VIOXX JUDGMENT THROWN OUT BY HOUSTON COURT OF APPEALS
Merck & Co., Inc. v. Ernst, No. 14-06-00835-CV (Tex. App. - Houston [14th Dist.] May 29, 2008) (Opinion
by Chief Justice Adele Hedges) (product liability) (Vioxx judgment, damages awarded by jury, reversed on the
ground that there was insufficient evidence to support causality element of plaintiff’s cause of action; causal
connection not proven)
O P I N I O N
Merck & Co., Inc., appeals from a jury verdict in a personal_injury and wrongful_death suit filed by Carol Ernst in
which she alleged that ingestion of Vioxx caused the sudden cardiac death of her husband, Bob Ernst. Merck
raises four issues in which it challenges the legal and factual sufficiency of the evidence to support the jury’s
verdict on causation, strict liability, negligence, malice, and damages. Merck further contends that the trial court
erred in instructing the jury and in admitting certain evidence. Finding the evidence to be legally insufficient on the
issue of causation, we reverse the trial court’s judgment and render judgment that appellee take nothing.
Vioxx, known generically as rofecoxib, belongs to a general class of pain relievers known as non steroidal anti
inflammatory drugs (“NSAIDs”). NSAIDs work by inhibiting cyclooxygenase (“COX”), an enzyme that stimulates
synthesis of prostaglandins, which are chemicals produced in the body that promote certain effects. Traditional
NSAIDs, such as Advil (ibuprofen), Aleve (naproxen), and Voltaren (diclofenac), have been longstanding treatment
options for patients needing relief from chronic or acute inflammation and pain associated with osteoarthritis,
rheumatoid arthritis, and other musculoskeletal conditions. This relief, however, has historically come with
significant adverse side effects. Specifically, traditional, or non_selective, NSAIDs greatly increase the risk of
gastrointestinal perforations, ulcers, and bleeds (”PUBs”). This risk is further increased when high doses are
ingested, which is often necessary to remedy chronic or acute inflammation and pain.
In the early 1990s, scientists discovered that the COX enzyme had two formsCCOX 1 and COX 2Ceach of which
appeared to have several distinct functions. Scientists believed that COX 1 affected the synthesis or production of
prostaglandins responsible for protection of the stomach lining, whereas COX 2 mediated the synthesis or
production of prostaglandins responsible for pain and inflammation. This belief led scientists to hypothesize that
“selective” NSAIDs designed to inhibit COX 2, but not COX 1, could offer the same pain relief as non_selective
NSAIDs with a reduced risk of fatal or debilitating PUBs. In addition, scientists believed that such drugs might also
prove beneficial for the prevention or treatment of other conditions, such as Alzheimer’s disease and certain
cancers where evidence suggests that inflammation may play a causative role. In light of these scientific
developments, pharmaceutical companies began developing new drugs known as “COX 2 inhibitors” or “coxibs.”
Merck developed a COX 2 inhibitor and named it Vioxx.
On November 23, 1998, Merck submitted a new drug application for Vioxx to the Food and Drug Administration
(“FDA”) and requested an expedited review of its application. Six months later, on May 20, 1999, the FDA
approved Vioxx as safe and effective for treatment of osteoarthritic pain, menstrual pain, and acute pain based on
the data and label supplied by Merck.
Vioxx was subjected to a number of studies and tests both before and after its initial approval. In March of 2000,
Merck received the preliminary results of the Vioxx Gastrointestinal Outcomes Research (“VIGOR”) study. VIGOR
was an 8,000 patient trial designed to assess the relative incidence of gastrointestinal PUBs in rheumatoid arthritis
patients treated with Vioxx as compared to those treated with the drug naproxen. While VIGOR demonstrated that
patients taking Vioxx suffered fewer serious gastrointestinal PUBs than patients taking naproxen, it also showed
that patients on Vioxx suffered a statistically significant increase of serious cardiovascular thrombotic events
compared to patients taking naproxen.
At the time of the VIGOR trial, two studies were being conducted on the effects of Vioxx on Alzheimer’s patients.
