed] for [the] FDA to formulate a
formal opinion on the issue of a precaution around these data.” Pl. Br., Ex. 12, at
A1498; id., Ex. 33, at A1970-71. As stated, supra, the data specifically involves
atypical femoral fractures.
Then, in March 2010, the Agency stated that its review of the data “did not
show an increase in th[e] risk” of atypical femoral fractures from bisphosphonate use.
Def. Br., Ex. 5, A1508. FDA officials did not change their assessment until October
2010, a month after the Task Force issued its Report, id., Ex. 10, at A1118-19; id.,
Ex. 6, at A1392-93, which “clarif[ied] the features of atypical femoral fractures,” Pl.
Br., Ex. 20, at A1392, and “help[ed] the [Agency] understand [them] a little bit better.”

61

Id. at A1396. But even then, the FDA did not observe a definitive causal link. Indeed,
this series of events would not have occurred “if the agency already had a sufficient
basis, in May 2009, to approve a warning” in the Precautions section. Def. Br., at 26.
Neither would the FDA’s own interpretation of the CRL in this litigation: it rejected
Defendant’s warning for “the lack of adequate data to support [it],” not “because of . .
. the term ‘stress fractures.’” FDA Brief as Amicus Curiae, at *31-32. Nevertheless,
Plaintiffs argue that the FDA’s informal email communications are not “Laws” in the
sense of the Supremacy Clause, and in any event, Defendant “ignores the full context
of what [the] FDA told [it]” at the time. Pl. Br., at 25.
Plaintiffs are correct that informal communications do not constitute “Laws”
with the power to preempt. In re Avandia, 945 F.3d at 760 (holding that “an informal
phone conversation with an FDA official is not an ‘agency action taken pursuant to
the FDA’s congressionally delegated authority’”) (quoting Merck, 139 S.Ct. at 1679).
Yet, importantly, Defendant does not argue that the FDA’s informal communications
themselves establish preemption, only that they “shed light on” the meaning and
scope of the CRL, which is “Law” with preemptive effect. Def. Br., at 30. I agree that
it is appropriate to consider the communications for that limited purpose. See, e.g.,
Fosamax, 852 F.3d at 293 (stating that the preemption inquiry involves an
“evaluative inference about human behavior based on correspondence[] [and] agency
statements”); Order at 1, No. 14-1900 (3d Cir. Nov. 25, 2019) (remanding to this Court
“to determine the effect of the FDA’s [CRL] and other communications”); Center for
Drug Evaluation & Research, FDA, CDER 21st Century Review Process: Desk

62

Reference Guide 37 (2014) (explaining that the “[d]evelopment of final labeling” is “an
iterative

process

between

the

applicant

and

FDA”

involving

significant

correspondence); FDA, Guidance for Review Staff and Industry: Good Review
Management Principles and Practices for PDUFA Products 21 (Apr. 2005)
(addressing “communication between the FDA and applicants” during “labeling
discussions”); 21 C.F.R. § 10.85(k) (providing that an FDA employee’s written
statement, which constitutes “an informal communication,” “does not necessarily
represent the formal position of FDA,” a statement that by its terms contemplates
that certain employee statements may do so); In re Incretin, 2021 WL 880316, at *1617 (considered and credited such evidence); Swanson v. Abbott Lab’ys, No. 14-1052,
2017 WL 5903362, at *4 (S.D. Ohio Nov. 28, 2017) (same). Indeed, in Justice Alito’s
concurrence here, he suggested that informal communications between the FDA and
drug manufacturers should be considered in the preemption analysis. Merck, 139
S.Ct. at 1685; see In re Zofran (Ondansetron) Prods. Liab. Litig., 541 F. Supp. 3d 164,
194 (D. Mass. 2021).
Focusing on the sequence of communications and announcements from the
same period, the CRL does not tell the whole story without the proper context gleaned
from other FDA communications. The FDA received data regarding atypical femoral
fractures long before 2008, and specifically sought more information in June 2008 on
“atypical femoral subtrochanteric femur fractures,” a request with which Defendant
complied. Def. Br., Ex. 5, at A1508 (“All available case reports and clinical trial data
were requested.”); Fosamax, 852 F.3d at 296 (“It is undisputed that the FDA was

63

aware of the possible link between Fosamax and atypical fractures well before
September 2010.”). Defendant proposed amending both the Precautions and Adverse
Reactions sections of the Fosamax label in September 2008, to include an appropriate
warning about atypical femoral fractures, which was “important” to do given their
clinical significance, even if it “was not possible with the present data” to establish
causation, and even if the FDA was in the process of reviewing the issue. Pl. Br., Ex.
38, at A1349; Def. Br., Ex. 1, at A2756. The FDA rejected Defendant’s Precautions
warning in May 2009. In correspondence before sending the CRL, agency officials
stated that the “conflicting nature of the [scientific] literature does not provide a clear
path forward” on a Precautions warning, “more time [would] be need[ed] for [the]
FDA to formulate a formal opinion on the issue of a precaution” based on the data,
and Defendant’s “elevation of [the warning] to a precaution” was “prolonging
review.” 18 Pl. Br., Ex. 33, at A1971. Then, after sending the CRL, the FDA expressed
no desire to consider revisions despite Defendant’s repeated inquiries to that end.
As late as March 2010, the FDA continued to believe that the “available” data
“did not show an increase in th[e] risk” of atypical femoral fractures from
bisphosphonate use, instructed doctors to “continue to follow” the existing Fosamax

