United States Court of Appeals
                            For the Eighth Circuit
                        ___________________________

                                No. 17-2630
                        ___________________________

                                       Jo Levitt

                        lllllllllllllllllllllPlaintiff - Appellant

                                           v.

                              Merck & Company, Inc.

                       lllllllllllllllllllllDefendant - Appellee
                                      ____________

                     Appeal from United States District Court
                for the Western District of Missouri - Kansas City
                                 ____________

                          Submitted: September 27, 2018
                             Filed: February 4, 2019
                                  ____________

Before COLLOTON, GRUENDER, and GRASZ, Circuit Judges.
                        ____________

GRUENDER, Circuit Judge.

       Merck & Company, Inc. (“Merck”) manufactured and distributed Vioxx as a
medication to relieve pain and inflammation between 1999 and 2004. Plaintiff Jo
Levitt began taking Vioxx during the summer of 1999. She suffered cardiovascular
injuries in March and May of 2000 while taking the medication. Her doctor
continued to prescribe her Vioxx until 2002, around the time Merck changed its label
to disclose a risk of cardiovascular injuries associated with the use of Vioxx. Merck
removed Vioxx from the market in 2004.

       Levitt filed a personal injury lawsuit against Merck on September 29, 2006.
Merck filed a motion for judgment on the pleadings, arguing that Levitt’s claims were
barred by Missouri’s five-year statute of limitations. The district court granted
Merck’s motion and determined as a matter of law that, because “Plaintiff’s claims
accrued prior to September 2001,” her September 29, 2006 suit was time-barred.
Levitt appealed.

       We review a district court’s grant of judgment on the pleadings de novo.
Clemons v. Crawford, 

, 1124 (8th Cir. 2009). The movant has the
burden of “clearly establish[ing] that there are no material issues of fact and that it is
entitled to judgment as a matter of law.” Porous Media Corp. v. Pall Corp., 

, 1079 (8th Cir. 1999). At this stage of the proceedings, we view all facts
pleaded by Levitt as true and grant her all reasonable inferences. See 

.

       Judgment on the pleadings may be granted “on the basis that the
governing . . . statute of limitations expired.” Thach v. Tiger Corp., 

,
960 (8th Cir. 2010). The parties agree that Missouri’s statute of limitations applies,
and we review the district court’s determinations of Missouri law de novo. See
Jurrens v. Hartford Life Ins. Co., 

, 922 (8th Cir. 1999). “The party
asserting the affirmative defense of the running of the applicable statute of limitations
has the burden of not only pleading but proving it.” Lomax v. Sewell, 

,
552 (Mo. Ct. App. 1999). In Missouri, the statute of limitations for personal injury
claims is five years after the cause of action accrues. Mo. Ann. Stat. § 516.120.
“[T]he cause of action shall not be deemed to accrue when the wrong is done or the
technical breach of contract or duty occurs, but when the damage resulting therefrom
is sustained and is capable of ascertainment.” Mo. Ann. Stat. § 516.100.

                                           -2-
       The Missouri Supreme Court defined “capable of ascertainment” as when “the
evidence [is] such to place a reasonably prudent person on notice of a potentially
actionable injury.” Powel v. Chaminade Coll. Preparatory, Inc., 

,
582 (Mo. 2006) (emphasis removed). This “objective” test is from the standpoint of
a “reasonable person in [plaintiff’s] situation.” 

586. Both the “character
of the condition . . . and its cause” must be capable of ascertainment. Elmore v.
Owens-Illinois, Inc., 

, 436 (Mo. 1984). “[W]hen contradictory or
different conclusions may be drawn from the evidence as to whether the statute of
limitations has run, it is a question of fact for the jury to decide.” 

.

       We conclude that there remains such “a question of fact for the jury to decide”
because “contradictory or different conclusions may be drawn” as to whether “the
evidence was such to place a reasonably prudent person on notice of a potentially
actionable injury” before September 29, 2001. See 

Between 1999 and
2004, scientists became increasingly confident that the use of Vioxx entailed a risk
of cardiovascular injuries. Beginning in early 2000, the media covered reports of a
possible link between Vioxx and cardiovascular incidents. See, e.g., Edward R.
Silverman, Merck Shares Fall on Vioxx Study, Star-Ledger (Newark, N.J.), April 29,
2000, at 17; Rita Rubin, Data: Vioxx Might Raise Heart Risk, USA Today, February
9, 2001; Could Vioxx and Celebrex Increase Heart Risk?, Wash. Post, March 13,
2001, at T6. In late 2000, the Vioxx Gastrointenstinal Outcomes Research (VIGOR)
study indicated a potential risk of myocardial infarction associated with the use of
Vioxx. See Claire Bombardier et al., Comparison of Upper Gastrointestinal Toxicity
of Rofecoxib and Naproxen in Patients with Rheumatoid Arthritis, 343 New Eng. J.
Med. 1520, 1523 (2000).

