ARGUED: George Clemon Freeman, Jr., Hunton & Williams, Richmond, Virginia, for Appellant. Patrick Allen Malone, Stein, Mitchell & Mezines, Washington, D.C., for Appellee. ON BRIEF: David F. Dobbins, John D. Winter, Patterson, Belknap, Webb & Tyler LLP, New York City; Robert W. Sparks, Jones, Day, Reavis & Pogue, New York City, for Appellant. Robert F. Muse, Stein, Mitchell & Mezines, Washington, D.C.; Sandra L. Hughes, Chevy Chase, Maryland, for Appellee.

Before WILKINS and LUTTIG, Circuit Judges, and JOSEPH F. ANDERSON, Jr., United States District Judge for the District of South Carolina, sitting by designation.

Affirmed by published opinion. Judge ANDERSON wrote the opinion, in which Judge WILKINS and Judge LUTTIG joined.

Antonio Benedi suffered severe liver damage allegedly as a result of combining Extra-Strength Tylenol and alcohol. Benedi filed suit against McNeil-P.P.C. ("McNeil"), the manufacturer of the drug, for negligent failure to warn of the possible dangers of mixing Tylenol with alcohol and for breach of implied warranties. After a three-day trial, the jury returned a verdict for Benedi. McNeil appeals claiming that judgment as a matter of law should be entered in its favor because Benedi failed to prove that Tylenol caused his injuries and because the Tylenol label contained adequate warnings. In the alternative, McNeil seeks a new trial based upon the district court's allegedly erroneous admission of certain evidence and testimony. For the reasons discussed below, we affirm the judgment of the district court.

Benedi was admitted to the hospital on February 10, 1993 in a coma and near death due to liver and kidney failure. Benedi, who regularly consumed three to four glasses of wine a night during the week and sometimes more on the weekend, had been taking Extra-Strength Tylenol in normal, over-the-counter doses since February 5, 1993 for flu-like aches.1 Blood tests performed shortly after his admission to the hospital revealed that Benedi still had therapeutic amounts of Tylenol in his blood, although he had not taken any for several days. He did not, however, have any alcohol in his blood. The hospital doctor determined that Benedi's injuries were most likely caused by acetaminophen (Tylenol) toxicity.

On the night of February 12-13, 1993, Benedi underwent an emergency liver transplant. The new liver, which did not match Benedi's blood type, restored Benedi's liver and kidney functioning. He was released from the hospital on March 22, 1993. However, Benedi may have to undergo kidney dialysis in a few years because the drugs that he must take to keep his body from rejecting the new liver may cause deterioration in his kidneys.

The hospital's pathologist examined Benedi's damaged liver and concluded that the injury was characteristic of acetaminophen toxicity. He did not find any evidence of a viral infection in the liver. The transplant surgeon also examined Benedi's damaged liver and confirmed the pathologist's opinion. At trial, several liver disease experts also confirmed Benedi's treating physicians' diagnosis of acetaminophen toxicity.

The infectious disease specialist who treated Benedi at the hospital noted that Benedi had tested negative for herpes simplex virus on the day he was admitted, but that two days later he tested positive. The specialist concluded that the later positive result was likely due to contamination from blood transfusions. During Benedi's first two days in the hospital, he received eight units of blood from eight different donors. The transplant surgeon testified that if herpes had destroyed Benedi's liver, the disease would have also rapidly infected and destroyed the transplanted liver.

Both parties retained liver toxicity experts to determine whether herpes simplex could be ruled out as a cause of Plaintiff's liver disease. These experts compared slides of Benedi's liver with control slides of a liver infected by the herpes simplex virus. McNeil's expert testified that the slides did not reveal the "signature" for acetaminophen liver toxicity, and that the slides of Benedi's liver matched the control slides. McNeil's expert took photos of the slides of Benedi's liver and the control slides and testified to their similarity at trial. He concluded that the herpes simplex virus was the cause of Benedi's liver damage.