Those studies differed from the VIGOR trial in that the control group was taking a placebo, not naproxen. After
receiving the VIGOR results, Merck requested that the blind controls be removed from those studies so that the
scientists at Merck could investigate whether the patients who were taking Vioxx in the Alzheimer’s studies
experienced similar risks of cardiovascular thrombotic events. The results of both of the Alzheimer’s studies
revealed that patients taking Vioxx did not suffer a greater risk for cardiovascular thrombotic events.
In light of the new data obtained in the VIGOR study, Merck submitted a proposed label change for Vioxx to the
FDA in June of 2000. After approximately 18 months of negotiation with Merck over the content and organization
of a new Vioxx label, the FDA approved a revised label on April 11, 2002. The new label incorporated the VIGOR
data and noted that such data “should be taken into consideration and caution should be exercised when Vioxx is
used in patients with a medical history of ischemic heart disease.”
On September 23, 2004, an external safety board monitoring the results of a separate long term study entitled
Adenomatous Polyp Prevention on Vioxx (“APPROVe”), which was designed to assess whether Vioxx could help
prevent the recurrence of precancerous colon polyps, informed Merck that the interim data from this study also
showed a significantly increased rate of cardiovascular events in the Vioxx arm as compared to the placebo arm of
the study. One week later, on September 30, 2004, Merck voluntarily withdrew Vioxx from the market.
On September 15, 2000, Dr. Brent Wallace prescribed a daily 25_milligram dose of Vioxx to Bob Ernst to alleviate
tendinitis pain in Ernst’s hands. On May 6, 2001, approximately one hour after Ernst went to bed, appellee noticed
that he was unconscious and was having trouble breathing. Appellee called emergency medical personnel and
began cardiopulmonary resuscitation (“CPR”). Ernst never regained consciousness and was pronounced dead
shortly after arriving in the emergency room. An autopsy was performed the following morning, and the report
listed his death as cardiac arrhythmia secondary to coronary atherosclerosis.
Appellee sued Merck alleging that the ingestion of Vioxx caused the death of her husband. Appellee’s suit was
tried to a jury, which found that the design and marketing of Vioxx was defective, that Merck’s negligence
proximately caused Ernst’s death, and that the harm to Ernst resulted from malice attributable to Merck. The jury
awarded a total of $24,450,000 in compensatory damages and assessed $229,000,000 in exemplary damages.
Pursuant to section 41.008 of the Texas Civil Practice and Remedies Code, the trial court reduced the assessment
of exemplary damages and entered judgment for appellee in the sum of $26,100,000. From that judgment, Merck
In its first issue, Merck argues that appellee failed to present legally sufficient evidence of causation. At trial,
appellee alleged that Ernst’s death was caused by a blood clot triggered by Vioxx. Merck first argues that appellee
failed to present competent evidence of the existence of such a clot.
The test for legal sufficiency “must always be whether the evidence at trial would enable reasonable and fair
minded people to reach the verdict under review.” City of Keller v. Wilson, 168 S.W.3d 802, 827 (Tex. 2005).
Legal sufficiency review in the proper light must credit favorable evidence if reasonable jurors could, and disregard
contrary evidence unless reasonable jurors could not. Id. Although the reviewing court must consider evidence in
the light most favorable to the judgment, and indulge every reasonable inference that would support it, if the
evidence permits only one inference, neither jurors nor the reviewing court may disregard it. Id. at 822. We
sustain a legal_sufficiency challenge when the record discloses one of the following situations: (1) complete
absence of evidence establishing a vital fact; (2) the court is barred by rules of law or of evidence from giving
weight to the only evidence of a vital fact; (3) the evidence offered to prove a vital fact is no more than a mere
scintilla; or (4) the evidence conclusively establishes the opposite of a vital fact. Id. at 810.
A thrombotic cardiovascular event is a myocardial infarction or sudden cardiac death triggered by a thrombus, or
blood clot. The autopsy report reflected that Ernst suffered mild to severe atherosclerosis in the left anterior artery
and left circumflex coronary arteries and listed the cause of death as “cardiac arr[h]ythmia secondary to coronary
atherosclerosis.” Both parties’ experts testified that the usual cause of a myocardial infarction is a blood clot
formed in response to a rupture or fissure of atherosclerotic plaque in the inner lining of the artery. It is
undisputed that no blood clot or fissure of atherosclerotic plaque was found during the autopsy.
