The dispute in the underlying action centers upon an advertising campaign launched by defendant-appellant Bristol-Myers Squibb Company ("Bristol-Myers"), proclaiming that Aspirin-Free Excedrin ("AF Excedrin"), an over-the-counter ("OTC") analgesic it manufacturers, relieves pain better than Extra-Strength Tylenol ("ES Tylenol"), an OTC analgesic manufactured by plaintiff-appellee McNeil-P.C.C., Inc. ("McNeil"). Believing that Bristol-Myers' comparative claim of superiority was false, McNeil commenced an action in the United States District Court for the Southern District of New York pursuant to Section 43(a) of the Lanham Act, 15 U.S.C. Sec. 1125(a) (1988), seeking preliminary and permanent injunctive relief enjoining Bristol-Myers from making its superiority claim.

The district court (John M. Cannella, Judge ), consolidated the hearing on McNeil's motion for a preliminary injunction with the trial on the merits. See Fed.R.Civ.P. 65(a)(2). Following a three-day bench trial, the district court ruled that McNeil had met its burden of proving the falsity of Bristol-Myers' advertising claim by demonstrating that Bristol-Myers' own test data established that AF Excedrin did not outperform ES Tylenol to a statistically significant degree. See McNeil-P.C.C., Inc. v. Bristol-Myers Squibb Co., 755 F. Supp. 1206, 1219 (S.D.N.Y.1990). Accordingly, the district court permanently enjoined Bristol-Myers from making its superiority claim. Bristol-Myers appeals from this judgment, contending, among other things, that the district court erred by improperly allocating and applying the burden of proof. Additionally, Bristol-Myers challenges several of the district court's factual findings.

For the reasons set forth below, the judgment of the district court is affirmed.

Defendant-appellant Bristol-Myers and plaintiff-appellee McNeil are direct competitors in the market for OTC analgesics. Bristol-Myers sells Excedrin pain relievers. McNeil, the market leader in OTC analgesics, sells Tylenol products, including ES Tylenol. A two-tablet dose of ES Tylenol contains 1000 milligrams ("mg.") of acetaminophen. In the Spring of 1990, Bristol-Myers began marketing AF Excedrin, a two-tablet dose of which contains 1000 mg. of acetaminophen and 130 mg. of caffeine. Thus, the only difference between AF Excedrin and ES Tylenol is that AF Excedrin contains caffeine.

To market its new pain reliever, Bristol-Myers began a comprehensive advertising campaign. Initially, Bristol-Myers sent promotional literature to drug retailers and other trade professionals that touted AF Excedrin's superiority over ES Tylenol. The advertisement proclaimed that "in doctor supervised clinical studies ... [AF] Excedrin was shown to provide greater headache relief" than ES Tylenol. According to the advertisement, AF Excedrin "works better" than ES Tylenol because AF Excedrin contains caffeine, which "acts as an analgesic adjuvant--boosting the effect of the acetaminophen to provide greater headache relief." Bristol-Myers subsequently directed its promotional efforts at the general public. On August 30, 1990, Bristol-Myers commenced a television advertising campaign. Its television commercial featured a headache sufferer who explained that she used to take ES Tylenol, but then tried AF Excedrin and discovered that AF Excedrin "works better." Id. at 1209. Unlike the trade advertising, this commercial did not refer to clinical studies or to caffeine's adjuvancy as the basis of the superiority claim.

Because McNeil believed that Bristol-Myers' "works better" claim was false, and therefore violated section 43(a) of the Lanham Act, 15 U.S.C. Sec. 1125(a), McNeil commenced the underlying action seeking preliminary and permanent injunctive relief enjoining Bristol-Myers from making that claim. The district court consolidated the preliminary injunction hearing with the trial on the merits. See Fed.R.Civ.P. 65(a)(2). Since the only difference between AF Excedrin and ES Tylenol is that AF Excedrin contains caffeine, the district court noted that, in order to prevail on its Lanham Act claim, McNeil had to demonstrate by a preponderance of the evidence that "the presence of caffeine in AF Excedrin fails to make AF Excedrin superior to ES Tylenol in relieving headache pain." Id. at 1211.

