IRVING R. KAUFMAN, Circuit Judge:

The ongoing competition between McNeilab, Inc. (McNeil), a wholly owned subsidiary of Johnson & Johnson, Inc., and American Home Products Corporation (AHP), rival pain reliever manufacturers, has brought anything but relief to the federal courts. Instead, repeated and protracted litigation has created a substantial headache. The competitive battlefield has shifted from the shelves of supermarkets and drugstores to the courtroom. Once again, we are called upon to restrain deceptive advertising practices and to protect the public from inaccurate safety claims.

Plaintiff-appellee McNeil manufactures Tylenol, the leading aspirin-free over-the-counter (OTC) pain reliever. The active ingredient in Tylenol is acetaminophen. Defendant-appellant American Home Products manufactures Advil, a competing product. The active ingredient in Advil is ibuprofen. Tylenol and Advil compete directly, along with aspirin, in the lucrative OTC internal analgesic market. To distinguish their products in consumers’ minds, each manufacturer stresses the diverse qualities of its drug’s ingredients, particularly with respect to safety and effectiveness.

Ibuprofen has been available in the United States as a prescription drug since 1974. The Food and Drug Administration approved it for over-the-counter distribution in 1984. Like aspirin, ibuprofen is a non-steroidal anti-inflammatory drug that inhibits the body’s production of prostaglan-dins. Prostaglandins sensitize nerve endings to pain and inflamed joints; thus these drugs prevent pain and swelling. Because prostaglandins also protect the stomach lining, however, impeding their production may lead to stomach irritation or ulcers. Ibuprofen is less likely to cause these side effects than aspirin. Acetaminophen, on the other hand, does not affect prostaglan-din production; it reduces suffering by raising the pain threshold. Consequently, acetaminophen is relatively free of adverse gastrointestinal side effects.

AHP introduced Advil in 1984. McNeil, perceiving a threat to its market position, responded with heavy promotion of Tylenol. It distributed innumerable free sam- *36 pies to physicians and engaged in widespread print and television advertising. The competition swiftly turned bitter, however, as AHP filed suit in the United States District Court for the Southern District of New York, charging McNeil with false advertising. The principal target of AHP’s claim was a safety profile McNeil provided physicians visually linking ibuprofen to aspirin in a comparative display. The effect, AHP asserted, was to suggest that ibuprofen shared aspirin’s high propensity to irritate the stomach. McNeil counterclaimed, raising allegations that AHP’s multi-media advertising campaign concealed its product’s side effects.

In April 1987, after a lengthy trial, Judge Conner held that both parties had violated section 43(a) of the Lanham Act, 15 U.S.C. § 1125(a) (1982). American Home Products Corp. v. Johnson & Johnson, 654 F.Supp. 568 (S.D.N.Y.1987) (Advil 1). Specifically, he found McNeil had falsely equated ibuprofen’s risks with aspirin’s and had exaggerated acetaminophen’s safety superiority over ibuprofen. AHP, he determined, misled consumers by claiming ibuprofen was less susceptible than acetaminophen to adverse reactions with other drugs. Advil I, 654 F.Supp. at 591. Judge Conner also indicated that while neither Advil nor Tylenol is likely to cause minor “subjective” stomach pain, Advil is more likely to cause serious “objective” stomach injuries such as ulcers, hemorrhages or concealed bleeding. Advil I, 654 F.Supp. at 584. The court then enjoined a variety of advertising by both parties.

Within days, AHP launched a new televised campaign proclaiming, “if you worry about the discomfort of stomach upset as I do, I find that nothing is better than Advil. Not even Tylenol.” (emphasis added). AHP also introduced a commercial stating that “like Tylenol, Advil doesn’t upset my stomach.” (emphasis added). Thus, AHP linked Advil directly to Tylenol. In response, McNeil brought this Lanham Act claim, seeking damages and injunctive relief.

McNeil charged the slogans would mislead and confuse consumers by falsely representing that Advil and Tylenol cause stomach upset with the same frequency. McNeil also alleged the commercials contradicted Judge Conner’s determination in Advil I that Advil is more likely than Tylenol to cause severe side effects. AHP responded that the comparisons were neither false nor misleading because ibuprofen and acetaminophen produce similar minor side effects. Consumers, AHP contended, would not interpret “stomach upset” to include serious gastrointestinal injury, but, rather, would distinguish between subjective and objective stomach problems. Accordingly, the decision in this case depended upon consumer understanding of the phrase “stomach upset.”