18
As explained supra, the FDA clearly understood Defendant to be warning of the injury
discussed in the literature, which is the same injury as Plaintiffs allegedly suffered. See, e.g.,
Pl. Br., Ex. 10, at A1145 (asking Defendant for more data concerning “the occurrence of
atypical fractures” in a June 2009 email titled “Fosamax Information Request – Atypical
Fractures”); Def. Br., Ex. 3, at A1498 (describing Defendant’s PAS as the “currently pending
[Supplemental Label Revision] for atypical fracture” in an April 2009 email); Def. Br., Ex. 15,
at T660:5-8 (Plaintiff’s expert, Dr. Cheryl Blume, testified that Defendant “approach[ed] the
FDA with respect to atypical femur fractures”); Def. Br., Ex. 2, at A1500 (calling the fractures
at issue “atypical” in the CRL); Def. Br., Ex. 1, at A2752-52 (defining, for the FDA, how the
PAS uses the term stress fracture).

64

label, and specifically noted a December 2008 study showing “similar numbers of
atypical subtrochanteric femur fractures relative to classic osteoporosis” in patients
not treated with bisphosphonates. The FDA made this Drug Safety Announcement
pursuant to its Congressionally delegated authority. 21 U.S.C. § 355(r). The FDA also
stated that it was “working closely with outside experts to gather additional
information that may provide more insight.” Def. Br., Ex. 5, at A1508. Construed in
light of these various FDA communications, the CRL clearly rejected Defendant’s
warning, in part, because the FDA doubted the underlying science causally
connecting bisphosphonate use and atypical femoral fractures.
It is also telling that, in the process of rejecting Defendant’s Precautions
warning, the FDA approved an Adverse Reactions warning for “low-energy femoral
shaft and subtrochanteric fractures.” The reason for the Agency’s decision in this
regard may very well be the different causal thresholds governing each section of the
label. Indeed, the Precautions section requires “reasonable evidence of a causal
association” to add a warning about an adverse event. 21 C.F.R. § 201.57(c)(6)
(emphasis added). The Adverse Reactions section requires only “some basis to believe
there is a causal relationship.” 21 C.F.R. § 201.57(c)(7) (emphasis added).
Finally, the FDA, itself, believes that it rejected Defendant’s warning for “the
lack of adequate data to support [it],” and not “because of . . . the term ‘stress
fractures.’” FDA Brief as Amicus Curiae, at *31-32. Plaintiffs challenge this evidence
because it is a “legal interpretation[] . . . submitted by government lawyers under a
subsequent administration, nearly a decade after the fact,” which represents the

65

views of the Office of the Solicitor General not the FDA. Pl. Br., at 22. I disagree on
both points. First, “[I] have no reason to suspect that the Solicitor General’s
representation of [the FDA’s] views reflects anything other than ‘the agency’s fair and
considered judgment.’” Geier, 529 U.S. at 884. Second, an agency’s fair and considered
judgment as to the meaning of its own regulation and actions deserves some measure
of deference. Auer v. Robbins, 519 U.S. 452, 461-62 (1997).
On the first point, it is appropriate to consider the FDA’s views because
Congress delegated to that agency the authority to implement federal drug
regulations, it has expertise in that highly “technical” subject matter, and it is wellequipped to navigate “the relevant history and background” on such a “complex and
extensive” issue. Geier, 529 U.S. at 883 (giving “some weight” to agency’s view in a
preemption case on similar grounds). Or, stated differently, the FDA is “likely to have
a thorough understanding of its own regulation and objectives” with respect to any
CRL it issues. Id.; Medtronic, 518 U.S. at 496 (relying, in part, on the FDA’s
interpretation of a provision’s preemptive effect).
On the second point, I am not strictly foreclosed from crediting the FDA’s
reading of the CRL simply because the Agency advances it in litigation, particularly
in light of all the other pertinent evidence. I am aware that in Kisor v. White, 139
S.Ct. 2400 (2019), the Supreme Court warned that “a court should decline to defer to
a merely convenient litigation position or post-hoc rationalization advanced to defend
past agency action against attack,” such as a brand-new interpretation presented for
the first time in legal briefs. Id. at 2417-18 (quotations and alterations omitted); see

66

also Christopher v. SmithKline Beecham Corp., 567 U.S. 142, 155 (2012); Bowen v.
Georgetown Univ. Hospital, 488 U.S. 204, 213 (1988). But, Kisor sets forth “[t[he
general rule,” not an “entirely foreclosed . . . practice.” Kisor, 139 S.Ct. at 2417 n.6.
For example, the Auer Court deferred to a “new regulatory interpretation presented
in an amicus curiae brief in [the Supreme Court].” Id. Where the agency is not a party
to the litigation and has expressed its views only at the Court’s invitation, as here,
there is no reason to question whether it has provided its “fair and considered
judgment” rather than an after-the-fact rationalization. Id. (citing Auer, 519 U.S. at
462).
In sum, when viewed in light the FDA’s communications, the CRL rejected
Defendant’s Precautions warning because the FDA doubted the evidence linking
bisphosphonate use to atypical femoral fractures in a causal sense. In other words,
when it issued the CRL, the Agency believed that Fosamax’s current label adequately
reflected the results of the Agency’s continuous and comprehensive evaluation of the
risks associated with using Fosamax.
4. The Regulatory Regime
Finally, the parties use the regulatory regime, indeed some of the same
provisions, to draw opposite inferences as to the meaning of the CRL. This is “highly
relevant” and bears discussion. Merck, 139 S.Ct. at 1685 (Alito, J., concurring in the
judgment) (“On remand, I assume that the Court of Appeals will consider the effect
of [21 U.S.C. § 355(o)(4)(A)].”); Seufert, 187 F. Supp. 3d at 1175 (“[A] clear evidence