To meet its burden of proof, McNeil relied on expert statistical analyses of two studies performed by Bristol-Myers in 1988 and 1989 that directly compared the efficacy of AF Excedrin and ES Tylenol in relieving headache pain ("AF Excedrin studies"). According to McNeil, a proper analysis of the AF Excedrin studies' data demonstrated that AF Excedrin did not outperform ES Tylenol to a statistically significant degree and therefore Bristol-Myers' superiority claim was false. Bristol-Myers countered that the data from the AF Excedrin studies indeed showed that AF Excedrin relieved headache pain significantly better than ES Tylenol.

Before evaluating the parties' contentions concerning the results of the AF Excedrin studies, the district court considered the relevance of other studies offered by Bristol-Myers to rebut McNeil's assertion that Bristol-Myers' superiority claim was false, namely: (1) Bristol-Myers' studies comparing original formula Excedrin with ES Tylenol ("Excedrin studies"), (2) a study conducted by Whitehall Laboratories comparing Anacin, which contains aspirin and caffeine, with aspirin ("Whitehall study"), and (3) a "meta-analysis study" that analyzed the results of thirty clinical studies involving the adjuvant effect of caffeine on a variety of analgesics. Bristol-Myers contended that these studies established the validity of its superiority claim. McNeil argued that the studies were not pertinent to determining whether AF Excedrin outperformed ES Tylenol.

The district court agreed with McNeil's view and found that these other studies were not germane to Bristol-Myers' "works better" claim in the instant case. The Excedrin studies compared Excedrin, containing 500 mg. of aspirin, 500 mg. of acetaminophen, and 130 mg. of caffeine, with ES Tylenol. Based on expert testimony and documentary evidence, the district court concluded that studies testing the adjuvant effect of caffeine on a combination of analgesics, such as the aspirin and acetaminophen contained in Excedrin, were not relevant to determining caffeine's interaction with and effect on acetaminophen alone. Id. at 1213. Accordingly, the court found that the results of the Excedrin studies were not relevant to whether caffeine acted as an adjuvant in AF Excedrin, which contains only acetaminophen. For similar reasons, the district court rejected the results of the Whitehall study.

The district court also determined that the results of the meta-analysis study were not relevant because it was based on doses of caffeine and analgesic that were beyond OTC levels. Moreover, evidence at trial showed that only seven out of the thirty studies incorporated into the meta-analysis concerned acetaminophen. Judge Cannella therefore accorded the meta-analysis little or no weight.

Because the district court found these other studies not pertinent to the dispute in the instant action, the key factual inquiry at trial became whether the AF Excedrin studies demonstrated AF Excedrin's superiority over ES Tylenol. The AF Excedrin studies were "crossover" studies. As explained by the district court, "[i]n a crossover study the patients take one study drug in 'period one' and another study drug in 'period two.' " Id. at 1214. The principal advantage of such studies is that each patient receives and evaluates the efficacy of both study drugs, thereby serving " 'as his [or her] own control.' " Id. Inherent in crossover studies, however, is the danger of a "carryover" effect, whereby the physical or psychological effects of a drug in period one carry into and distort the test results in period two. Id. at 1215. If carryover is present in a crossover study, the experts substantially agreed that only the period one data is reliable. Id. at 1216.

At trial, the parties agreed that an analysis of both the period one and period two data from the AF Excedrin studies showed that AF Excedrin relieved pain better than ES Tylenol to a statistically significant degree. However, McNeil argued that the AF Excedrin studies were distorted by psychological carryover, thereby invalidating the period two results. According to McNeil, a proper analysis of the period one data--the only reliable data--revealed no significant difference between the two analgesics. Consequently, McNeil contended that AF Excedrin and ES Tylenol were equivalent pain relievers and that Bristol-Myers' claim of superiority was therefore false.