In contrast, plaintiff's expert testified that these photos were misleading. After examining the actual slides, he concluded that the slide of Benedi's damaged liver showed negative for the herpes simplex virus.

At trial, Benedi called two liver disease specialists who both testified that a warning of the possible danger to heavy drinkers from combining alcohol and acetaminophen should have been placed on the Tylenol label since the mid-1980s. These experts described exactly how the alcohol-acetaminophen mixture can become a toxin in the liver. They cited numerous treatises and articles published in medical journals prior to 1993 that describe the increased risk of liver injury when acetaminophen is combined with alcohol. One of plaintiff's experts referred to sixty reports that McNeil had received by the end of 1992 documenting cases of liver injury associated with combining therapeutic doses of Tylenol with alcohol.

The provisions of the Federal Food, Drug, and Cosmetic Act, 21 U.S.C. Secs. 301-395 (1972 & Supp.1995), as well as the regulations of the Food and Drug Administration ("FDA"), govern the content of labels and warnings on drugs. In 1972, the FDA convened a panel of independent experts to evaluate over-the-counter analgesics, such as aspirin and acetaminophen. The panel determined that chronic alcoholics were susceptible to liver damage from acetaminophen, but that acetaminophen was safe in normal over-the-counter doses. The panel continued its research and hearings, and, in 1988, the FDA adopted the panel's findings and decided against requiring a warning. Nevertheless, at the suggestion of the FDA, McNeil agreed in the summer of 1993 (after Benedi's injury) to include a warning that persons who regularly consume three or more alcoholic drinks a day should consult a physician before using Tylenol or any other pain reliever.

Benedi filed suit against McNeil in March 1994, and the case went to the jury on theories of breach of implied warranty and negligent failure to warn. After a three-day trial, the jury awarded Benedi $7,850,000 in compensatory damages and $1,000,000 in punitive damages. The district court reduced the punitive damage award to $350,000, the statutory maximum under Virginia law. Va.Code Ann. Sec. 8.01-38.1 (Michie 1992). Finding that the issues of causation and warnings were properly submitted to the jury and that the evidence supported an award of punitive damages, the district court denied McNeil's post-trial motions.

McNeil appeals alleging that Benedi failed to prove that Tylenol caused his injuries and that the Tylenol label contained adequate warnings. Accordingly, McNeil asks us to reverse the judgment below and enter judgment as a matter of law in its favor. In the alternative, McNeil seeks a new trial because, it claims, the trial court erroneously admitted certain evidence and testimony. Specifically, McNeil argues that the district court erred in admitting summaries and case reports relating to prior cases of liver damage resulting from alcohol-acetaminophen interaction. McNeil also claims prejudice from plaintiff's introduction of a subsequent remedial measure by McNeil--the summer 1993 warning. Finally, McNeil argues that the evidence does not support an award of punitive damages.

McNeil appeals the district court's denial of its motion for judgment as a matter of law and its motion for a new trial. We review a denial of a motion for judgment as a matter of law de novo. Benesh v. Amphenol Corp. (In re Wildewood Litigation), 52 F.3d 499, 502 (4th Cir.1995). The court must determine whether substantial evidence exists upon which the jury could find for the appellee. Id. "The question is whether a jury, viewing the evidence in the light most favorable to [Benedi], could have properly reached the conclusion reached by this jury." Id. A denial of a motion for a new trial is reviewed for abuse of discretion. Id. Likewise, we review evidentiary rulings for an abuse of discretion. United States v. D'Anjou, 16 F.3d 604, 610 (4th Cir.), cert. denied, --- U.S. ----, 114 S. Ct. 2754, 129 L. Ed. 2d 871 (1994). Finally, whether substantial evidence existed to submit the issue of punitive damages to the jury is a question of law and is, therefore, reviewed de novo. Richmond, Fredericksburg & Potomac R.R. v. United States, 945 F.2d 765, 768-69 (4th Cir.1991), cert. denied, 503 U.S. 984, 112 S. Ct. 1667, 118 L. Ed. 2d 388 (1992).