67

analysis must account for the regulatory framework governing prescription drug
labeling.”).
Plaintiffs argue that it is “dispositive” that the FDA “omi[tted]” any explicit
discussion of the science linking bisphosphonates to atypical femoral fractures,
because the agency must “describe all of the specific deficiencies that [it] has
identified” when it sends a CRL. Pl. Br., at 20-21; 21 C.F.R. § 314.110(a)(1). To the
extent that the FDA did not specifically raise causation as an issue, it cannot form
any part of the basis for the agency’s rejection, unless the agency “wrote a false
[letter].” Pl. Br., at 20.
Defendant reads the regulations differently. Def. Br., at 25-27; Def. Rep. Br.,
at 11-12. According to Defendant, the FDA has a duty to mandate a label change if
it “becomes aware of new information” that “should be included in the labeling.” 21
U.S.C. § 355(o)(4)(A). 19 Defendant reasons that the FDA did not do so until October
2010, which implies that the FDA could not support a change before then and/or
concluded that the Fosamax label conveyed the proper risk profile to the public at the
time. Likewise, according to Defendant, the FDA will not reject a warning for
“editorial” reasons, 21 C.F.R. § 314.105(b), and will “make every reasonable effort to
communicate” any “easily correctable deficiencies” to a manufacturer “promptly,” 21
C.F.R. § 314.102(b), including by suggesting remedies or recommending actions. 21
C.F.R. § 314.110(a). In light of these provisions, Defendant submits that based on the
19
Prior to October 2018, § 355(o)(4)(A)’s language contained slight differences not
relevant here. See Substance Use-Disorder Prevention That Promotes Opioid Recovery and
Treatment for Patients and Communities Act, Pub. L. 115-271, § 3041(b), 132 Stat. 39423943, effective Oct. 24, 2018.

68

FDA’s statutory obligations, “[h]ad the Agency believed a [Precautions] change was
justified earlier,” or that the problem with Defendant’s warning was fixable, such as
Defendant’s “stress fractures” language, “it would have taken [the necessary] steps,”
Def. Br., at 26, similar to the steps the FDA took as to Defendant’s Adverse Reactions
warning, Def. Br., Ex. 1, at A2732, and when Defendant, again, proposed stress
fractures language in December 2010 in response to the Agency’s mandated label.
First, while the CRL did not use certain terminology, which would have made
it less ambiguous, this Court has found, see supra, that the CRL did in fact reject
Defendant’s proposed warning based on causation, and therefore, Plaintiff’s
argument in this context must be rejected out of hand. But, even if I were to accept
Plaintiffs’ position, one must assume that the FDA had reasonable evidence
warranting a Precautions warning, but was so troubled by Defendant’s use of the
term “stress fracture” that it rejected a warning without offering any suggestions or
revisions. To make such an assumption would effectively overlook the FDA’s raison
d’etre to regulate drug safety, its independent legal duty to notify a manufacturer as
soon as it “becomes aware of new safety information that [it] believes should be
included in the labeling of a drug” and “initiate discussions to reach an agreement . .
. on labeling,” 21 U.S.C. § 355(o)(4)(A), and the “presumption of regularity”
accompanying its actions. Rather, “in the absence of clear evidence to the contrary,
[FDA officials] have properly discharged their official duties.” United States v.
Chemical Foundation, Inc., 272 U.S. 1, 14-15, (1926) (quoted in Merck, 139 S.Ct. at
1684 (Alito, J., concurring in part)). In other words, it is improbable that the FDA

69

declined to approve Defendant’s Precautions warning, or failed to propose a solution
to the problem it perceived with the language, i.e., stress fracture, all while the FDA
had sufficient causal evidence linking bisphosphonates to atypical femoral fractures
and thus exposing patients to the risk of severe injury in the interim. Accord Zofran,
2021 WL 2209871, at *32 (“[T]he Court will not assume that the FDA failed to
perform, in fact blatantly ignored, its statutory duties to review and monitor the drug
for human safety . . . . Accepting plaintiffs’ argument would suggest that the FDA . .
. turned a blind eye to evidence that Zofran causes birth defects. That is highly
unlikely, to say the least. And it is also unlikely . . . that it refused to take up the
issue with Novartis based on the technical point that Novartis had not sought to
change that specific section.”).
The more likely scenario is that the FDA’s actions taken in this case convey
doubts that the Agency had about the underlying science, a deficiency no revision or
edits could solve; hence, the Agency did not propose any. The FDA’s subsequent
inaction—it did not mandate a label change until October 2010, despite substantial
ongoing review both internally and by the Task Force—confirms its then-existing
perspective on the science, not that it was merely troubled by Defendant’s
phraseology of its proposed warning. Ridinis v. Maurince, No. 15-00020, 2020 WL
1264178, at *21 (W.D. Mo. Mar. 16, 2020) (holding that “continued inaction . . . in
light of the known issues and the ongoing give-and-take between [a manufacturer]
and the FDA” can constitute “clear evidence”); In re Incretin-Based Therapies Prod.
Liab. Litig., 142 F. Supp. 3d 1108, 1123-24 (S.D. Cal. 2015) (“The FDA’s subsequent