The parties vigorously contested whether the AF Excedrin studies were tainted by carryover and presented the court with conflicting expert testimony regarding the proper statistical test for ascertaining the presence of carryover. McNeil's experts, Professor Nan Laird, an expert in biostatistics, and Professor Michael G. Kenward, an expert in crossover studies, both of whom the court found to be credible witnesses, advocated a "sums and differences" test. They testified that this test was more reliable than any other carryover test and that it conclusively demonstrated the existence of carryover. The validity of the "sums and differences" test was also recognized in the scientific research paper upon which the AF Excedrin studies were based. Id. at 1217. McNeil next presented a "two-sequence" test, which, though less reliable than the sums and differences method, also pointed to the presence of carryover. Professor Laird further demonstrated the presence of carryover with graphic illustrations of the period one and period two data. To rebut McNeil's proof of carryover, Bristol-Myers offered its own array of statistical tests, including a "four headache analysis" and a "goodness-of-fit" test. According to Professor Dennis Gillings, Bristol-Myers' expert in biostatistics, the "four headache analysis" was the most reliable carryover test and showed no evidence of carryover. However, based on Professor Kenward's criticisms of the methodology of the four headache analysis, the court found the results of that test to be less persuasive than those of the sums and differences analysis. Id. at 1218. The court also rejected the results of the goodness-of-fit test because it did not test directly for carryover. Consequently, the district court found that McNeil had proved the presence of carryover in the AF Excedrin studies.

The truth or falsity of Bristol-Myers' "works better" claim therefore turned on an analysis of the period one data. Professors Laird and Kenward analyzed the period one data and found no significant statistical difference between AF Excedrin and ES Tylenol in relieving headache pain. This statistical analysis of the voluminous period one data led these experts to conclude that AF Excedrin provided the same, not superior, pain relief. Bristol-Myers offered its own statistical analyses of the period one data to demonstrate that AF Excedrin significantly outperformed ES Tylenol. The court considered Bristol-Myers' analyses and rejected them.

Accordingly, the district court held that the results of McNeil's statistical analyses satisfied its burden of proving the falsity of Bristol-Myers' superiority claim. The court therefore granted the requested injunctive relief and permanently enjoined Bristol-Myers from advertising that AF Excedrin works better than ES Tylenol. This appeal followed.

I. Bristol-Myers' Challenges to the District Court's Application of the Burden of Proof.

On appeal, Bristol-Myers contends that the district court misapplied the burden of proof. Specifically, Bristol-Myers argues that the district court improperly permitted McNeil to prevail in its Lanham Act action merely by criticizing Bristol-Myers' studies rather than by adducing its own studies to demonstrate the falsity of Bristol-Myers' superiority claim. Bristol-Myers also complains that by rejecting the other scientific studies that Bristol-Myers submitted to rebut McNeil's charge of falsity, and by rejecting the period two data of the AF Excedrin studies, the court impermissibly shifted the burden of proof to Bristol-Myers.

Section 43(a) of the Lanham Act provides in relevant part:

Any person who, on or in connection with any goods or services, ... uses in commerce ... any false designation of origin, false or misleading description of fact, or false or misleading representation of fact, which--

* * * * * *

(2) in commercial advertising or promotion, misrepresents the nature, characteristics, qualities, or geographic origin of his or her or another person's goods, services or commercial activities,

shall be liable in a civil action by any person who believes that he or she is or is likely to be damaged by such act.

15 U.S.C. Sec. 1125(a).1 To obtain permanent injunctive relief against a false or misleading advertising claim pursuant to section 43(a), a plaintiff must demonstrate by a preponderance of the evidence that an advertisement is either literally false or that the advertisement, though literally true, is likely to mislead and confuse consumers. See Johnson & Johnson v. GAC Int'l, Inc., 862 F.2d 975, 977 (2d Cir.1988); Coca-Cola Co. v. Tropicana Products, Inc., 690 F.2d 312, 317 (2d Cir.1982). Where the advertising claim is shown to be literally false, the court may enjoin the use of the claim "without reference to the advertisement's impact on the buying public." Coca-Cola, 690 F.2d at 317; see also PPX Enterprises, Inc. v. Audiofidelity Enterprises, Inc., 818 F.2d 266, 272 (2d Cir.1987) (citation omitted); American Home Products Corp. v. Johnson & Johnson, 577 F.2d 160, 165 (2d Cir.1978) (citation omitted). Additionally, a plaintiff must show that it will suffer irreparable harm absent the injunction. See, e.g., Tambrands, Inc. v. Warner-Lambert Co., 673 F. Supp. 1190, 1195 (S.D.N.Y.1987). In the instant case, Bristol-Myers does not dispute that if McNeil sustains its burden of proving that Bristol-Myers' "works better" claim is literally false, McNeil will have established the requisite harm. See McNeilab, Inc. v. American Home Products Corp., 848 F.2d 34, 38 (2d Cir.1988) (where the challenged advertisement directly, but falsely, proclaims the superiority of defendant's product over plaintiff's, irreparable harm may be presumed). Nevertheless, Bristol-Myers contends that McNeil has failed to meet its burden of proving the falsity of Bristol-Myers' "works better" claim.