       But the scientific community’s analysis remained tentative. In August 2001,
a peer-reviewed article in the Journal of the American Medical Association analyzed
“all published, English-language, randomized, double-blind trials of COX-2

                                          -3-
inhibitors [like Vioxx] from January 1998 to February 2001” and concluded that “the
available data raise[d] a cautionary flag about the risk of cardiovascular events with
COX-2 inhibitors” such as Vioxx. Debabrata Mukherjee et al., Risk of
Cardiovascular Events Associated with Selective COX-2 Inhibitors, 286 JAMA 954,
954-955 (2001). But this comprehensive analysis recognized “significant
limitations”:

      There remains considerable uncertainty in any post hoc analysis. The patient
      populations in these trials . . . with rheumatoid arthritis have a higher risk of
      MI [myocardial infarction]. . . . Currently, no data exist on cardiovascular
      safety for the sporadic, intermittent use of these agents by individuals for
      musculoskeletal pain, which appears to be the most frequent pattern of use. . . .
      [D]efinitive evidence of [a cardiovascular] adverse effect will require a
      prospective randomized clinical trial.

And Merck publicly argued at that time that no sound conclusions relating
to cardiovascular risks could be drawn from the available studies. See Thomas M.
Burton & Gardiner Harris, Note of Caution: Study Raises Specter of Cardiovascular
Risk For Hot Arthritis Pills—Vioxx and Celebrex Marketers Dispute the Research,
Sought to Downplay It—A Spurned Appeal to JAMA, Wall St. J., August 22, 2001, at
A1.

       The Missouri Supreme Court has not addressed how Powel’s “reasonably
prudent person” test applies where an injury was only beginning to be linked to a
certain cause by developing science. Therefore, “we must attempt to predict what the
court would decide if it were to address the issue” based on “relevant state precedent,
analogous decisions, considered dicta, . . . and any other reliable data.” Jurrens v.
Hartford Life Ins. Co., 

, 922 (8th Cir. 1999).

      The Missouri Court of Appeals’s recent decision in Giles v. Carmi Flavor &
Fragrance Co., Inc. is most instructive. 

(Mo. Ct. App. 2015). Giles
involved a worker at a food facility who was exposed to diacetyl, a chemical used in

                                         -4-
butter flavoring. In 1999 Giles began to experience lung problems, including
difficulty breathing, shortness of breath, coughing, and wheezing. 

Around
2001 and 2002, Giles and his primary care physician had actual knowledge about “the
possible connection between butter flavoring and certain lung diseases,” but his
doctor did not make a diagnosis on this basis. 

In 2004, Giles switched
primary care physicians, and between 2004 and 2011 his new doctor diagnosed Giles
with dyspnea (shortness of breath) and asthma. 

Finally, in 2011, he
was referred to a third doctor who diagnosed Giles with bronchiolitis obliterans
caused by diacetyl. 

       The trial court in Giles granted summary judgment based on Missouri’s five-
year statute of limitations. The Missouri Court of Appeals reversed, finding that there
was a question of fact as to whether the cause of Giles’s lung condition was capable
of ascertainment when “the scientific community [was] only beginning to piece
together a connection between the bronchiolitis obliterans and diacetyl.” 

       In dismissing Levitt’s claims, the district court relied on two pre-Powel cases.
In Ahearn v. Lafayette Pharmacal, Inc., the plaintiff exhibited symptoms of a back
condition in 1975 but “did not learn of the causal relationship between the drug and
the disease until 1983.” 

, 503 (Mo. Ct. App. 1987). She filed her
lawsuit in 1984. 

Court of Appeals determined that the statute of
limitations barred her lawsuit because “the medical community, as early as the
1940’s, was aware of the possible causation link” and “[t]here was considerable
literature available to the medical community which suggested this causal
connection.” 

Buttice v. G.D. Searle & Co., the plaintiff had an
intrauterine device (IUD) inserted in 1976 and removed a year later. 

, 564 (E.D. Mo. 1996). Between 1976 and 1983, she suffered from a pelvic
infection, blocked fallopian tubes, and infertility. 