In preparation for trial, McNeil designed two surveys to document its contentions of consumer confusion. McNeil intended to show potential consumers the challenged commercials and then ask them a series of questions gauging their reactions. At Judge Conner’s suggestion, the studies were submitted for inter partes challenge and court approval to ensure their eviden-tiary value. In a letter dated June 24, 1987, Judge Conner raised several objections to the surveys, rewriting at least one question. He did not order McNeil to accept his alterations; he simply indicated the revisions would enhance the value of the evidence. McNeil adopted Judge Conner’s proposals.

On July 27, the district court judge convened an evidentiary hearing to consider McNeil’s preliminary injunction motion. Immediately before the hearing, however, AHP voluntarily discontinued those commercials stating “nothing is better than Advil. Not even Tylenol.” The slogan, AHP conceded, “unexpectedly” implied that Advil is safer than Tylenol. Consequently, the hearing addressed only those advertisements containing AHP’s assertion that “like Tylenol, Advil doesn’t upset my stomach.”

The hearing, however, never reached a decision on the merits. After all evidence was introduced, the court renewed its efforts to induce a settlement. Judge Conner proposed that AHP clarify its remaining commercials to specify that the similarity between Advil and Tylenol extends only to “minor” or “occasional” stomach upset. AHP agreed, but McNeil balked at the *37 proposed settlement, fearing that the modified slogans would still mislead consumers. Despite McNeil’s reservations, AHP agreed to alter its commercials to state, “like Tylenol, Advil doesn’t give me minor [occasional] stomach upset.” (emphasis added). The court then denied a preliminary injunction, reasoning that McNeil’s evidence of consumer confusion was directed to the former language and therefore was not probative of the new wording. Because the decision turned on the lack of relevant evidence, it did not determine whether the modified commercials would confuse consumers.

In August, AHP began airing advertisements incorporating the agreed upon changes. McNeil immediately amended its complaint to challenge the adjusted script. The original commercial — “like Tylenol, Advil doesn’t upset my stomach” — baldly asserted equivalence between Advil and Tylenol. The new versions were subtler, McNeil argued, but no less misleading. By claiming parity between Advil and Tylenol for minor stomach problems, AHP necessarily implied similarity for serious problems. Thus, the modified phrasing did not improve upon the original slogan.

Judge Conner convened a second eviden-tiary hearing in October to consider McNeil’s challenge to the adjusted language. McNeil supplemented the record of the first hearing with three new surveys indicating consumer confusion from the updated advertisements. It also adduced testimony from a prominent gastroenterologist that his patients used the phrase “stomach upset” in a generic fashion to describe all types of stomach ailments, whether subjective or objective. Moreover, McNeil introduced testimony that the television networks were initially reluctant to air the modified commercials due, at least in part, to a propensity to confuse viewers. As a result, Judge Conner preliminarily enjoined AHP’s claim that Advil is “like Tylenol” with regard to “minor” or “occasional” stomach upset, McNeilab, Inc. v. American Home Products Corp., 675 F.Supp. 819 (S.D.N.Y.1987) (Advil II), and this appeal ensued.

A preliminary injunction may issue upon a showing of: (a) a risk of irreparable harm and (b) either (1) a likelihood of success on the merits or (2) the existence of sufficiently serious questions going to the merits to make them fair ground for litigation and a balance of hardships decidedly favoring the party seeking relief. Charles of the Ritz Group Ltd. v. Quality King Distribs., Inc., 832 F.2d 1317, 1320 (2d Cir.1987); Jackson Dairy, Inc. v. H.P. Hood & Sons, Inc., 596 F.2d 70, 72 (2d Cir.1979) (per curiam). A preliminary injunction is reviewed for abuse of discretion, which occurs when the district court relies upon “clearly erroneous findings of fact or on an error of law in issuing the injunction.” Hanson Trust PLC v. SCM Corp., 774 F.2d 47, 54 (2d Cir.1985); Coca-Cola Co. v. Tropicana Prods., Inc., 690 F.2d 312, 315-16 (2d Cir.1982). See Fed.R.Civ.P. 52(a).