McNeil first contends that Benedi failed to prove that combining alcohol with therapeutic doses of acetaminophen causes liver disease. McNeil argues that the district court erred in admitting Benedi's allegedly unreliable expert testimony. McNeil asserts that without this expert testimony, Benedi lacked sufficient evidence to submit the issue of causation to the jury.

In the recent case of Daubert v. Merrell Dow Pharmaceuticals, Inc., --- U.S. ----, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993), the Supreme Court set forth the proper standard for admission of expert testimony under the Federal Rules of Evidence. Until Daubert, the courts were split on this issue--some had applied, and some had rejected, the "general acceptance" standard as set forth in Frye v. United States, 293 F. 1013 (D.C.Cir.1923). The Daubert Court noted that the Federal Rules of Evidence, not Frye, govern the admissibility of expert scientific evidence, because Frye predated the Federal Rules of Evidence by half a century. Rule 702 of the Federal Rules of Evidence provides: "If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise." Fed.R.Evid. 702.

In Daubert, the Court elaborated on this rule, noting that scientific testimony or evidence must be both relevant and reliable. "Proposed testimony must be supported by appropriate validation--i.e., 'good grounds,' based on what is known. In short, the requirement that an expert's testimony pertain to 'scientific knowledge' establishes a standard of evidentiary reliability." Daubert, --- U.S. at ----, 113 S. Ct. at 2795. In determining whether to admit expert scientific testimony, the district court must determine whether the expert's testimony is based on scientific knowledge that will assist the trier of fact in understanding or determining a fact in issue. Id. at ----, 113 S.Ct. at 2796 (citing Fed.R.Evid. 104(a)).

The Court offered the following general guidelines:

Ordinarily, a key question to be answered in determining whether a theory or technique is scientific knowledge that will assist the trier of fact will be whether it can be (and has been) tested....

Another pertinent consideration is whether the theory or technique has been subjected to peer review and publication....

Additionally, in the case of a particular scientific technique, the court ordinarily should consider the known or potential rate of error, and the existence and maintenance of standards controlling the technique's operation.

Finally, "general acceptance" can yet have a bearing on the inquiry.

Id. at ---- - ----, 113 S. Ct. at 2796-97 (citations omitted). In offering these guidelines, the court emphasized that it was not formulating a rigid test or checklist, relying instead on the ability of federal judges to properly determine admissibility. Id. at ----, 113 S.Ct. at 2797 ("[t]he inquiry envisioned by Rule 702 is, we emphasize, a flexible one"). In conclusion, the Court held that the Federal Rules of Evidence, especially Rule 702, "assign to the trial judge the task of ensuring that an expert's testimony both rests on a reliable foundation and is relevant to the task at hand." Id. at ----, 113 S.Ct. at 2799.

McNeil contends that because Benedi's experts did not rely upon epidemiological data in formulating their opinions, their testimony is inadmissible under Daubert. However, we do not read Daubert as restricting expert testimony to opinions that are based solely upon epidemiological data. Daubert merely requires that the expert testimony be both relevant and reliable; and Daubert clearly vests the district courts with discretion to determine the admissibility of expert testimony. Under the Daubert standard, epidemiological studies are not necessarily required to prove causation, as long as the methodology employed by the expert in reaching his or her conclusion is sound.

Benedi's treating physicians based their conclusions on the microscopic appearance of his liver, the Tylenol found in his blood upon his admission to the hospital, the history of several days of Tylenol use after regular alcohol consumption, the liver enzyme blood level, and the lack of evidence of a viral or any other cause of the liver failure. Benedi's other experts relied upon a similar methodology: history, examination, lab and pathology data, and study of the peer-reviewed literature. We conclude that the district court did not abuse its discretion when it determined that the methodology employed by Benedi's experts is reliable under Daubert. We will not declare such methodologies invalid and unreliable in light of the medical community's daily use of the same methodologies in diagnosing patients.

