Centocor, Inc. v. Hamilton

delivered the Opinion of the Court.

Under the learned intermediary doctrine, the manufacturer of a pharmaceutical product satisfies its duty to warn the end user of its product’s potential risks by providing an adequate warning to a “learned intermediary,” who then assumes the duty to pass on the necessary warnings to the end user. See, e.g., Gravis v. Parke-Davis & Co., 502 S.W.2d 868, 870 (Tex.Civ.App.-Corpus Christi 1973, writ ref'd n.r.e.). In this case, we consider the applicability of the learned intermediary doctrine to a patient’s claims against a prescription drug manufacturer, whose product allegedly caused a serious injury. We hold that the doctrine generally applies within the context of a physician-patient relationship and allows a prescription drug manufacturer to fulfill its duty to warn end users of its product’s potential risks by providing an adequate warning to the prescribing physician. We further hold that the court of appeals erred by *143creating an exception to the learned intermediary doctrine for direct-to-consumer (DTC) advertising. Although the patient alleged various common law causes of action, all of the patient’s claims turn on the prescription drug manufacturer’s failure to warn. Therefore, the learned intermediary doctrine applies to all of the patient’s claims, and the patient was required to show that an inadequate warning to the prescribing physicians caused the patient’s injuries. Because the patient presented no evidence that the allegedly inadequate warning was a producing cause of her physicians’ decisions to prescribe the prescription drug, her claims fail as a matter of law. Accordingly, we reverse the court of appeals’ judgment in part and render judgment that the plaintiffs take nothing.

I. Background

In March 2003, Patricia and Thomas Hamilton sued Centocor, Inc., a prescription drug manufacturer and subsidiary of Johnson & Johnson, claiming that Cento-cor provided “inadequate and inappropriate warnings and instruction for use” of its prescription drug Remicade, which made Remicade “defective and unreasonably dangerous,” and seeking damages for injuries that Patricia allegedly incurred from using the drug.1 In August 2006, the Hamiltons amended their claims and added Patricia’s prescribing and treating physicians as defendants. They claimed that Centocor was liable for, among other things, (1) “manufacturing, promoting, distributing and/or selling Remicade®,” which was “defective and unreasonably dangerous” because of “inadequate and inappropriate warnings and instructions for use”; (2) negligence; (3) gross negligence; (4) fraud; and (5) malice. The Hamiltons claimed that Remicade caused Patricia to suffer a serious drug-induced side effect called lupus-like syndrome. The Hamil-tons also alleged that Patricia’s medical providers failed to adequately warn Patricia of the risks associated with Remicade and failed to obtain her informed consent to the treatment.

In the course of her prescribed treatments, Patricia’s treating physician, Michael Bullen, M.D., showed her an informational video that he received from Centocor. The Hamiltons alleged that Centocor’s video over-emphasized the benefits of Remicade and intentionally omitted warnings about the potential side effect of lupus-like syndrome. They argued that the video bypassed the physician-patient relationship and required Centocor to warn Patricia directly of Remicade’s potential risks and side effects, thereby making Centocor liable for Patricia’s injuries. The jury found in favor of the Hamiltons, and the trial court entered judgment for approximately $4.6 million. The court of appeals reversed the award of 'future pain and mental anguish damages but affirmed the remainder of the trial court’s judgment, adopted a DTC advertising exception to the learned intermediary doctrine, and held that the record contained sufficient expert evidence to prove that Centocor’s actions caused Patricia’s injuries.

A. Patricia’s Medical History Prior to 2001

Patricia Hamilton has a complicated medical history. For more than two decades, she has suffered from Crohn’s disease, recurring joint pain, arthritis, and several other ailments. Crohn’s disease is a chronic, lifelong inflammatory condition that can affect any part of the digestive system. There is no cure for the disease; *144however, patients have several treatment options, which seek to control intestinal inflammation. Over the years, Patricia underwent various procedures to treat the disease and mitigate its effects. By 2001, Patricia had part of her small intestine, colon, and rectum removed, and she lived with a colostomy. During a resection surgery — a procedure to reconnect her bowels after removing damaged tissues — Patricia contracted hepatitis C from a blood transfusion. She was also diagnosed with sar-coidosis.2

B. Dr. Hauptman Treats Patricia’s Crohn’s Disease

In September 2001, Patricia experienced a “flare” in her Crohn’s disease and sought treatment from Ronald Hauptman, M.D., a gastroenterologist who was practicing in Corpus Christi. To confirm that Patricia’s symptoms were caused by her Crohn’s disease, Dr. Hauptman. tracked Patricia’s reported abdominal pains for several weeks and ordered a series of tests, including a CAT scan and an upper GI.3 By December 2001, Dr. Hauptman confirmed that Patricia was experiencing a moderate flare in her Crohn’s disease.