In 1995 she filed a
lawsuit alleging that her IUD had caused these injuries. 

determined
that her injuries were “capable of ascertainment more than five years before” in part

                                         -5-
because “[b]y the early to mid 1980’s, medical texts regularly referenced the reported
association between IUD use and pelvic inflammatory disease . . . and the resulting
possibility of infertility.” 

567.

        Relying on Buttice and Ahearn, the district court concluded that the “case
law . . . only requires knowledge of ‘a possible link’ between the injury and product.”
Levitt v. Merck Sharp & Dohme Corp., 

, 387-88 (W.D. Mo. 2017)
(emphasis added) (quoting Buttice, 938 F. Supp at 567). Such a standard was not
specifically discussed in Powel, and it was rejected in 

(“[W]e reject the argument that summary judgment should be affirmed because the
medical community was aware of the potential that diacetyl exposure could cause
lung disease in 2001 and thus that Giles was on notice of a potentially actionable
claim at that time.”). Giles distinguished Ahearn and Buttice as cases where “the
science had long provided a causal connection between the diagnosis and the alleged
wrong” (Ahearn) and where a “causal theory [for plaintiff’s injury] was capable of
ascertainment” about a decade before the cause of action was filed (Buttice).2 

.

       But here, as in Giles, the causal theory linking Vioxx to heart problems was
only beginning to emerge in the scientific community before September 29, 2001.
Therefore, we predict that the Missouri Supreme Court would conclude that mere
knowledge in the medical community of a possible link does not as a matter of law
place “a reasonably prudent person in [Levitt’s] position . . . on notice of a potentially
actionable injury.” See 

. Powel focuses on when “the
evidence was such to place a reasonably prudent person on notice of a potentially

      2
        Giles also cites a federal district court opinion that questioned Ahearn’s
viability and concluded “that medical community knowledge of possible causation
is inconclusive” for statute of limitations purposes. 

n.9
(quoting Miller v. Mobay Corp., 

, 178 (W.D. Mo. 1990)).


                                           -6-
actionable injury.”3 

(emphasis added). As Giles and
Powel indicate, Missouri law treats this as “a question of fact for the jury to decide”
when “contradictory or different conclusions may be drawn” concerning the
implications of the scientific evidence for a reasonable plaintiff. 

(quoting 

).

       Here, the state of the scientific literature connecting Vioxx to cardiovascular
injuries before September 29, 2001 was more akin to Giles than to Ahearn and Buttice
because at that time science was only “beginning to piece together” a causal theory.
See 

Most compellingly, only one month before September 29, 2001, the
American Medical Association’s flagship journal found that “considerable
uncertainty” remained as to the apparent connection between Vioxx and
cardiovascular injuries, “no data” existed on whether “the most frequent pattern of
use” of Vioxx caused any heart problems at all, and more studies were needed. See
Mukherjee et 

. In addition, in the month prior to September 29,
2001, Merck’s representatives and scientists were publicly contesting the validity of
the scientific claims suggesting a possible heart risk from the use of Vioxx, and
Merck had yet to add any warnings of cardiovascular risks to Vioxx’s label.

      To be sure, the national news media covered reports of potential heart risks
from taking Vioxx starting in 2000, and some plaintiffs filed cases in the months
before September 2001 alleging that Vioxx caused heart problems. See, e.g., J.A.
931, Reid v. Merck & Co., No. BC254630 (Cal. Super. Ct. July 23, 2001). Such
publicly available information is relevant to whether “notice of a potentially




      3
        As Giles emphasizes, the “test to determine when a cause of action has accrued
is to ascertain the time when plaintiff could have first maintained the action to a
successful 

(quoting 

).


                                         -7-
actionable injury” existed, but in this case it is not determinative as a matter of law.4
See 

(“The record simply does not bear out, as a matter
of law, that Giles could have maintained the action to a successful result prior to
2007, unlike in Ahearn . . . .”). Thus, Merck failed to “clearly establish[] that there
are no material issues of fact,” Porous Media 

, as to
“whether the statute of limitations has run,” 

.

        We do not require definitive proof of causation, a consensus in the scientific
community, or an admission by the putative defendant for the statue of limitations to
begin to run. But with all reasonable inferences given to Levitt, and following the
most recent and “relevant state precedent, analogous decisions, considered dicta,
. . . and any other reliable data” to predict what the Missouri Supreme Court would
decide after Powel, we conclude that judgment on the pleadings was improper. See

. As Giles emphasizes, “[a]lthough it may be the case that
a reasonably prudent person in [plaintiff’s] position would have been on notice of a
potentially actionable injury . . . , this conclusion does not follow as a matter of law.”

. We thus reverse the district court’s dismissal on this basis
and remand the case for further proceedings consistent with this opinion.