In City of Greenville v. W.R. Grace & Co., 827 F.2d 975 (4th Cir.1987), this court recognized that expert testimony need not be based upon identical case studies or epidemiological data. In that case, the City of Greenville brought suit after Grace used building materials containing asbestos in constructing Greenville's city hall building. The City's expert testified that even very low levels of asbestos could cause serious harm. He based this testimony upon scientific inferences drawn from data showing a correlation between asbestos-related diseases and exposure to asbestos at relatively high levels. The City's expert stated "that in the absence of scientific studies concerning exposure to low levels of asbestos, one technique accepted in the scientific community for predicting the risks associated with low-level exposures is to extrapolate the risk downward from results obtained in studies involving high-level exposures." Id. at 980 n.2. Grace argued that the plaintiffs evidence was insufficient, however, to prove that low levels of asbestos posed a serious risk for the building's occupants.

We held that the plaintiff's expert's testimony could reasonably support the jury's finding for the City and that the defendant should not be allowed "to escape liability simply because no occupant of the city hall has yet developed an asbestos-related disease, or because there are, as yet, no epidemiological studies concerning the health risks associated with asbestos contamination of office buildings." Id.

In the instant case, the district judge found the experts on both sides qualified and allowed their testimony under Federal Rule of Evidence 702. The district court acted within its discretion when it found that Benedi's experts satisfied Daubert 's test of relevance and reliability; therefore, the district judge properly admitted their testimony.

Because Benedi's experts presented sufficient evidence on the issue of causation, the district judge properly submitted that issue to the jury. It was the jury's role to assess the weight and credibility of the evidence, and the jury found that Benedi proved causation. In reviewing the jury's verdict, we must determine whether a jury, viewing the evidence in the light most favorable to Benedi, could have properly found that Tylenol caused Benedi's liver failure. See Benesh v. Amphenol Corp. (In re Wildewood Litigation), 52 F.3d 499, 502 (4th Cir.1995). We find that ample evidence existed from which a reasonable jury could find for Benedi; therefore, the district judge properly denied McNeil's motion for judgment as a matter of law.

Next, McNeil contends that the district court erred in admitting into evidence case reports, known as Drug Experience Reports (DERs), and summary charts of the DERs. The district court found that the reports were not hearsay, because Benedi offered them to prove notice, and not for the truth of the matter asserted in them. First, McNeil asserts that the district court should have excluded the reports under Federal Rule of Evidence 403 because the danger of unfair prejudice outweighed any probative value the reports might have.2 Second, McNeil contends that the cases reported were not similar to Benedi's case and, therefore, are not admissible to prove notice. McNeil points out that alcohol history was unknown in some of the reports and that the acetaminophen dose was unknown in others. McNeil also notes that most of the reports involving alcohol consumption involved chronic alcoholics, but that Benedi was not an alcoholic. Finally, McNeil claims that these reports are inadmissible hearsay because Benedi relied upon them for the truth of the matter contained in them, and not just to prove notice.

McNeil relies upon Ellis v. International Playtex, Inc., 745 F.2d 292 (4th Cir.1984), to support its argument that the reports were inadmissible. In Ellis, however, this court merely held that the lower court did not abuse its discretion in excluding prior consumer complaints under Rule 403. We recognized that the complaints in dispute there related solely to the issue of notice, and that the district court's statement to the jury that the defendant had received prior complaints eliminated the need for the complaints themselves to be admitted. Id. at 305. Ellis does not require all district courts presented with prior reports to rule in the same manner. Rather, district judges have the discretion either to admit the prior reports or to exclude them under Rule 403 and instruct the jury that the defendant received notice through prior reports.

More on point is Worsham v. A.H. Robins Co., 734 F.2d 676 (11th Cir.1984), in which the court of appeals upheld the trial court's admission of prior adverse reaction reports because the reports went to the issue of notice. The Worsham court held that the reports were properly admitted into evidence as proof of the defendant's state of mind and, therefore, did not constitute hearsay. Id. at 687. The district court instructed the jury that it could only consider the reports as proof of notice and not as evidence that the information contained in them was true. See also Kehm v. Procter & Gamble Mfg. Co., 724 F.2d 613, 625-26 (8th Cir.1983) (holding that prior complaints of adverse reactions to a product are admissible to prove notice).