70

inaction regarding drug labeling supports the conclusion that the FDA [did] not
consider available scientific evidence of a causal association sufficient to warrant
inclusion in the labeling . . . . [This] is highly persuasive given the FDA’s
comprehensive review of pancreatic safety and ability to mandate a label change.”);
see also Merck, 139 S.Ct. at 1685 (Alito, J., concurring in the judgment) (implying that
FDA inactions in light of “[its § 355(o)(4)(A)] duty arguably affect the pre-emption
analysis”).
More to the point, the FDA “communicate[s] with applicants about scientific,
medical, and procedural issues that arise” when it reviews a request for regulatory
action. 21 C.F.R. § 314.102(a). More specifically, § 314.110(a)(2) imposes a “complete
description” requirement when the agency sends a CRL; § 355(o)(4)(A) imposes an
“obligation to initiate a label change” if the FDA believes one is warranted; §
314.110(a)(3) states that a CRL reflects the “FDA’s complete review of the data
submitted,” not merely the particular labeling language proposed; and under §
314.105(b), the FDA may approve an application with “minor deficiencies” contingent
on appropriate corrections. Taken together, these provisions warrant the following
inference as to the FDA’s intention when it issued the CRL: the Agency did not believe
there was reasonable scientific evidence of a causal association between
bisphosphonate use and atypical femoral fractures, or else it would have suggested
edits to that end, or simply mandated a warning using language that the FDA
thought was more appropriate, similar to what the Agency ultimately did in 2010.

71

What is more, the FDA red-lined Defendant’s proposed “stress fractures”
language between October 2010, when the Agency imposed a label change, and
January 2011, when Defendant implemented the Agency’s Precautions warning aswritten. Were such language the sole problem with the 2008 warning, then the FDA
could have simply stricken it, as it did two years later, or approved it on the condition
that Defendant implement edits pursuant to 21 C.F.R. § 314.105(b). But, an issue
existed in 2008 that did not exist in 2010, one that could not be resolved with any
revisions: in 2008, the FDA was unconvinced of the causal link between
bisphosphonate use and atypical femoral fractures. The Agency’s contrasting
approaches to Defendant’s proposed “stress fractures” language cannot be reconciled
otherwise. Accordingly, it follows from the regulatory regime that the FDA rejected
Defendant’s warning for lack of reasonable evidence of causation.
iii.

The Scope of the CRL

Having determined the context of the CRL, I next determine the FDA’s likely
response to another proposed warning based on how it did respond in the CRL. See,
e.g., Fosamax, 852 F.3d at 293 (stating that the preemption inquiry requires
“pars[ing]” the FDA’s actual response “to discern what it suggests about the FDA’s
likely response”); Dobbs v. Wyeth Pharmaceuticals, 797 F. Supp. 2d 1264, 1277 (W.D.
Ok. 2011) (finding preemption even though the FDA later determined that sufficient
evidence existed to justify a warning, in part because it was “highly persuasive” that
the FDA rejected a similar warning before).

72

1. A Revised PAS
Plaintiffs first argue that the FDA would have approved a differently worded
Precautions warning had Defendant simply removed the “stress fractures” language
and resubmitted its PAS. Pl. Br., at 29 (“FDA invited further action from [Defendant]
on at least four occasions, over several months, in various formats (email, formal
letter, telephone call). Thus, ‘the ball was back in [Defendant’s] court to submit a
revised, corrected proposal.’”) (quoting Fosamax, 852 F.3d at 299).
Judge Pisano disagreed, finding that “the FDA would have rejected a stronger
Precautions warning because the FDA did reject a stronger Precautions warning” in
the CRL. OTSC Opinion, 2014 WL 1266944, at *16. Construed in light of the FDA’s
communications and the regulations governing prescription drug labeling, I also find
that: the CRL denied Defendant’s Precautions warning because the FDA doubted the
causal connection, if any, between bisphosphonates and low-energy femur fractures,
and to that extent, the letter foreclosed the possibility that the FDA would have
approved a differently worded warning in a revised PAS, without any substantial
change in science.
Plaintiffs’ evidence to the contrary is unavailing. Indeed, while the FDA
mentioned working with Defendant in April 2009 “to decide on language” for a
warning in the Precautions section, the Agency conditioned that response with
explicit language that only “if it [was] warranted,” an important qualification
signaling the uncertain state of the underlying science. In fact, in the same email, the
FDA instructed Defendant to “hold off” on a Precautions warning, which was

73

“prolonging review,” so that it could “close out” its PAS and “agree quickly” to changes
in the Adverse Reactions section, another sign that the Agency was not prepared to
approve any revised Precautions language.
Plaintiffs also point to various interactions between May and July 2009,
purporting to show that Defendant declined to engage with the FDA after the CRL
was issued. Pl. Br., at 28-29. Each interaction, put into context, however, is an
attempt by Defendant to initiate further discussion, which the Agency rebuffed. For
instance, in a June 2009 phone call, Defendant asked for “a teleconference” to discuss
revisions to its Precautions warning. Pl. Br., Ex. 13 (“I asked . . . would the Division
be open to a teleconference to discuss what may be acceptable.”). But the FDA
responded that Defendant must “formally” request one. Id. (“[The FDA official]
replied such requests should be made formally through a submission to the file.”).
Soon thereafter, by email, Defendant reiterated its desire to discuss a
Precautions warning. Id., Ex. 14 (“Per your recommendation from a previous
conversation this [potential] meeting would be requested formally as a Type C
meeting.”). Days later, however, the FDA informed Defendant in another phone call
that it must “address both issues high-lighted in the [CRL] to initiate a new review
cycle . . . or . . . withdraw the previous PAS.” Id., Ex. 15 (emphasis added); 21 C.F.R.
§ 314.110(b) (requiring a drug manufacturer to “address[] all deficiencies identified”
by the FDA if it chooses to resubmit its application). One issue was the “inadequate
justification” for the warning, which embodied the FDA’s then-existing skepticism on