      4
        The dissent argues that this is a question of law because the “the standard ‘is
an objective one’” and “[t]here is no dispute about the scientific literature, media
reports, corporate annual reports, and publicly-filed lawsuits that were in the public
domain before September 2001.” Post, at 12-13. We agree that the existence of such
materials is undisputed, as it was undisputed Giles. But Giles shows that the import
of such materials can be reasonably disputed. In such a scenario, Powel and Giles
call upon the jury to decide whether a “reasonably prudent person” in plaintiff’s
position—as opposed to Levitt herself—was on “notice of a potentially actionable
injury.” 

(quoting 

).


                                           -8-
COLLOTON, Circuit Judge, dissenting.

      I conclude that the district court correctly applied Missouri law in ruling that
Jo Levitt’s claim against Merck & Company was barred by the statute of limitations.
I would therefore affirm the judgment.

       Missouri applies an “inquiry notice” standard to begin the running of a statute
of limitations. Powel v. Chaminade Coll. Preparatory, Inc., 

, 583
(Mo. 2006). There is no requirement that the plaintiff have actual knowledge of her
injury and its cause. Wheeler v. Eftink, 

, 604 (Mo. Ct. App. 2016).
A cause of action accrues when the “evidence was such to place a reasonably prudent
person on notice of a potentially actionable injury.” 

.

       In this case, Levitt suffered cardiovascular injuries in March and May 2000
while taking Vioxx, a medication manufactured by Merck. She underwent double
coronary bypass surgery. Around the same time, information in the public domain
placed a reasonably prudent person on notice of a causal theory that linked Vioxx to
cardiovascular injuries. The district court properly took judicial notice of several
peer-reviewed journal articles, media reports, and public court records, some of which
were even cited in Levitt’s complaint. See, e.g., Miller v. Redwood Toxicology Lab.,
Inc., 

, 931 n.3 (8th Cir. 2012); Levy v. Ohl, 

, 991 (8th Cir.
2007); Kushner v. Beverly Enters., 

, 831 (8th Cir. 2003); Wash. Post v.
Robinson, 

, 291-92 (D.C. Cir. 1991).

       In 2000, an article in the New England Journal of Medicine reported that in a
clinical study of 8,076 people taking arthritis drugs, patients using Vioxx suffered
more heart attacks than patients using a comparator agent, naproxen. News reports
on this study explained that a clinical trial of Vioxx had “detected statistically
‘significant’ cardiovascular problems,” (Newark Star-Ledger, April 2000), that



                                         -9-
“patients on Vioxx were more likely to suffer heart attacks and other cardiovascular
complications than those on naproxen,” (USA Today, Feb. 2001), that “patients
taking Vioxx have a higher risk of heart attack,” (CBS The Early Show, May 2001),
and that takers of Vioxx “had significantly more heart attacks” than users of the
alternative. (Washington Post, March 2001). See J.A. 639, 898-99, 901, 904.

       In August 2001, a review of literature published in the Journal of the American
Medical Association reiterated that the clinical study showed a “significantly
increased risk of cardiovascular event” with use of Vioxx rather than the alternative,
and opined that “the available data raise a cautionary flag about the risk of
cardiovascular events” with Vioxx. The authors found “a potential increase in
cardiovascular event rates” for Vioxx and “urge[d] caution in prescribing” the drug
to certain patients.

       As early as July 2001, other plaintiffs sued Merck alleging that Vioxx caused
serious cardiovascular events. One complaint from July 23, 2001, alleged that Vioxx
had “been linked to several severe and life threatening medical disorders,” including
heart attack and stroke. J.A. 931, Reid v. Merck & Co., Case No. BC254630 (Cal.
Sup. Ct. July 23, 2001). Levitt herself alleged that Merck’s 2001 Annual Report
disclosed that a number of federal and state lawsuits had been filed against Merck
with respect to Vioxx, alleging gastrointestinal bleeding and cardiovascular events.
R. Doc. 1, at 9.

      Despite all of this information in the public domain before September 2001, the
court holds that a reasonable jury could find that a reasonably prudent person who
had suffered cardiovascular injuries while taking Vioxx was not on inquiry notice of
a potential claim against Merck. The rationale apparently is that the scientific data
allowed for uncertainty about whether Vioxx caused cardiovascular injuries, and
Merck was disputing the causal link.



                                        -10-
       In my view, the court’s approach calls for more than the “inquiry notice”
required to start the statute of limitations under Missouri law. The Missouri Court of
Appeals, Eastern District, long ago ruled that a “possible causation link” based on
literature in the medical community that “suggested” a “causal connection” was
sufficient as a matter of law to start the limitations period running. Ahearn v.
Lafayette Pharmacal, Inc., 

, 503 (Mo. Ct. App. 1987).