Dr. Hauptman testified that it was important to treat the Crohn’s flare quickly to mitigate the risk that Patricia would lose more of her bowels. Based on Patricia’s existing medical regimen and her reported allergic reactions to one type of anti-inflammatory medication used to treat Crohn’s disease, Dr. Hauptman testified that Patricia’s only two options to treat the Crohn’s flare were through steroids or Remicade infusions. According to Dr. Hauptman, he consulted with Patricia about the available treatments and explained the risks and benefits of each approach. Based in part on Patricia’s desire to avoid steroid treatments, which had previously caused severe adverse effects, Dr. Hauptman prescribed three treatments of Remicade, a relatively new drug that had been developed since Patricia’s surgery in 1997, administered at six-week intervals of 400 milligrams each.

C. Remicade

Remicade is a prescription drug, manufactured by Centocor, that is approved by the Food and Drug Administration (FDA) for the treatment of Crohn’s disease and rheumatoid arthritis.4 An immunomodulator medication, Remicade is designed to suppress the immune system’s inflammatory response to the affected bowel. Patients receive Remicade treatments through intravenous infusions — the medication is injected through an IV catheter in the patient’s arm.

1. The FDA Approval Process

Barbara Matthews, M.D., an FDA administrator from 1994 to 2000, testified as an expert on the FDA approval process. Dr. Matthews was the clinical reviewer of Centocor’s application for FDA approval of Remicade and testified about her knowledge of the drug and her review of the safety and clinical data supporting Cento-*145cor’s application. According to Dr. Matthews, once the FDA approves a drug for prescription use, the drug manufacturer drafts a package insert, which contains the clinical information, warnings, and other information known about the drug. The FDA then reviews the proposed package insert, makes revisions, and ultimately approves the insert for distribution with the drug. According to Dr. Matthews, the purpose of a package insert

is to describe both the safety and efficacy that were reported to [the] FDA and ... provide[ ] information to the physician regarding the types of events, the serious[ ] nature of some of the events, the incidents of the events and, yes, the physician uses [this information to assess the] risk to the patient when they prescribe the medication.

[[Image here]]

The degree of risk to the individual patient really depends on the physician’s knowledge of the patient and then also the information that’s in the label but the label really doesn’t link directly that patient to the degree of risk.

Additionally, once a drug has received FDA approval, the manufacturer must submit periodic safety update reports to the FDA. These post-approval reports contain cumulative summaries of the drug’s safety information, including updated clinical studies and any other medical findings published about the drug. Depending on the number, nature, and severity of events reported for a given adverse reaction, the FDA may recommend that the manufacturer (1) continue to monitor the events, (2) change the warning label, or (3) conduct additional studies. Because the FDA requires continuing studies of the safety and efficacy of the prescription drug, it is common for the package insert to undergo revisions as new information becomes available.

2. The 2001 Remicade Package Insert

At the time of Patricia’s initial Remicade prescription in December 2001, Centocor provided Patricia’s doctors with a package insert that warned Remicade’s use could lead to certain adverse reactions. The package insert included the following warning information regarding lupus-like syndrome:

PRECAUTIONS:

Autoimmunity

Treatment with REMICADE may result in the formation of autoantibodies and, rarely, in the development of a lupus-like syndrome. If a patient develops symptoms suggestive of a lupus-like syndrome following treatment with REMICADE, treatment should be discontinued (see ADVERSE REACTIONS, Autoantibodies/Lwpus-like Syndrome).

[[Image here]]

ADVERSE REACTIONS:

A total of 771 patients were treated with REMICADE in clinical studies. In both rheumatoid arthritis and Crohn’s disease studies, approximately 6% of patients discontinued REMICADE because of adverse experiences. The most common reasons for discontinuation of treatment were dyspnea, urticaria and headache. Adverse events have been reported in a higher proportion of patients receiving the 10 mg/kg dose than the 3 mg/kg dose.

[[Image here]]

Autoantibodies/Lupus-like Syndrome In the ATTRACT rheumatoid arthritis study through week 54, 49% of REMI-CADE-treated patients developed antinuclear antibodies (ANA) between screening and last evaluation, compared to 21% of placebo-treated patients. Anti-dsDNA antibodies developed in ap*146proximately 10% of REMICADE-treat-ed patients, compared to none of the placebo-treated patients. No association was seen between REMICADE dose/schedule and development of ANA or anti-dsDNA.