74

causation. Defendant received the same response when it asked the Agency to keep
its PAS open pending further discussion.
Likewise, in its Adverse Reactions CBE amendment in July 2009, Defendant
stated that it “still believe[d]” in a Precautions warning about “low-energy fracture[s]”
and anticipated requesting a formal meeting on that issue per the FDA’s prior
instructions. Id., Ex. 16. Defendant never did so, and Plaintiffs demand an adverse
inference for it. But, Plaintiffs overlook the fact that withdrawal is a lawful response
to a CRL. 21 U.S.C. § 314.70(c)(6)(iii)(A). Moreover, a formal meeting is not a
prerequisite to preemption. Dolin, 901 F.3d at 814 (rejecting plaintiff’s argument that
defendant’s failure to request a formal meeting with the FDA after receiving a CRL
barred preemption, which “misunderstands the preemption standard”); see also
PLIVA, 564 U.S. at 619-20 (rejecting plaintiff’s argument that defendant’s failure to
ask the FDA to change the brand-name label barred preemption for a generic
manufacturer, because what matters is that the manufacturer “cannot independently
satisfy its state duties without the Federal Government’s special permission and
assistance”).
Plaintiffs’ argument is essentially that Defendant could have, perhaps,
theoretically, changed the FDA’s decision had Defendant somehow insisted on
engaging with the Agency or invoked an available procedural mechanism rather than
withdraw its PAS, but “the possibility of [that] possibility” is certainly not enough to
“defeat[] pre-emption.” PLIVA, 564 U.S. at 626 n.8; cf. In re Actos (Pioglitazone) Prod.
Liab. Litig., No. 11-2299, 2014 WL 4364832, at *20 (W.D. La. Sept. 2, 2014) (rejecting

75

manufacturer’s preemption defense because of substantial evidence that the
manufacturer declined various FDA efforts to require a stronger warning); Dorsett v.
Sandoz, Inc., 699 F. Supp. 2d 1142, 1159 (S.D. Cal. 2010) (“Defendants offer nothing
but theoretical assumptions of what the FDA would have done, and that is not enough
to warrant a finding of preemption.”). Indeed, Plaintiffs have presented no evidence
that the FDA made any suggestions, at the time it issued the CRL, that it would
somehow change its decision regarding the proposed warning if Defendant made
certain changes. Rather, the Agency rejected the warning based on a lack of scientific
evidence, and it follows that the FDA would not have approved a Precautions warning
had Defendant simply omitted the “stress fractures” language and resubmitted its
PAS.
2. A CBE Amendment
Plaintiffs also suggest that Defendant, on its own initiative, could have
amended the Precautions section of the Fosamax label through a CBE amendment.
The CBE process permits a drug manufacturer to unilaterally add a Precautions
warning to its label, but only if “newly acquired information” provides “reasonable
evidence of a causal association of a clinically significant adverse reaction linked to a
drug.” 21 C.F.R. §§ 314.70(c)(6)(iii), 201.57(c)(6)(i). The question of whether newly
acquired information exists is fact-intensive, but because it is “part and parcel of the
broader legal [preemption] question,” Merck, 139 S.Ct. at 1680, it is incumbent upon
this Court to decide. Lyons v. Boehringer Ingelheim Pharmaceuticals, Inc., 2020 WL
5835125, at *8 (N.D. Ga. Sept. 29, 2020) (collecting cases holding same).

76

“Newly acquired information” can take many forms. Information previously
known to a manufacturer, but not submitted to the FDA, may suffice,73 Fed. Reg. at
49,606, as well as “data derived from new clinical studies, reports of adverse events,
or new analyses of previously submitted data (e.g., meta-analyses) if the studies,
events, or analyses reveal risks of a different type or greater severity or frequency than
previously included in submissions to FDA.” Wyeth, 555 U.S. at 569 (emphasis added);
21 C.F.R. § 314.3(b). This “accounts for the fact that risk information accumulates
over time and that the same data may take on a different meaning in light of
subsequent developments.” Wyeth, 555 U.S. at 569. Notably, however, the CBE
process does not exempt the proposed change from the FDA’s substantive
requirements, nor does it eliminate FDA jurisdiction—two points that Plaintiffs
acknowledge. Indeed, the FDA retains authority to review amendments submitted
through the CBE process, and it will reject a CBE amendment if, among other things,
it concludes that there is insufficient evidence of a link between the drug and the
adverse event or the proposed change “requires approval prior to distribution.” 73
Fed. Reg. 2848, 2851; 21 C.F.R. 314.70(c)(5)(i); see also id. 314.70(c)(7) (“If the agency
disapproves the [CBE], it may order the manufacturer to cease distribution of the
drug product(s) made with the manufacturing change.”); 71 Fed. Reg. 3922, 3934
(“FDA reviews all such submissions and may later deny approval of” a CBE; “[t]hus,
in practice, manufacturers typically consult with FDA prior to adding risk
information to labeling”). Case law also highlights this important characteristic of
the CBE process. See Wyeth, 555 U.S. at 571 (“Of course, the FDA retains authority