We find that the district court did not abuse its discretion in admitting the DERs and case summaries, because the plaintiff offered the evidence solely to prove notice. In addition, the district judge properly gave a limiting instruction to the jury that it could only consider the DERs and case summaries as evidence of notice to McNeil, and not for the truth of the matter contained in them.

We also find that the district judge did not err in refusing to exclude these reports under Rule 403. The probative value of these reports was not outweighed by the danger of unfair prejudice, because the reports were highly probative on the issue of notice and because McNeil was free to, and did, offer testimony rebutting the significance of these reports.

Furthermore, the dissimilarities between Benedi's situation and those reported in the DERs do not affect the admissibility of the evidence, but rather go to the weight that the jury gives to the evidence. Kehm, 724 F.2d at 626. McNeil explored these dissimilarities at length. Whether the dissimilarities were significant was for the jury to determine. When prior incidents are admitted to prove notice, the required similarity of the prior incidents to the case at hand is more relaxed than when prior incidents are admitted to prove negligence. The incidents need only be sufficiently similar to make the defendant aware of the dangerous situation. Cf. Gardner v. Southern Ry. Systems, 675 F.2d 949, 952 (7th Cir.1982); Young v. Illinois Central Gulf R.R., 618 F.2d 332, 339 (5th Cir.1980).3 Because all of the reports at issue involved liver injury associated with therapeutic doses (near or slightly above the recommended doses)4 of Tylenol in alcohol drinkers, we find that the DERs were sufficiently similar to Benedi's case to be admissible on the issue of notice.

Next, McNeil contends that insufficient evidence existed to submit to the jury the issue of negligent failure to warn. In support of its position, McNeil relies upon Thomas v. Hoffman-LaRoche, Inc., 949 F.2d 806 (5th Cir.), cert. denied, 504 U.S. 956, 112 S. Ct. 2304, 119 L. Ed. 2d 226 (1992), in which the plaintiff sued the defendant for failure to provide adequate warning. The jury returned a $1,000,000 verdict for the plaintiff, but the district court granted the defendant's motion for judgment notwithstanding the verdict and set aside the jury verdict. Thomas is not determinative in the present case, however, because Thomas is distinguishable from this case in several important respects. First, Thomas involved a prescription drug; therefore, in order to prove that the inadequate warning caused the plaintiff's injuries, the Plaintiff had to prove that an adequate warning would have convinced her physician not to prescribe the drug. Second, the defendant in Thomas notified physicians of the possible risk associated with its product. Finally, the plaintiff in Thomas argued that in order to adequately warn physicians, the defendant needed to contraindicate its product.5

McNeil contends that the adverse reports do not establish that therapeutic doses of acetaminophen combined with moderate amounts of alcohol can cause liver damage. Accordingly, McNeil argues that it had no knowledge of such a risk, which is required before it can be held liable for negligent failure to warn. See Owens-Corning Fiberglas Corp. v. Watson, 243 Va. 128, 413 S.E.2d 630, 634 (1992) ("[A] manufacturer has a duty to warn only if it knows or has reason to know that its product is dangerous.").

McNeil states that, according to the FDA, case reports do not establish that a particular drug is reasonably associated with a particular injury. McNeil also cites Weinberger v. Hynson, Westcott & Dunning, Inc., 412 U.S. 609, 93 S. Ct. 2469, 37 L. Ed. 2d 207 (1973), and E.R. Squibb & Sons, Inc. v. Bowen, 870 F.2d 678 (D.C.Cir.1989), in support of this position. However, the FDA and these cases merely state that case reports cannot serve as the basis for approving a drug; they do not say that these case reports do not give rise to a duty to warn.