[[Image here]]

In clinical studies, three patients developed clinical symptoms consistent with a lupus-like syndrome, two with rheumatoid arthritis and one with Crohn’s disease. All three patients improved following discontinuation of therapy and appropriate medical treatment. No cases of lupus-like reactions have been observed in up to three years of long-term follow-up (see PRECAUTIONS, Autoimmunity).

The package insert also included a table noting that serious adverse reactions — including systemic lupus erythematosus syndrome — occurred at frequencies of less than 2% “by body system in all patients treated with REMICADE.”

D. Lupus-Like Syndrome

According to Mary Olsen, M.D., an expert witness hired by Centocor but called by the Hamiltons, lupus-like syndrome, also called drug-induced lupus, has similar characteristics to the autoimmune disorder systemic lupus erythematosus (SLE), except that lupus-like syndrome is caused by a drug. All of the testifying experts generally agreed that symptoms of both lupus-like syndrome and SLE include joint pain and swelling, weight gain, fatigue, unusual weakness, leukopenia, lymphopenia, rash, oral ulcers, fever, and pericarditis. Patients with Crohn’s disease or rheumatoid arthritis, another autoimmune disorder, could also present similar symptoms to lupus-like syndrome, making it sometimes difficult to diagnose SLE or drug-induced lupus.

Physicians can conduct lab tests to check for the presence of anti-nuclear antibodies (ANA), double-stranded DNA antibodies (anti-dsDNA), and antihistone antibodies, which are specific indicators that may help a physician diagnose the presence of an autoimmune condition. Although no antibody is definitive of lupus-like syndrome, positive tests for ANA or anti-dsDNA may indicate the patient has lupus-like syndrome. Physicians also use electrophoresis and immunoelectrophoresis as other immunology blood tests to help diagnose lupus. According to Afilia Ertan, M.D., an expert witness for Centocor, the anti-dsDNA test is the most important indicator for diagnosing drug-induced lupus. Additionally, Dr. Olsen testified that antihistone antibodies are classically seen in patients with drug-induced lupus and “as a rule, an antihistone antibody often is a flag [that indicates] a drug-induced problem.” Because Remicade may produce ANA or anti-dsDNA in patients, however, Dr. Olsen explained that it is often difficult to diagnose lupus-like syndrome. A doctor must look to both the laboratory tests and the clinical presentation of symptoms. If a patient has drug-induced lupus rather than SLE, removing the patient from the drug should improve the patient’s lupus-like symptoms.

E. Dr. Bullen and the Remicade Infusions

Dr. Hauptman prescribed three infusions of Remicade over a six-week period from December 2001 to January 2002 and referred Patricia to Dr. Bullen for treatment. Dr. Bullen is an infectious disease specialist and, at that time, operated an infusion clinic in Corpus Christi where Patricia received the Remicade infusions. As the non-prescribing, treating physician, neither Dr. Bullen nor his staff discussed with Patricia the risks inherent in Remi-cade, but they informed her of the poten*147tial risks directly associated with the infusion process. At trial, Dr. Bullen did not recall confirming that Dr. Hauptman had informed Patricia of Remieade’s risks. Dr. Bullen stated that he was aware that Rem-icade could cause lupus-like syndrome, but admitted that he probably did not give Patricia any warnings or instructions concerning the risk of developing lupus-like syndrome.

Polly Swinney, a registered nurse at Bullen’s infusion clinic, took Patricia’s patient history, advised Patricia on the potential infusion-related side effects,5 and monitored Patricia during her Remicade treatments at the infusion center. Before Patricia received her first infusion, Dr. Bullen’s clinic performed a tuberculin skin test, per Dr. Hauptman’s orders, because of the serious risks related to treating tuberculosis-infected patients with Remi-cade.6 After confirming that Patricia did not have tuberculosis, Patricia received her first Remicade infusion at Dr. Bullen’s infusion clinic on December 19, 2001.

F. Centocor’s Informational Video

After Swinney connected Patricia’s IV and started the first Remicade infusion, she showed Patricia an informational video about Remicade and the treatment process, which Centocor had provided to Dr. Bullen. Centocor had submitted the video to the FDA, but the FDA neither approved nor disapproved it.