To prove that an advertising claim is literally false, a plaintiff must do more than show that the tests supporting the challenged claim are unpersuasive. Procter & Gamble Co. v. Chesebrough-Pond's Inc., 747 F.2d 114, 119 (2d Cir.1984). Rather, the plaintiff must demonstrate that such tests are "not sufficiently reliable to permit one to conclude with reasonable certainty that they established" the claim made. Id.; see also Alpo Petfoods, Inc. v. Ralston Purina Co., 720 F. Supp. 194, 213 (D.D.C.1989) ("[r]epresentations found to be unsupported by accepted authority or research or which are contradicted by prevailing authority or research, may be deemed false on their face and actionable under Section 43(a) of the Lanham Act."), aff'd in part, rev'd in part on other grounds, 913 F.2d 958 (D.C.Cir.1990).

Bristol-Myers argues that the district court erred by permitting McNeil to satisfy its burden of proof, not by offering its own studies, but by merely showing that Bristol-Myers' tests were unpersuasive. This contention is unavailing. We do not believe that McNeil had to conduct its own studies to establish the falsity of Bristol-Myers' "works-better" claim. Bristol-Myers' claim was part of a comprehensive advertising campaign directly comparing the efficacy of AF Excedrin and ES Tylenol. As Bristol-Myers' initial trade advertising made plain, the "works better" claim was bottomed on the results of the AF Excedrin studies. McNeil could therefore meet its burden of proof by demonstrating that these studies did not establish that AF Excedrin provided superior pain relief. See Procter & Gamble, 747 F.2d at 119; Alpo Petfoods, Inc., 720 F. Supp. at 213; cf. Tambrands, 673 F. Supp. at 1194 (defendants' advertising claim deemed false where defendant's own test results refuted the claim). Moreover, we note that even if the tests were not directly referred to in connection with Bristol-Myers' superiority claim, McNeil still could have relied on and analyzed data generated by Bristol-Myers as scientific proof that the challenged advertisement was false.

In the present case, McNeil has sustained its burden of proving Bristol-Myers' superiority claim to be false. Despite Bristol-Myers' arguments to the contrary, McNeil did more than merely critique and expose the methodological weaknesses of the AF Excedrin studies. McNeil affirmatively demonstrated to the satisfaction of the district court that a proper analysis of the period one data--which the court found to be the only relevant and reliable scientific data--conclusively established that AF Excedrin did not relieve pain better than ES Tylenol to a statistically significant degree. Such proof was sufficient to establish that Bristol-Myers' superiority claim was false. See Vidal Sassoon, Inc. v. Bristol-Myers Co., 661 F.2d 272, 278 (2d Cir.1981); American Home Products Corp. v. Johnson & Johnson, 654 F. Supp. 568, 576 (S.D.N.Y.1987); American Home Products Corp. v. Johnson & Johnson, 436 F. Supp. 785, 802-03 (S.D.N.Y.1977), aff'd, 577 F.2d 160 (2d Cir.1978).

Similarly unpersuasive is Bristol-Myers' contention that the district court impermissibly shifted to it the burden of proving the truth of its "works better" claim. By rejecting the other scientific studies Bristol-Myers offered to rebut McNeil's charge of falsity, and by rejecting the period two data of the AF Excedrin studies, the district court merely determined what evidence it would consider on the issue of falsity. These determinations in no way shifted the burden of proof. Accordingly, we hold that Judge Cannella clearly articulated the appropriate burden of proof and placed that burden squarely on McNeil.