Because we review a denial of a motion for judgment as a matter of law de novo, we must determine whether substantial evidence exists upon which a reasonable jury could find for Benedi. Benesh v. Amphenol Corp. (In re Wildewood Litigation), 52 F.3d 499, 502 (4th Cir.1995). We find that sufficient evidence existed to support the jury's finding that McNeil negligently failed to warn consumers of the risks associated with combining therapeutic doses of acetaminophen with alcohol.

Benedi introduced expert testimony that the Tylenol label should have contained a warning against taking the medication while consuming alcohol. Benedi also introduced the DERs as evidence that McNeil had notice of the dangers associated with combining therapeutic doses of acetaminophen with alcohol, yet failed to warn of these dangers on Tylenol's label. Finally, Benedi presented evidence that the FDA's 1988 "tentative final monograph"6 on Tylenol only considered data collected up to 1980 because McNeil did not provide approximately forty case reports that it collected from 1980-1988 to the FDA's monograph staff until after publication of the 1988 monograph. The evidence supports the jury's decision that McNeil negligently failed to warn consumers.

Next, McNeil contends that the district court erroneously admitted into evidence a subsequent remedial measure and that the district court therefore should have granted its motion for a new trial. Rule 407 of the Federal Rules of Evidence provides that subsequent remedial measures are "not admissible to prove negligence or culpable conduct in connection with the event. This rule does not require the exclusion of evidence of subsequent measures when offered for another purpose, such as proving ownership, control, or feasibility of precautionary measures, if controverted, or impeachment." Fed.R.Evid. 407.

At trial Benedi called Dr. Anthony Temple, McNeil's medical director and corporate representative, as an adverse witness. Benedi's counsel questioned Dr. Temple about discussions between McNeil and the FDA in December 1992 concerning an alcohol warning on a new Tylenol product, Extended Relief Tylenol, which was very similar to the product at issue in this case, Extra-Strength Tylenol. In December 1992, an FDA official proposed to McNeil that the new product contain the following warning: "If you consume more than two alcohol-containing drinks on a daily basis, consult your physician before using this product." These events all occurred prior to plaintiff's injury and evidence about them was properly put before the jury without objection. However, Benedi's counsel then went further and asked Dr. Temple if McNeil placed the warning on the new product, and he responded that McNeil had done so, although it used a slightly different warning: "[I]f you regularly consume three or more alcohol-containing drinks every day you should consult your physician before using this or any other pain reliever." Benedi's counsel then questioned Dr. Temple about the similarities between the new product and the product involved in this case, and at that point McNeil objected.7 At side-bar, the judge instructed Benedi's counsel that because the warning was placed on the new product after Benedi's injury, he would not allow any post-February 18, 1993 activity of the FDA to be injected into the case. McNeil did not ask the court to strike the statement or to give a cautionary instruction.

Before closing arguments, the court instructed the parties that it would allow plaintiff to refer to the pre-injury discussions between McNeil and the FDA, but that it would not allow plaintiff to refer to the post-injury label change. In closing argument, plaintiff's counsel only referred to the 1992 discussions between the FDA and McNeil concerning the FDA's proposed warning on the new acetaminophen product.

In denying McNeil's motion for a new trial, the district court determined that the evidence was not offered solely on the issue of negligence or breach of warranty, but rather, the testimony came about as a result of defendant's expert's position on the prior informal proposal to add an alcohol warning to the product's label. The district court further noted that the sole reference to the warning that McNeil ultimately placed on its new product in late 1993 was the subject of only a passing reference during the trial and was not argued to the jury. In short the district court found that the error, if any, was harmless.