Dr. Bullen, Dr. Matthews, and Swin-ney generally agreed that the main purpose of the video was to educate patients and make them more comfortable with the infusion process. The video, titled “Patient Guide to Remicade® (infliximab) IV Administration,” was viewed by the jury during trial. It depicts the effects of Remicade on several people and includes statements from a doctor, identified as Alan Safdi, M.D., who explains the Remicade infusion process and warns about some of the drug’s side effects. The video shows several patients receiving Remicade infusions, provides answers to common questions about the treatment process, and shows the patients continuing their daily routines after the treatment while the bottom of the screen states: “RESULTS MAY VARY.” Although Dr. Safdi states that “there are very little side effects that people need to watch for” and adverse reactions are “extremely rare,” he also instructs patients to contact their medical providers if they have any discomfort and states that there have been some reports of serious, life-threatening side effects.

In addition to- Dr; Safdi’s verbal warnings, the video provides several written warnings and disclaimers at the end of the production. It instructs patients to contact their healthcare provider if they have any questions and provides a Remicade website address for further information. The video warns of various risks associated with the infusion process, advises that “[plhysicians should discuss with their patients all potential side effects that may occur during these infusions,” and cautions that the “video should not be used as a substitute for talking with your doctor.” For the treatment of fistulizing Crohn’s *148disease,7 the video warning states that “[t]he safety and efficacy of therapy continued beyond three doses have not been studied.” After listing a series of potential side effects, the warning instructs patients to “see the accompanying Full Prescribing Information.” It is undisputed that the video did not mention lupus-like syndrome as a potential side effect of Remicade.

Swinney testified that the infusion clinic received the videotapes from Centocor in cellophane-wrapped boxes that usually contained the video, informational brochures about Remicade, and package inserts that provided more extensive details about the drug. She claimed that after showing Patricia the video, she placed it back in the box with the written materials on top and gave it to Patricia. At trial, Patricia denied receiving any written information about Remicade, but stated that she never looked in the box that Swinney gave her or reviewed any additional written information about Remicade. During one of Patricia’s infusions, Swinney gave Patricia a second Centocor video to give to Patricia’s sister, who had rheumatoid arthritis. The second video contained the same visual content but was enclosed in different packaging material. The box containing the second video had a plastic sleeve on the inside cover that contained the Remicade package insert and an informational brochure.8

Patricia reported an excellent response to her Remicade treatments at Dr. Bul-len’s infusion clinic, and the treatments helped relieve the symptoms of her Crohn’s disease. After her first two Remi-cade infusions, Dr. Hauptman performed a colonoscopy on Patricia, which revealed that Patricia no longer had abnormalities in her small intestine. Following her third Remicade infusion, Dr. Hauptman believed that Patricia’s Crohn’s disease was in remission. Dr. Hauptman then planned to continue monitoring Patricia’s condition before determining whether Patricia needed a Remicade maintenance dose every eight weeks. Throughout her follow-up appointments with Dr. Hauptman, Patricia reported that she was having no problems with her bowels, leading Dr. Hauptman to conclude that the Remicade treatments were successful and that Patricia’s Crohn’s disease remained in remission. It is undisputed that, since taking Remicade, Patricia’s Crohn’s disease has been asymptomatic.

G. Dr. Pop-Moody Treats Patricia’s Arthritis

In the weeks following her initial treatments with Remicade, Patricia experienced severe arthritis-like pains in her joints. Patricia’s family physician referred her to a local rheumatologist, Adriana Pop-Moody, M.D. During her initial visit with Dr. Pop-Moody, Patricia explained her recent treatments with Remicade and told Dr. Pop-Moody that it had dramatically improved her Crohn’s condition and that her arthritis pains had markedly improved after her first three doses of Remicade. In April 2002, Dr. Pop-Moody prescribed treatments including additional Remicade infusions at regular intervals to treat Patricia’s joint pains. Between April 2002 and September 2003, Patricia received fourteen additional Remicade infusions at the Corpus Christi Medical Center.

*149Patricia continued to experience severe joint pain. Each treatment provided temporary relief, but the relief periods dwindled. In an attempt to improve Patricia’s condition, Dr. Pop-Moody decreased the time between infusions and increased the dosage of Remicade. Dr. Pop-Moody and Patricia remained convinced that Remi-cade was providing Patricia temporary relief from her joint pains. In her practice, Dr. Pop-Moody routinely diagnosed and treated patients with lupus. Even though Patricia tested positive for SLE in June 2002 — a condition similar to drug-induced lupus and a potential side effect of Remi-cade — because of Patricia’s complicated condition, Dr. Pop-Moody did not diagnose Patricia with drug-induced lupus at that time. In April 2008, Dr. Pop-Moody reviewed treatment options with Patricia and discussed that she might need to stop the Remicade infusions. Dr. Pop-Moody informed Patricia that she may have lupus-like syndrome, but Patricia’s medical records indicate that, despite this risk, Patricia desired to continue taking Remicade.