       The court believes that a recent decision of the Missouri Court of Appeals,
Western District, rejected the standard applied in Ahearn, but that case involved a
different issue. In Giles v. Carmi Flavor & Fragrance Co., 

(Mo. Ct.
App. 2015), the plaintiff was not even diagnosed with his injury outside the
limitations period, so the causal link between his injury and the defendant’s product
was immaterial to the decision. See 

(“The first two specialists who treated
Giles were unable to diagnose bronchiolitis obliterans even though they were made
aware of the potential occupational hazard. Under these facts, we cannot find as a
matter of law that Giles’s cause of action accrued prior to 2007.”) (emphases added).
Giles distinguished Ahearn on this basis, and the court never decided whether
particular scientific studies or journal articles suggesting a causal link were sufficient
to place an injured party on inquiry notice. 

Giles hinged on the fact that
“neither of two specialists was able to diagnose [the plaintiff] with a lung condition
related to [his] exposure” more than five years before his lawsuit. 

(emphasis added).

       Dictum in Giles does say that the science in Ahearn “had long provided a
causal connection between the diagnosis and the alleged wrong,” while the scientific
community in Giles was “only beginning to piece together a connection between the
bronchiolitis obliterans and diacetyl when Giles began showing symptoms of lung
disease.” 

provides little insight, however, because Giles says that
the plaintiff began showing symptoms in 1999, 

but does not describe what



                                          -11-
was known in the scientific community about a causal link at that time. Ahearn
described only a “possible causation link” based on literature that “suggested” a
causal connection, so Giles must have been referring to lesser knowledge than that
when it contrasted Ahearn.

        The better view of Missouri law is that studies, articles, and lawsuits in the
public domain before September 2001 placed Levitt on notice to inquire further about
whether Vioxx caused her cardiovascular injuries. To require definitive proof of
causation, a consensus in the scientific community, or an admission by the putative
defendant before the claim accrues would swallow the inquiry notice rule. And once
the information is available to a reasonably prudent person, it does not matter whether
the literature has been published for three weeks or three decades. When a plaintiff
should know about an injury and a variety of possible causes, she is on notice to
inquire further, and the statute of limitations begins to run, even though the exact
cause is yet unknown. See Ball v. Friese Const. Co., 

, 177 (Mo. Ct.
App. 2011). As the Iowa Supreme Court explained in an analogous situation, an
“inquiry notice” standard does not postpone the beginning of the limitations period
until a successful completion of the plaintiff’s investigation. Ranney v. Parawax Co.,

, 156 (Iowa 1998). Once the plaintiff knows of a possible causation
link, she is on notice to investigate, and she must complete an investigation and file
any action within the limitations period. 

says it is merely holding that the limitations issue cannot be decided
“as a matter of law,” but the standard “is an objective one,” and “where relevant facts
are uncontested, the statute of limitations can be decided by the court as a matter of
law.” 

. In Giles, there was a factual issue about whether
the plaintiff’s condition was capable of ascertainment before he was 

. But here, the relevant facts are uncontested. Levitt’s condition of
cardiovascular injury was diagnosed before September 2001. There is no dispute



                                         -12-
about the scientific literature, media reports, corporate annual reports, and publicly-
filed lawsuits that were in the public domain before September 2001. The only
fighting issue is a legal one: whether the publicly-available information placed a
reasonably prudent person with cardiovascular injury after taking Vioxx on inquiry
notice of a potential claim against Merck. By referring this issue to a jury, and
characterizing it as a “question of fact,” the court ironically contemplates a possible
ruling against Levitt based on scientific literature that the court says was “only
beginning to piece together” a connection between Vioxx and cardiovascular injury.
In my view, the better conclusion under Missouri law is that the relevant scientific
literature, reports, and lawsuits described a “possible causation link” and “suggested”
a causal connection, such that the limitations period began to run as a matter of law.
See 

.

       Levitt was on inquiry notice that she had a potential claim against Merck
before September 2001, but she did not file this action until September 2006. The
five-year statute of limitations had therefore expired. Levitt argues alternatively that
Merck fraudulently concealed her cause of action, but the district court correctly
rejected this claim: facts in the public domain placed Levitt on inquiry notice by
September 2001, and a putative defendant does not fraudulently conceal by disputing
a theory of causation. For these reasons, I would affirm the judgment.
                         ______________________________




                                         -13-