77

to reject labeling changes made pursuant to the CBE regulation in its review of the
manufacturer’s supplemental application, just as it retains such authority in
reviewing all supplemental applications.”); Mason v. SmithKline Beecham Corp., 596
F.3d 387, 392 (7th Cir. 2010) (“It is technically a violation of federal law to propose a
CBE that is not based on reasonable evidence.”).
Here, as of March 2010, the fact that the FDA still believed that “reasonable
evidence of a causal association” was lacking and that it rejected Merck’s proposed
Precaution in 2009, demonstrate that it would not have approved the same change
by way of a CBE amendment. See Dobbs, 797 F. Supp. 2d at 1276 (noting that the
FDA had rejected risk information added by a CBE amendment because “it did not
believe that a causal association” between the drug and the purported risk “has as
yet been definitively established”). Indeed, as a matter of procedure, in order for
Defendant to proceed with the CBE process after the FDA rejected its PAS—Merck
was required to produce data indicating a greater-than-previously-known risk of
atypical femoral fractures, which could establish “reasonable evidence” of a causal
association. Drescher v. Bracco Diagnostics, Inc., 2020 WL 699878, at *4 (D. Az. Jan.
31, 2020) (examining “whether Plaintiff has pled reasonable evidence of a causal
association sufficient to allow a CBE label change”); Dobbs, 797 F. Supp. 2d at
1272 (“The FDA has consistently defined reasonable evidence . . . as ‘when evidence
exists on the basis of which experts qualified by scientific training and experience can
reasonably conclude that the hazard is associated with the use of the drug.’”); 44 Fed.
Reg. 2848, 2851 (allowing a CBE amendment only for “known hazards and not

78

theoretical possibility”); id. at 49,604 (stating that this is how the FDA ensures
“scientifically accurate information appears in the approved labeling”); see also
Gibbons v. Bristol-Myers Squibb Co., 919 F.3d 699, 707 (2d Cir. 2019) (holding that
manufacturers are “limited in their ability to unilaterally change the labels on their
products” because they must comply with the CBE regulation’s causation thresholds);
Merck, 139 S.Ct. at 1677 (explaining that, “when the risks of a particular drug
become apparent, the manufacturer has a duty to provide a warning that adequately
describes the risk”) (emphasis added).
Certainly, no additional data between the period of time when the FDA issued
its CRL and when the Agency finally decided to issue a Precautions warning,
“reveal[ed] risks of a different type or greater severity or frequency” than the ones
which Defendant knew, informed the FDA, and sought to warn against in the first
instance. 21 C.F.R. § 314.3(b). Defendant submitted its PAS in September 2008,
concluding based on the research at the time that “[i]t is not possible with the present
data to establish whether” Fosamax “increases the risk of . . . low-energy
subtrochanteric and/or proximal shaft fractures.” Pl. Br., Ex. 38, A1349. And, while
Plaintiffs point to certain unidentified and unspecified case studies and articles,
which purportedly demonstrate a different risk profile for Fosamax with respect to
atypical femoral fractures that were available to Defendant between submission of
its PAS in September 2008 and the Task Force Report in October 2010, those case
studies and articles have neither been provided to the Court, nor summarized. Thus,
the Court cannot evaluate the conclusions reached by those articles and case studies,

79

nor can it even definitively determine whether Merck ever independently reviewed
or provided those materials to the FDA.
Moreover, even if those articles and case studies existed, in March 2010, the
FDA announced that it had not seen “a clear connection between bisphosphonate use
and a risk of atypical subtrochanteric femur fractures,” “an increase in the risk in
women using [bisphosphonates],” or “[different] numbers of atypical subtrochanteric
femur fractures” in “patients taking bisphosphonates” versus patients “not taking
bisphosphonates” after reviewing case reports and clinical trial data from all
bisphosphonate drug manufacturers. Pl. Br., Ex. 18, at A1508. These conclusions
mirror those in Defendant’s PAS, and to that extent, do not shed light on any “newly
acquired information” in the sense of the CBE regulation. In other words, even though
the FDA’s then-ongoing review was arguably more thorough than any review it might
have conducted under the CBE process—the Agency was compiling data from
multiple manufacturers, analyzing a variety of new reports, revisiting old ones,
conducting its own analyses, and working with outside experts on the Task Force—it
did not uncover definitive evidence linking Fosamax use to atypical femoral fractures
to a greater extent than Defendant originally indicated.
Then, in September 2010, the Task Force published its Report, which
developed a “provisional case definition” for the “features for complete and incomplete
atypical [femoral] fractures,” reassessed prior studies in light of that definition, and
reviewed a number of new articles/reports that Defendant had not previously
submitted to the FDA, but added nothing not already known. Shane et al., at 2267-