H. The Houston Doctors

Because of Patricia’s increasing joint pain and Dr. Pop-Moody’S inability to determine the cause of Patricia’s continuing ailments, Dr. Pop-Moody referred Patricia to rheumatologists at the University of Texas Health Science Center in Houston, where she saw Maureen D. Mayes, M.D., Noranna B. Warner, M.D., and Leslie Wilson, M.D. (collectively, the Houston Doctors). In response to Dr. Pop-Moody’S request, the Houston Doctors examined Patricia’s symptoms and Dr. Mayes and

Dr. Wilson made the following assessment in September 2003:

1. Symmetric polyarthritis involving the hands, elbows, shoulders, knees, and feet that could be consistent with lupus (potentially drug-induced by Remicade), although the literature is limited in supportive evidence of this entity. There have been several studies showing that the presence of double stranded DNA antibodies in patients who receive Remicade is not uncommon; however, there have been limited . cases of lupus-like syndrome seen with this medication. The patient, at this time, does appear to have a syndrome that could be classified as systemic lupus erythematosus. The clinical picture is less consistent with sarcoidosis or9 arthritis associated with sarcoidosis. The clinical presentation could be consistent with enteropathic arthritis. This could be arthritis associated with hepatitis C virus.

2. Leukopenia — possibly secondary to lupus-like syndrome or a side effect of Imuran.10

The Houston Doctors stopped Patricia’s Remicade treatments and, instead, prescribed steroids for her joint pains. At her October 2003 follow-up appointment, Patricia reported that she felt much better than she did at her previous visit. Dr. Warner sent Dr. Pop-Moody a status report, which documented Patricia’s improvement and stated in pertinent part:

*150The patient presented to our clinic for evaluation of possible lupus-like syndrome which may have been induced by Remicade therapy. Upon evaluation by her rheumatologist in Corpus Christi, the patient was found to have a positive ANA and positive double-stranded DNA antibodies. Other significant history included the episode of pericarditis in January of 2003 and the presence of leu-kopenia (absolute lymphopenia). Upon initial presentation to our clinic a few weeks ago, the patient’s physical exam showed significant tenderness on internal and external rotation of both of her shoulders. There was also some swelling of the fingers, particularly in the PIP joints bilaterally. There was decreased hand grip secondary to pain.

Today, the patient states that she is much improved since her previous visit....

[[Image here]]

ASSESSMENT: Symmetric polyarthri-tis involving hands, shoulders, knees, feet, which is consistent with a lupus-like syndrome (potentially drug-induced by Remicade). The patient has improved on an increased dose of Imuran 150 mg daily, increased from 100 mg daily. The patient also had increased her predni-sone dose from 10 mg daily to 15 mg daily.

Within a few months of ceasing the Remi-cade infusions, Patricia’s lupus-like symptoms subsided. Patricia’s arthritic symptoms also improved dramatically after she ceased taking Remicade.

II. Procedural Background

A. Trial Court Proceedings

The case proceeded to a multi-week jury trial. At trial, Patricia and her doctors gave conflicting testimony about their conversations concerning the risks and potential adverse effects associated with Remi-cade. Specifically, Dr. Hauptman and Dr. Pop-Moody testified that they fully informed Patricia about the risks of developing lupus-like syndrome while Patricia averred that she received no such warning from either doctor when they initially prescribed Remicade treatments. Although Patricia admitted that she was informed by her physicians of certain risks associated with Remicade, including the rare risk of cancer, she testified that her doctors made no mention of the risk of developing lupus-like syndrome. She further stated that the risk of lupus-like syndrome was something that she would have wanted to know, that it “would have impacted [her] decision,” and that “[t]he question of lupus would have made [her] stop and ask more questions before [she] made a decision [to take Remicade].” Regardless of the conflicting testimony, it is undisputed that all of Patricia’s doctors were aware of the risk of lupus-like syndrome when they chose to prescribe and treat Patricia with Remi-cade. And according to all of the doctors who testified on the subject, lupus-like syndrome can be difficult to diagnose.