Appellee’s theory of liability is that Vioxx increases the risk of thrombotic cardiovascular events. A key component
of this theory is what has become known as the Fitzgerald hypothesis, which was first posited by Garret Fitzgerald,
a researcher at Merck Research Laboratories. Toward the end of 1997, Merck was studying the renal effects of
Vioxx. In measuring urinary output of patients taking Vioxx, Fitzgerald found a decrease in prostaglandin
metabolites, indicating the possibility of an imbalance of prostacyclin and thromboxane. Fitzgerald theorized that
Vioxx creates an imbalance between thromboxane and prostacylcin in the blood vessels. Thromboxane promotes
platelet aggregation, vessel constriction, and proliferation of smooth muscle cells. Prostacyclin, by contrast,
opposes the action of thromboxane inhibiting platelet aggregation, facilitating vasodilation, and preventing
proliferation of smooth muscle cells. In blocking COX_2, Vioxx also blocks prostacyclin; therefore, Fitzgerald
hypothesized, the inhibition of COX_2 promotes an imbalance of prostacyclin and thromboxane in the blood
vessels, leading to the formation of blood clots. Appellee claims that this mechanism ultimately led to the
formation of a blood clot in Ernst’s coronary artery, causing his sudden death.
Appellee presented the testimony of three experts who testified that Vioxx caused Ernst’s death. Dr. Benedict
Lucchesi, a professor of pharmacology at the University of Michigan, testified as follows: He has performed
extensive research on COX_2 inhibitors, including Vioxx. He testified extensively about the Fitzgerald hypothesis
and the mechanism by which Vioxx can cause thrombotic cardiovascular events. Ernst “died of an arrhythmia,
precipitated by a transient ischemic event leading to ventricular fibrillation.” An “ischemic event” is an event that
decreases the amount of oxygen delivered to the heart. Dr. Lucchesi testified that Ernst most likely died when a
blood clot dissolved in the artery; he could have, however, suffered from very small clots that moved toward
smaller capillaries after causing the arrhythmia. No microemboli, or small blood clots, were found during the
Dr. Isaac Weiner, a board_certified cardiologist, testified as follows: Vioxx was a significant contributing factor in
causing Ernst’s death because a thrombus, or blood clot, could have formed on the underlying blockage in Ernst’s
arteries and blocked blood flow to the heart. Dr. Weiner opined that a blood clot might have dissolved through
fibrinolysis. Vioxx does not affect the electrical impulses of the heart directly, but leads to blockages in the
arteries, which cause cardiovascular events.
Dr. Maria Araneta, the assistant medical examiner who performed the autopsy, testified as follows: Ernst’s left
coronary arteries were 50_ to 75_percent blocked by calcified plaque. She did not find a blood clot when she
performed the autopsy on Ernst’s body. She did not see evidence of a myocardial infarction, or heart attack,
because Ernst died within an hour of initially showing symptoms. To affirmatively diagnose a myocardial infarction,
the physician must observe either death of the heart muscle tissue or the presence of cardiac enzymes. To
observe either death of the tissue or the presence of cardiac enzymes, the individual must have lived six to
eighteen hours after suffering the myocardial infarction. The most likely cause of Ernst’s arrhythmia, Dr.
Araneta testified, was an acute ischemic event. As the potential cause of such an event, she said, “Something
blocked that artery that was already narrowed, either a clot, a fissure, block or ruptured atheroma, none of which I
saw, but it - these things could be dissolved.” According to Dr. Araneta, the clot could have spontaneously
dissolved, been dislodged through vigorous CPR, or could have been too small to detect on autopsy. Not only did
she not see a clot, but she observed no rupturing or fissuring of the atherosclerotic plaque in the coronary
arteries. In fact, the plaque in Ernst’s arteries was intact and so hard that the arteries had to be de_calcified
before they could be sectioned for autopsy. Contrary to her report, at trial Dr. Araneta was “leaning towards more
likely than not” that Ernst suffered a myocardial infarction.