80

69. Still, according to the Report, “a causal association between [bisphosphonates]
and atypical fractures ha[d] not been established.” Id. The science merely supported
“evidence of a relationship between long-term [bisphosphonate] use and a specific
type of subtrochanteric and femoral shaft fracture.” Pl. Br., Ex. 2, at A1167 (emphasis
added). Defendant and the FDA, alike, had long recognized the same. See, e.g., Pl.
Br., Ex. 10, at A1145 (stating, in June 2008, that the Agency was “aware of reports
regarding the occurrence of subtrochanteric hip fractures in patients using
bisphosphonates,” that these were “reportedly rare in patients with osteoporosis not
on bisphosphonates,” and that it was “concerned about this developing safety signal”);
Fosamax, 852 F.3d at 275 (citing A1258) (forwarding an article stating that Fosamax
“may . . . potentially” increase the risk of such fractures); id. (citing A1237)
(forwarding an article stating that Fosamax “may be associated” with such
fractures”); id. (citing A1243) (forwarding an article stating that certain findings
“raise[d] the possibility” that Fosamax may lead to such fractures).
Given the conclusions in the Task Force Report, there was no “newly acquired
information” as defined in the CBE regulation on the basis of which Defendant could
have successfully submitted a CBE amendment. Accord In re Lipitor (Atorvastatin
Calcium) Marketing, Sales Practices & Prod. Liab. Litig., 185 F. Supp. 3d 761, 769
(D.S.C. 2016) (holding that a drug label cannot be changed based solely on
“information previously submitted to the FDA”); Dolin, 901 F.3d at 816 (“The [2011]
article contained the same figures as GSK’s 2006 analysis, which GSK submitted to
the FDA. There is no basis to conclude that this was a new analysis or that it was ‘not

81

previously submitted to the Agency.’”); Knight, 984 F.3d at 339 (explaining how a new
article showing a “correlation” is insufficient to defeat preemption because the FDA
already knew that); In re Incretin, 142 F. Supp. 3d at 1123 (stating that
“indeterminate” or “inconclusive” evidence is not “reasonable evidence” sufficient to
justify a CBE amendment); Seufert, 187 F. Supp. 3d at 1175 (stating that a CBE
amendment “demands more than an indeterminate or inconclusive relationship”);
McGrath, 393 F. Supp. 3d at 168 (“For the Court to draw the reasonable inference
that Bayer could have unilaterally amended the Magnevist label in compliance with
the FDA’s CBE regulation, the Complaint must plead more than the mere possibility
that Magnevist caused Plaintiff’s fibrosis and related injuries.”). The “new” evidence
published after Defendant submitted its PAS, and relied upon by Plaintiffs,
established the very relationship or connection Defendant had identified all along. 20
The FDA responded to the Task Force by issuing another Drug Safety
Announcement in September 2010, but with the same conclusion as before: the
Report would “facilitate future studies” assessing a causal link between “these

20
The Task Force Report also “suggest[ed] that the risk rises with increasing duration
of exposure.” Id. But this was not new information to the FDA either. Pl. Br., Ex. 2, at A1147
(“The duration of [bisphosphonate use] relative to onset of the fracture was 5.3 years mean
and 5 years median with a range from 3 months to 14 years.”); Merck, 139 S.Ct. at 1674
(describing a report from a doctor in 2002 stating that his hospital called atypical femoral
fractures the “Fosamax Fracture” because “100% of patients in his practice who have
experienced femoral fractures . . . were taking Fosamax . . . for over 5 years”) (emphasis
added); Merck, 139 S.Ct. at 1674 (“[Defendant began receiving adverse event reports from
the medical community indicating that long-term Fosamax users were suffering
atypical femoral fractures.”) (emphasis added); Fosamax, 852 F.3d at 275 (citing A1258,
A1237, A1243) (describing how Defendant “began to see numerous scholarly articles and case
studies documenting possible connections between long-term Fosamax use and atypical
femoral fractures”) (emphasis added).

82

unusual femur fractures” and bisphosphonate use, but “it is not clear if
bisphosphonates are the cause [of such fractures].” Def. Br., Ex. 9, at A1512. This,
too, echoes Defendant’s original assessment of the science/evidence and implies no
new risks or correlations of which the FDA was not already aware. McGrath, 2019
WL 2582530, at *5 (“Studies concluding it ‘remains unknown whether GBCAs induce
toxic effects’ and that ‘further studies are required to address possible clinical
consequences of gadolinium deposition . . . in patients with normal renal function’ do
not constitute reasonable or well-grounded scientific evidence of ‘clinically significant
adverse effects’ under the CBE regulation.”); Utts v. Bristol-Myers Squibb Co., 251 F.
Supp. 3d 644, 669 (S.D.N.Y. 2017) (same, but with respect to articles that “merely
express a desire for further investigation”), aff’d sub nom. Gibbons v. Bristol-Myers
Squibb Co., 919 F.3d 699 (2d Cir. 2019).
Finally, in October 2010, the FDA mandated a change to the Fosamax label,
yet again it rejected any causal link, which is “squarely in line” with its prior
conclusions and Defendant’s ongoing dialogue with the Agency. Lyons, 491 F. Supp.
3d at 1364. The Task Force Report merely made the FDA “confident” that atypical
femur fractures are “potentially more closely related to” long-term use of
bisphosphonates “than [the Agency] previously had evidence for.” Def. Br., Ex. 6, at
A1392-93 (emphasis added). The now-current Fosamax label, as written by the FDA,
refuses to go any further than Defendant’s proposal thirteen years ago: “Causality
has not been established as these fractures also occur in osteoporotic patients who
have not been treated with bisphosphonates.” Accord Matrixx Initiatives, Inc. v.