At the charge conference, the Hamiltons abandoned their original failure-to-warn claim and proposed separate jury questions on claims against Centocor for (1) fraud, (2) negligent misbranding, (3) negligent marketing, (4) negligent undertaking, (5) misrepresentation to Patricia’s prescribing physicians concerning the risk of lupus-like syndrome, and (6) misrepresentation to Patricia’s prescribing physicians concerning the risk of hepatitis C and liver damage. Centocor raised several objections to the Hamiltons’ proposed charge, including that it was error for the trial court to submit separately the claims for negligent misbranding, negligent marketing, negligent undertaking, misrepresentation regarding lupus-like syndrome, misrepresentation regarding hepatitis C and *151liver conditions, and fraud. Centocor argued that each of the Hamiltons’ claims were premised on a single failure-to-warn cause of action, the separate questions were duplicative, and the trial court should only submit one failure-to-warn claim to the jury. The trial court judge overruled each of Centocor’s objections and submitted all of the Hamiltons’ questions to the jury. Centocor also repeatedly raised the learned intermediary doctrine, contesting its duty to warn Patricia directly. It argued that the record lacked expert testimony to (1) support the Hamiltons’ claims that Centocor’s warnings to Patricia and her prescribing doctors were inadequate, and (2) prove that the allegedly inadequate product warning was the producing cause of Patricia’s injuries. Additionally, Cento-cor argued that there was no evidence Centocor breached the standard of care.

Before the trial court submitted the charge to the jury, the court granted a directed verdict in favor of Dr. Bullen and his infusion clinic, finding that Dr. Bullen and his staff had no duty to warn Patricia of the risks associated with Remicade because Dr. Bullen was not the prescribing physician.

The jury found Centocor liable for fraud, misrepresentation to Patricia’s doctors, negligent misbranding, negligent marketing to. Patricia’s doctors, and negligent undertaking. The jury awarded Patricia $1.2 million for past pain and mental anguish, $1 million for future pain and mental anguish, $1.1 million for past physical impairment, and $65,908 for past medical expenses. It also awarded Thomas $50,000 for loss of consortium and household services. The jury apportioned liability for the Hamiltons’ damages, finding Centocor 85% liable, Dr. Pop-Moody 10% liable, and Dr. Hauptman 5% liable. The jury determined that the. Hamiltons established fraud by clear and convincing evidence and awarded Patricia $15 million and Thomas $1 million in exemplary damages based on the fraud claim.

Before the trial court entered judgment on the jury’s verdict, Dr. Hauptman and Dr. Pop-Moody settled with the Hamil-tons, and the Hamiltons nonsuited those defendants.11 The trial court denied Cen-tocor’s motion for judgment notwithstanding the verdict. On February 13, 2007, the trial court apportioned responsibility, applied settlement credits, applied the exemplary damages caps, and entered judgment against Centocor, awarding a total of $4,687,461.70 to Patricia and $120,833.71 to Thomas in actual damages, punitive damages, and interest. The trial court entered judgment on the Hamiltons’ fraud claim only but stated that, in the alternative, the Hamiltons should recover actual damages under the various other theories presented to the jury.

B. Court of Appeals

Centocor timely appealed, arguing first that the learned intermediary doctrine precluded the Hamiltons’ recovery because, as a matter of law, Centocor’s warnings to Patricia’s prescribing physicians were adequate and Centocor had no duty to warn Patricia directly.12 310 S.W.3d 476, 499 (Tex.App.-Corpus Christi 2010). The *152court of appeals disagreed, however, and affirmed the trial court’s judgment on the Hamiltons’ fraud claim but reversed the damages award for future pain and mental anguish. Id. at 522.

The court of appeals first examined Cen-tocor’s duty under the learned intermediary doctrine, which generally limits á prescription drug manufacturer’s duty to warn of its product’s risks and potential side effects to prescribing physicians, who then serve as “learned intermediaries” and assume the duty to pass the warnings on to patients. See id. at 499-508. Relying on reasoning from the New Jersey Supreme Court’s opinion in Perez v. Wyeth Laboratories, 161 N.J. 1, 734 A.2d 1245, 1246^47 (1999), the court of appeals adopted “an exception to the learned intermediary doctrine when a drug manufacturer directly advertises to its consumers in a fraudulent manner.” 310 S.W.3d at 480-81. Guided by the Perez court’s decision, the court of appeals held “that when a pharmaceutical company directly markets to a patient, it must do so without fraudulently misrepresenting the risks associated with its product.” Id. at 508. The court therefore dismissed Centocor’s arguments based on the learned intermediary doctrine and, in a footnote, stated: “[W]e hold today that Centocor cannot rely on its adequate warnings to Patricia’s physicians when it directly misrepresented its product’s dangerous propensities to Patricia.” Id. at 508 & n. 18 (emphasis added).