In suggesting that the plaintiff's reference to the warning McNeil placed on its new product after Benedi's injury justifies a new trial due to prejudice, McNeil relies upon Werner v. Upjohn Co., 628 F.2d 848 (4th Cir.1980), cert. denied, 449 U.S. 1080, 101 S. Ct. 862, 66 L. Ed. 2d 804 (1981). The plaintiff in Werner suffered injuries in 1974 from a drug manufactured by Upjohn, and the primary issue at trial was Upjohn's allegedly negligent failure to provide an adequate warning. The district court allowed the plaintiff to introduce evidence of a warning which was placed on the product in 1975 and which expanded on the previous warning in several significant respects:8 The district court allowed the evidence under Rule 407 and instructed the jury that they could only consider the evidence on the issue of feasibility. However, plaintiff referred to this evidence several times during the trial and during closing argument without reference to feasibility. This court found that plaintiff's repeated references to the subsequent warning were clearly intended to prove negligence in not using the 1975 warning in 1974. Id. at 853. We held that Upjohn's failure to object to plaintiff's impermissible use of the 1975 warning did not preclude Upjohn's claim because an objection is not required when it would emphasize the impermissible point. Id. at 854. Finally, we recognized that Rule 407 allows evidence of subsequent remedial measures to prove feasibility; however, feasibility was not an issue in Werner, and the district court, therefore, committed reversible error in allowing evidence of the 1975 warning. Id. at 854-56.

McNeil contends that Werner mandates a new trial in this case. However, Werner is distinguishable from the present case in several respects. First, in this case the majority of plaintiff's questions to Dr. Temple focused on December 1992 discussions between the FDA and McNeil about a warning on its new acetaminophen product. These discussions predated Benedi's injury. Second, plaintiff's closing argument mentioned only the pre-injury discussions between McNeil and the FDA. Finally, once McNeil objected to plaintiff's reference to the content of the warning that was eventually placed on the new acetaminophen product in late 1993, plaintiff never mentioned the subsequent warning again. In contrast, Werner involved repeated, flagrant attempts to rely upon the subsequent warning to prove negligence.

We find that plaintiff's single reference to the subsequent warning in a trial in which the transcript exceeds 800 pages does not justify granting a new trial. The district judge confined all further testimony to pre-injury discussions between McNeil and the FDA on warnings, and plaintiff adhered to this instruction. The sole reference to the 1993 warning placed on the new product was harmless error in light of the properly admitted and far more damaging testimony that prior to Benedi's injury the FDA had suggested to McNeil that it place an alcohol warning on its new acetaminophen product. This latter testimony supports a finding that McNeil had prior knowledge that acetaminophen was dangerous when combined with alcohol and that its acetaminophen products needed an alcohol warning.

Finally, McNeil contends that the district court erred in submitting the issue of punitive damages to the jury. In order for a plaintiff to recover punitive damages under Virginia law, the plaintiff must prove either malice or conduct "so willful or wanton as to evince a conscious disregard of the rights of others." Booth v. Robertson, 236 Va. 269, 374 S.E.2d 1, 3 (1988).

We find that Benedi introduced sufficient evidence from which a reasonable jury could find that McNeil acted with reckless indifference to the health of its consumers. For example, Benedi introduced evidence that McNeil failed to give the FDA the 1980-1988 case reports until after the FDA had issued its 1988 final monograph, which concluded that an alcohol warning was not necessary. Benedi also introduced evidence that McNeil instructed its sales representatives to refrain from discussions with physicians about a scholarly article, which outlined the dangers associated with combining acetaminophen with alcohol. Finally, Benedi introduced evidence that McNeil sent a letter to pharmacists and hospitals stating that there was no link between liver damage and casual alcohol consumption combined with acetaminophen.

McNeil dedicates numerous pages of its brief to refuting Benedi's evidence of reckless conduct by McNeil and explaining the correct interpretation of its actions. Because reasonable people could differ in their evaluation of McNeil's conduct, it was the jury's job to weigh the testimony of both sides and determine the motive underlying McNeil's conduct. We will not overturn the jury's finding if substantial evidence exists to support it. We find that the district court did not err in submitting the issue of punitive damages to the jury because sufficient evidence existed on this issue that would allow a reasonable jury to find wanton and willful misconduct by McNeil.

For the foregoing reasons, the judgment of the district court is

AFFIRMED.