83

Siracusano, 563 U.S. 27, 44 (2011) (“The fact that a user of a drug has suffered an
adverse event, standing alone, does not mean that the drug caused that event.”).
In any event, the FDA’s review of Defendant’s CBE amendment would not have
been any less rigorous than its review of Defendant’s PAS, particularly since the FDA
was conducting its own review of causation at the time when Defendant had the
opportunity to submit a CBE amendment, and Defendant’s view of the scientific
evidence would not have been entitled to extra (or any) deference. Accord In re
Incretin, 142 F. Supp. 3d at 1125. In fact, drug manufacturers almost always consult
with the FDA before submitting CBE amendments to avoid future enforcement action
for an unwarranted warning. Id. There was much correspondence between the FDA
and Defendant here, none of which indicated that the Agency would permit
Defendant to implement a Precautions warning through the CBE process, but not
through the PAS process. In fact, and importantly, the FDA suggested that Defendant
submit a label change for the Adverse Reactions section through the CBE process,
but it did not make the same recommendation for the Precautions warning. Based on
these FDA communications, it is difficult to imagine that Defendant could have
successfully changed the Fosamax label through the CBE regulation after the FDA
rejected its PAS.
Contrary to Plaintiffs’ position, while Defendant of course could have tried to
submit a CBE amendment, regardless of futility, Merck, 139 S.Ct. at 1975; Pl. Br.,
Ex. 17, at T181:23-182:12, it need not do so merely to preserve its preemption defense.
A manufacturer is under no obligation to use the CBE process to change the

84

Precautions section of its label for any reason other than reasonable evidence of a
causal association. Wyeth, 555 U.S. at 571 (cautioning that the mere availability of a
CBE amendment does not defeat preemption); PLIVA, 564 U.S. at 628 n.8 (noting
that “the possibility of possibility” is not enough); Dolin, 951 F.3d at 890-91
(explaining how the phrase “would not have approved” in Wyeth implies that a drug
manufacturer may prove preemption without showing that it ever attempted to make
a label change); Cerveny, 783 Fed. App’x. at 804 n.8 (rejecting notion that “only
labeling changes sought by the manufacturer can lead to preemption”); Cerveny, 155
F. Supp. 3d at 1213-16 (explaining that lower courts have “universally rejected” the
notion “that [Wyeth] requires a showing that the manufacturer attempted to apply
the warning suggested by the plaintiff but that the labeling was ultimately rejected
by the FDA”); In re Incretin, 142 F. Supp. 3d at 1126 (“[Wyeth] does not require CBE
submission and rejection.”); Zofran, 2021 WL 2209871, at *32 (“Multiple courts have
found preemption where the manufacturer had not requested the precise warning
sought by the plaintiffs when the FDA had nonetheless made it clear that it would
not accept that label change.”).
A contrary rule would incentivize manufacturers to submit a CBE amendment
regardless of risk magnitude or scientific justification, which would impose an undue
burden on the FDA. Wyeth, 555 U.S. at 578-79 & n.11 (“[The] FDA has limited
resources to monitor the 11,000 drugs on the market.”); Seufert, 187 F. Supp. 3d at
1175 (“A rule to the contrary would encourage prophylactic labeling changes by
manufacturers, which, in turn, could inundate the FDA with labeling submissions.”);

85

FDA, FDAAA Implementation – Highlights Two Years After Enactment 7 (2010)
(finding just 363 CBE amendments between 2009 and 2010). Not to mention that “[i]t
is technically a violation of federal law to propose a CBE that is not based on
reasonable evidence.” Mason, 596 F.3d at 392; Drescher, 2020 WL 699878, at *4
(concluding same).
Moreover, the FDA does not approve CBE amendments simply out of an
abundance of caution, as Plaintiffs seem to suggest. The Agency regulates drug labels
for precisely the opposite reason: so as not to “cause meaningful risk information to
lose its significance.” 73 Fed. Reg. 2848, 2851 (“Exaggeration of risk, or inclusion of
speculative or hypothetical risks, could discourage appropriate use of a beneficial
drug . . . or decrease the usefulness and accessibility of important information by
diluting or obscuring it.”). Indeed, “[w]hile it is important for a manufacturer to warn
of potential side effects, it is equally important that it not overwarn because
overwarning can deter potentially beneficial uses of the drug by making it seem
riskier than warranted and can dilute the effectiveness of valid warnings.” Mason,
596 F.3d at 392. The FDA is thus appropriately wary of “the resulting information
overload [which] would make label warnings worthless to consumers.” Robinson v.
McNeil Consumer Healthcare, 615 F.3d 861, 870-71 (7th Cir. 2010); Muzichuck v.
Forest Laboratories, 2015 WL 235226, at *7 n.2 (N.D.W.V. Jan. 16, 2015) (“Public
policy recognizes a danger in ‘overwarning’ consumers of potential drug-related
risks.”). Accordingly, Defendant could not have met the relevant CBE criteria had it

86

submitted a Precautions warning through that regulation after the FDA rejected its
PAS.
CONCLUSION
Based on clear and convincing evidence, the Court finds that Defendant fully
informed the FDA of the justifications for its proposed warning, which was adequate
under state law and encompassed the injury Plaintiffs allege here. The FDA, in turn,
informed Defendant that it would not approve changing the Fosamax label to include
that warning in the CRL. Because the basis for the FDA’s rejection was insufficient
evidence of a causal link between Fosamax and atypical femoral fractures, the Court
is satisfied that the evidence is clear and convincing that the Agency would not have
approved a differently worded warning no matter how Defendant attempted to
submit one. Plaintiffs’ state law failure-to-warn claims are therefore preempted, and
Defendant’s Motion for Summary Judgment is GRANTED.

DATED: March 23, 2022

/s/ Freda L. Wolfson
Hon. Freda L. Wolfson
U.S. Chief District Judge

87