The court of appeals next considered Centocor’s argument that the Hamiltons failed to present legally and factually sufficient evidence of causation and held that the Hamiltons met their burden of proof on causation.13 Id. at 508-12. The court then overruled Centocor’s claims that the trial court erred because (1) no expert testified that the Remicade warnings were inadequate, (2) the side effect of lupus-like syndrome did not make Remicade unreasonably dangerous, and (3) the Hamiltons presented no evidence that a different warning by Centocor would have prevented Patricia’s injuries. Id. at 512-16. The court of appeals overruled or dismissed as *153moot Centocor’s remaining issues, reversed the trial court’s damages award for future pain and mental anguish,14 and affirmed the trial court’s judgment on the Hamiltons’ fraud claim. See id. at 516-22.

C. Centocor’s Petition for Review

Centocor timely petitioned this Court for review, and raises four issues on appeal: (1) the court of appeals erred by creating an advertising exception to the learned intermediary doctrine, and the doctrine applies, thereby limiting Cento-cor’s duty to warn to Patricia’s prescribing physicians only and barring the Hamiltons’ claims; (2) the Hamiltons failed to present any expert testimony that the warning in Centocor’s informational video was inadequate, and the appellate court erred by considering Dr. Matthews’s testimony about the FDA approval process and the FDA regulations as sufficient evidence to show that the video’s allegedly inadequate warning made Remicade unreasonably dangerous; (3) the Hamiltons failed to show any evidence of causation because they presented no expert testimony of causation or epidemiological studies, but instead relied on the package insert and expert witnesses’ unsupported references to clinical trials; and (4) the appellate court erred in affirming the Hamiltons’ fraud claim because there was no evidence of any mens rea on the part of Centocor, which was isolated from the patient by the intermediary doctors, and that Patricia could not have relied on any alleged misrepresentation because she continued to take Remicade even after suing Centocor. In response to Centocor’s first issue, the Hamiltons raise a conditional cross-issue, alleging that if we hold that the learned intermediary doctrine applies, we should reinstate the Hamiltons’ claims against Dr. Bullen, the non-prescribing, treating physician, because Centocor deliberately used Dr. Bullen to provide direct marketing materials outside of the context of the prescribing physician’s doctor-patient relationship. We granted both petitions. 54 Tex.Sup.CtJ. 1578 (Aug. 26, 2011).15 We address the parties’ arguments in turn.

III. Learned Intermediary Doctrine

Generally, a manufacturer is required to provide an adequate warning to *154the end users of its product if it knows or should know of any potential harm that may result from the use of its product. See, e.g., Bristol-Myers Co. v. Gonzales, 561 S.W.2d 801, 804 (Tex.1978). In certain contexts, however, the manufacturer’s or supplier’s duty to warn end users of the dangerous propensities of its product is limited to providing an adequate warning to an intermediary, who then assumes the duty to pass the necessary warnings on to the end users. See, e.g., Aim v. Aluminum Co. of Am., 717 S.W.2d 588, 590-92 (Tex.1986); Richard C. Ausness, Learned Intermediaries and Sophisticated Users: Encouraging the Use of Intermediaries to Transmit Product Safety Information, 46 Syraouse L. Rev. 1185, 1195-96 (1996). It is firmly established in Texas that whether a duty exists is ordinarily a legal matter for the court to decide. See, e.g., Humble Sand & Gravel, Inc. v. Gomez, 146 S.W.3d 170, 181 (Tex.2004). Within the context of prescription drug manufacturers, the underlying premise for the learned intermediary doctrine is that prescription drugs are complex and vary in effect, depending on the unique circumstances of an individual user, and for this reason, patients can obtain them only through a prescribing physician. See Reyes v. Wyeth Labs., 498 F.2d 1264, 1276 (5th Cir.1974), cert. denied, 419 U.S. 1096, 95 S.Ct. 687, 42 L.Ed.2d 688 (1974).

Centocor argues that the learned intermediary doctrine applies and therefore it had no duty to warn Patricia directly of the risks and potential side effects associated with Remicade. It asserts that it provided warnings of specific side effects and risks associated with Remicade to Patricia’s doctors through the FDA-approved package insert, which Patricia did not challenge as inadequate. Centocor further argues that the appellate court erred by adopting a DTC advertising exception. Centocor claims that DTC advertising does not threaten the physician-patient relationship, but helps educate consumers about available medications — sometimes causing patients to seek treatment for ailments they otherwise would not have treated or sometimes even discovered. Additionally, Centocor points out that the FDA and other courts have recognized that over-warning can confuse the public and, ultimately, can harm treatment efforts by scaring patients from taking the necessary risks associated with some prescription drugs.