Admitting that there is no direct evidence of a blood clot or myocardial infarction, appellee contends that there is
sufficient circumstantial evidence to support the jury’s verdict. The circumstantial evidence on which appellee
relies is the testimony that a clot could have been present, but could have disappeared through various means
prior to the autopsy. In claims supported by only meager circumstantial evidence, the evidence does not rise
above a scintilla if jurors would have to guess whether a vital fact exists. City of Keller, 168 S.W.3d at 813. When
the circumstances are equally consistent with either of two facts, neither fact may be inferred. Id. When the
circumstantial evidence of a vital fact is meager, we must consider not just favorable but all the circumstantial
evidence, and competing inferences as well. Id. at 814. Therefore, we address the circumstantial evidence that
although a clot was not found, it could have disappeared.
First, Dr. Araneta opined that a clot could have naturally dissolved through a process called fibrinolysis. However,
Dr. Araneta admitted that she would not expect a clot to naturally dissolve after death. Dr. Lucchesi explained that
certain types of clots could naturally move through the blood vessels, but did not specify how quickly this process
could take place in the body. Dr. Craig Pratt, director of the coronary care unit at Methodist Hospital and a
professor of medicine at Baylor College of Medicine, testified that the possibility of a clot dissolving on its own was
not a viable hypothesis because the natural process of fibrinolysis takes 24 to 48 hours and only continues while
the patient is alive. Further, clot_busting drugs, which are administered to accelerate this natural process, usually
take at least an hour to dissolve a clot.
Next, Dr. Araneta and Dr. Lucchesi speculated that vigorous CPR could have dislodged a blood clot or caused it to
fragment into such small pieces that it could not be detected on autopsy. Dr. Araneta testified that emboli could
have been dislodged and fragmented into such small pieces that she could not have seen them on her
microscope. She described this theory as a “guess,” her “estimate,” and a “possibility.” Dr. Lucchesi admitted that
he was aware of no published literature suggesting that CPR could dislodge or move clots.
An expert’s bare opinion will not suffice to support a jury’s verdict. Merrell Dow Pharm. v. Havner, 953 S.W.2d 706,
711 (Tex. 1997). The substance of the testimony must be considered. Id. Factors to which a reviewing court
should look in determining the reliability of scientific testimony are: (1) the extent to which the theory has been or
can be tested; (2) the extent to which the technique relies upon the subjective interpretation of the expert; (3)
whether the theory has been subjected to peer review and/or publication; (4) the technique’s potential rate of
error; (5) whether the underlying theory or technique has been generally accepted as valid by the relevant
scientific community; (6) the nonjudicial uses that have been made of the theory or technique; and (7) any other
factor that is helpful to determine the reliability of the scientific evidence. Id. at 714 (citing E.I. duPont de Nemours
and Co. v. Robinson, 923 S.W.2d 549, 557 (Tex. 1995)).
Causation opinions based on possibility, speculation, and surmise are no evidence. Havner, 953 S.W.2d at 711-
12. Expert opinions must be supported by facts in evidence, not conjecture. Marathon Corp. v. Pitzner, 106 S.W.
3d 724, 729 (Tex. 2003).
Dr. Araneta’s and Dr. Lucchesi’s opinions are not supported by scientifically reliable facts in evidence. Appellee’s
experts postulate that Ernst suffered a thrombotic cardiovascular event, but at the time of his death, less than one
hour later, the thrombus had been dislodged and fragmented into such small pieces that it could not be seen on a
microscope. Because there is no proof that a clot existed and was dislodged, the experts’ opinions are mere
speculation. Moreover, we cannot disregard Dr. Thomas Wheeler’s uncontroverted testimony that there is no
support in scientific or medical literature for such a theory. Dr. Wheeler conceded that he had neither
independently learned of CPR dislodging or fragmenting a clot, nor had he read any anecdotal case reports
supporting the theory. He further testified that after hearing Dr. Araneta’s and Dr. Lucchesi’s testimony, he
researched the possibility of CPR’s dislodging a clot and found no scientific or medical support for the theory. Dr.
Wheeler explained that if a clot had been fragmented by CPR, the anatomy of the heart would not have permitted
the fragments to travel down toward smaller capillaries, but would have pushed the fragments to a position in the
coronary artery where they could have been detected.