II. Bristol-Myers' Challenges to the District Court's Factual Determinations.

Bristol-Myers next challenges various findings made by the district court. According to Bristol-Myers, the district court erred by finding that the other scientific studies it offered to rebut McNeil's claims of falsity were irrelevant or entitled to little or no evidentiary weight. Additionally, Bristol-Myers contends that the court erred in finding carryover in the AF Excedrin studies and in finding that the period one data showed no statistically significant difference between the pain relief provided by AF Excedrin and ES Tylenol. We find these contentions unpersuasive.

Initially, we note that the district court's determinations regarding the weight to accord to Bristol-Myers' other studies, such as the Excedrin studies, are factual determinations that will not be disturbed unless "clearly erroneous." See Fed.R.Civ.P. 52(a); Procter & Gamble, 747 F.2d at 119. A finding is "clearly erroneous" when, considering the entire record, the reviewing court is "left with a definite and firm conviction that a mistake has been committed." Anderson v. City of Bessemer City, 470 U.S. 564, 573, 105 S. Ct. 1504, 1511, 84 L. Ed. 2d 518 (1985) (quoting United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S. Ct. 525, 541, 92 L. Ed. 746 (1948)); see also Avis Rent A Car System, Inc. v. Hertz Corp., 782 F.2d 381, 384 (2d Cir.1986). An appellate court may not, however, reverse a trial court's factual findings simply because the appellate court would have weighed the evidence differently in the first instance. See Anderson, 470 U.S. at 574, 105 S.Ct. at 1511; Inwood Laboratories, Inc. v. Ives Laboratories, Inc., 456 U.S. 844, 856-57, 102 S. Ct. 2182, 2189-90, 72 L. Ed. 2d 606 (1982); Sheehan v. Purolator Inc., 839 F.2d 99, 105 (2d Cir.), cert. denied, 488 U.S. 891, 109 S. Ct. 226, 102 L. Ed. 2d 216 (1988). This deferential standard of review "rests upon the unique opportunity afforded the trial court judge to evaluate the credibility of witnesses and to weigh the evidence." Inwood Laboratories, 456 U.S. at 855, 102 S.Ct. at 2189; see also Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 123, 89 S. Ct. 1562, 1576, 23 L. Ed. 2d 129 (1969).

Applying these well-established principles to the instant case, we conclude that the district court's determinations regarding the evidentiary weight and value of the Excedrin, Whitehall, and meta-analysis studies were not clearly erroneous. The district court heard and evaluated documentary evidence and expert testimony regarding the relevance of these studies to the issue of whether caffeine made AF Excedrin relieve pain better than ES Tylenol. Ultimately, the court found that only the AF Excedrin studies were probative of that issue. The court's decision to accord the other studies little or no weight is amply supported in the record and is therefore not clearly erroneous. See American Home Products Corp., 577 F.2d at 170; see also Anderson, 470 U.S. at 574, 105 S.Ct. at 1511; Sheehan, 839 F.2d at 105.

Bristol-Myers also contends that the district court erred in finding carryover in the AF Excedrin studies. The court was presented with conflicting expert testimony concerning the proper test for carryover. Professors Laird and Kenward, whom the court found to be credible witnesses, advocated a "sums and differences" test that conclusively demonstrated the presence of carryover. The court chose to accord this test more weight than the competing tests, such as the "four headache analysis" offered by Bristol-Myers' expert, Professor Gillings. Additionally, both Professor Laird's graphic illustration and her "two-sequence" test also pointed to the presence of carryover. Although the carryover issue presented a close factual question at trial, the district court's finding of carryover is supported by the record and we do not believe that the court erred in its evaluation and weighing of the evidence. See American Home Products Corp., 577 F.2d at 170; Anderson, 470 U.S. at 574, 105 S.Ct. at 1511. Similarly, we believe that, based on the expert testimony presented, the district court did not err in finding that the period one data showed no statistically significant difference in efficacy between AF Excedrin and ES Tylenol, and, therefore, that AF Excedrin did not "work better" than ES Tylenol.

We have considered Bristol-Myers' remaining contentions and find them to be without merit. Based on the foregoing, we affirm the judgment of the district court.