In response, the Hamiltons argue that Centocor cannot benefit from the learned intermediary doctrine because it provided an inadequate and misleading warning to the “learned intermediary” — Patricia’s prescribing and treating physicians. The Hamiltons contend that Centocor misrepresented the risks and side effects associated with Remicade to Patricia and her doctors by conveying instances of lupus-like syndrome observed only in clinical studies instead of all reported cases, thereby preventing Centocor from relying on the defense of the learned intermediary doctrine. Alternatively, if the learned intermediary doctrine applies, the Hamiltons urge us to adopt the DTC advertising exception and affirm the court of appeals’ judgment because when a drug manufacturer directly markets its product to patients, that manufacturer should have a duty, at minimum, to present non-misleading information about the drug and must be liable for its fraudulent or intentionally misleading marketing.

A. The Learned Intermediary Doctrine in Texas Jurisprudence

The learned intermediary doctrine has been part of Texas jurisprudence for many years. See, e.g., Gravis v. Parke-Davis & Co., 502 S.W.2d 863, 870 (Tex.Civ.App.-Corpus Christi 1973, writ ref'd n.r.e.). In *155Gravis, the court of appeals held that it was unreasonable for the law to impose a duty on prescription drug manufacturers to “specifically warn each and every patient that receives drugs prescribed by the physician or other authorized persons” and outlined the underlying rationale for the doctrine:

The entire system of drug distribution in America is set up so as to place the responsibility of distribution and use upon professional people. The laws and regulations prevent prescription type drugs from being purchased by individuals without the advice, guidance and consent of licensed physicians and pharmacists. These professionals are in the best position to evaluate the warnings put out by the drug industry. Our holding in no way relieves the drug company in their duty to warn or to provide a product free of defects.

Id. Since the Thirteenth Court of Appeals’ opinion in Gravis — the same court from which we consider the instant case — many Texas courts of appeals have applied the learned intermediary doctrine in prescription drug produets-liability cases. See, e.g., Wyeth-Ayerst Labs. Co. v. Medrano, 28 S.W.3d 87, 91 (Tex.App.-Texarkana 2000, no pet.) (“In prescription drug cases, the courts have found that it is reasonable for the manufacturer to rely on the health care provider to pass on its warnings. This is reasonable because the learned intermediary understands the propensities and dangers involved in the use of a given drug, and as the prescriber, he stands between this drug and the ultimate consumer.”).16

We first discussed the doctrine in Aim, v. Aluminum Co. of America, a case involving an aluminum bottle cap manufacturer’s *156duty to warn end users of hazards associated with its product. 717 S.W.2d at 590-92. Aim sued the manufacturer of soda bottle caps, claiming that an aluminum bottle cap popped off a soda bottle and struck him in the eye. Id. at 590. AJm claimed that Alcoa, the manufacturer of the machine that fastened the caps to the soda bottles, had a duty to warn him of the risk that a cap could pop off. Id. Although Alcoa manufactured the bottle-fastening machine, .the machine was owned and operated by an independent bottler. Id. at 589-90. Alcoa did not control the bottling process or sell the bottled soda, nor did it have any practical way of reaching consumers with any warning. Id. at 592. Because of Alcoa’s limited connection to the end user of the consumer product, we recognized the need for an intermediary:

[A] manufacturer or supplier may, in certain situations, depend on an intermediary to communicate a warning to the ultimate user of a product. However, the mere presence of an intermediary does not excuse the manufacturer from warning those whom it should reasonably expect to be endangered by the use of its product. The issue in every ease is whether the original manufacturer has a reasonable assurance that its warning will reach those endangered by the use of its product.

Id. at 591. We then analogized Aim’s position to that of a bulk supplier “who sells a product to another manufacturer or distributor who in turn packages and sells the product to the public.” Id. at 592. Because the bulk-supplier rationale applied to Alcoa’s duty to warn in that case, we explained:

Alcoa should be able to satisfy its duty to warn consumers by proving that its intermediary ’was adequately trained and warned, familiar with the propensities of the product, and capable of passing on a warning. But, if Alcoa failed to adequately warn and train [the intermediary] or if [the intermediary] was incapable of passing on the received warning, Alcoa would not have discharged its duty to the ultimate consumer.

Id. While Alcoa did not have a duty to warn Aim directly, it still had a duty to warn the bottler, and we concluded that the record contained some evidence to support the jury’s finding that Alcoa’s warning to the bottler was inadequate. Id. at 593-95.

Although