Appellee contends that the failure to find a blood clot does not defeat causation, arguing that Dr. Lucchesi ruled
out all non_thrombotic causes of Ernst’s arrhythmia. Dr. Lucchesi testified that Ernst was under 65 years old,
exercised regularly, had not smoked for more than 15 years, and did not have a family history of heart disease.
Appellee contends that by ruling out other potential causes of sudden cardiac death, the evidence is legally
sufficient to show Ernst suffered a myocardial infarction triggered by a blood clot. Dr. Lucchesi’s testimony that
Ernst did not exhibit risk factors for cardiovascular disease does not support the reasonable conclusion that Ernst
suffered a thrombotic cardiovascular event.
Appellee points to Dr. Araneta’s testimony that she rarely finds a clot on autopsy; therefore, appellee argues the
failure to find a clot does not rule out the possibility of a myocardial infarction. Dr. Araneta’s testimony, however,
referred to cases in which the autopsy revealed pathological evidence of myocardial infarction, i.e., dead or
obviously infarcted muscle tissue. Every expert agreed that Ernst died too suddenly for his heart muscle tissue to
show signs of death. Moreover, the trial court excluded any evidence that ingestion of Vioxx is associated with
non_thrombotic cardiovascular events because the theory was not supported by scientifically reliable evidence.
The reliable expert testimony supports the theory that Vioxx causes cardiovascular events by causing blood clots,
which lead to myocardial infarction or sudden cardiac death. See Merck & Co. v. Garza, 04_07_00234_CV 2008
WL 2037350 at *2 (Tex. App.CSan Antonio May 14, 2008, no pet. h.) (not released for publication) (recognizing
expert testimony that “the formation of clots is the type of problem caused by Vioxx”). Appellee directs our
attention to Dr. Araneta’s and Dr. Lucchesi’s opinions that “something” triggered the ischemic event suffered by
Ernst. However, there is no competent evidence that a blood clot triggered by Vioxx ingestion was the “something”
causing his death.
The epidemiological evidence supports the conclusion that Vioxx use at a certain dose and duration is associated
with an increased risk of thrombotic cardiovascular events. The experts’ speculation that a clot “could have”
existed, but “could have” dissolved, been dislodged, or fragmented gives rise to nothing more than conjecture.
Crediting all favorable evidence that reasonable jurors could believe and disregarding all contrary evidence except
that which they could not ignore, we find no evidence that Ernst suffered a thrombotic cardiovascular event, i.e., a
myocardial infarction triggered by a blood clot. Accordingly, appellee failed to show that the ingestion of Vioxx
caused her husband’s death. Merck’s first issue is sustained.
The judgment of the trial court is reversed and judgment is rendered that appellee take nothing.
/s/ Adele Hedges
Judgment rendered and Opinion filed May 29, 2008.
Panel consists of Chief Justice Hedges and Justices Anderson and Brown.
 Claire Bombardier, et al., Comparison of Upper Gastrointestinal Toxicity of Rofecoxib and Naproxen in Patients
with Rheumatoid Arthritis, 343 New Eng. J. Med. 1520 (Nov. 23, 2000).
 S.A. Reines, et. al., Rofecoxib No effect on Alzheimer’s Disease in a 1_year, Randomized, Blinded, Controlled
Study, 70 Neurology 66 (Sept. 16, 2003); Leon J. Thal, et. al., A Randomized, Double_Blind, Study of Rofecoxib in
Patients with Mild Cognitive Impairment, Neuropsychopharmacology (2005).
 Robert S. Bresalier, et. al., Cardiovascular Events Associated with Rofecoxib in a Colorectal Adenoma
Chemoprevention Trial, 352 New Eng. J. Med. 1092 (Mar. 17, 2005).
 In 1998, Merck made the FDA advisory board aware of the Fitzgerald hypothesis. The FDA concluded that
while the hypothesis was a theoretical concern, there was no evidence that it was true.
 All experts agreed on this conclusion.
 Dr. Wheeler is the head of pathology at Baylor College of Medicine.
 Because Merck’s first issue is dispositive, we need not address its remaining issues or the issues raised in