AHP directs its appeal to the district court’s factual findings. Specifically, it charges that the surveys were flawed and thus could not support a finding of likelihood of success on the merits. AHP claims: (1) the studies did not sufficiently focus on the advertisements’ comparative claims; (2) they contained leading questions regarding the meaning of “stomach damage”; (3) the court relied on ambiguous testimony interpreting the results to find consumer confusion; and (4) a survey measuring confusion among ulcer sufferers utilized flawed sampling techniques by including only those who had participated in previous studies rather than drawing upon the general population of ulcer sufferers.

McNeil responds that Judge Conner correctly concluded a “not insubstantial fraction” of viewers would perceive the new language as claiming overall equivalence between Advil and Tylenol and thus would be misled regarding their relative safety. It also notes that this court has traditionally undertaken close scrutiny of drug safety claims. See Syntex Laboratories, Inc. v. Norwich Pharmacal Co., 437 F.2d 566, 569 (2d Cir.1971) (applying a “stricter standard” in Lanham Act cases where confusion “could result in physical harm to the consuming public ... through adverse reaction to inadvertently prescribed or dispensed drugs”). Finally, McNeil urges deference to Judge Conner’s fact finding, ob *38 serving that he participated in drafting the initial surveys and presided over hearings consuming five days directed primarily at the efficacy of the survey evidence.

AHP’s challenges raise serious questions related to the usefulness of McNeil’s research. If the surveys were so flawed that they lacked relevance, they would have been inadmissible under Rules 401 and 402 of the Federal Rules of Evidence and thus could not have supported Judge Conner’s decision. See, e.g., Universal City Studios, Inc. v. Nintendo Co., 746 F.2d 112, 118 (2d Cir.1984). If, however, the research data, alone or in conjunction with other evidence, made “the existence of any fact that is of consequence to the determination of the action more probable or less probable,” Fed.R.Evid. 401, then the surveys were admissible. Fed.R.Evid. 402.

Judge Conner recognized the surveys’ deficiencies but determined that the evidence was relevant despite its flaws. The results, he concluded, were “supported by the other evidence adduced by McNeil and by common sense.” Advil II, 675 F.Supp. at 825. We have previously upheld reliance on disputed survey data that were supported by testimony. See American Home Prods. Corp. v. Johnson & Johnson, 577 F.2d 160, 167 (2d Cir.1978). And, although AHP’s objections undermine the precision of the surveys, they do not deplete them of all value. The district court, therefore, had ample basis to find the research relevant, and did not abuse its discretion in admitting the evidence.

In addition, Judge Conner expressly acknowledged that AHP’s contentions were “not without significant merit.” Advil II, 675 F.Supp. at 825. Accordingly, he considered only the surveys' general thrust, and not their actual quantitative results in determining that McNeil had established a likelihood of success on the merits. Advil II, 675 F.Supp. at 825. This determination, consequently, was not clearly erroneous and must be affirmed.

One further issue merits discussion. The district court wrote, “Where false or misleading advertising in violation of the Lanham Act is shown, irreparable harm is presumed.” Advil II, 675 F.Supp. at 826. The case it relied upon for this proposition, however, states, “The likelihood of injury and causation will not be presumed, but must be demonstrated in some manner.” Coca-Cola, 690 F.2d at 316 (emphasis added); see also Johnson & Johnson v. Carter-Wallace, Inc., 631 F.2d 186, 189-90 (2d Cir.1980). This apparent confusion need not detain us; an important distinction must be drawn between the instant case and Coca-Cola and Johnson & Johnson.

Both Coca-Cola and Johnson & Johnson involved misleading, non-comparative commercials which touted the benefits of the product advertised but made no direct reference to any competitor’s product. The injury in such cases accrues equally to all competitors; none is more likely to suffer from the offending broadcasts than any other. The Lanham Act, however, only authorizes actions by one “who believes that he is or is likely to be damaged.” 15 U.S.C. § 1125(a). Thus, we required some indication of actual injury and causation to satisfy Lanham Act standing requirements and to ensure a plaintiff’s injury was not speculative. See Johnson & Johnson, 631 F.2d at 189.

This case, by contrast, presents a false comparative advertising claim. Thus, the concerns voiced in Coca-Cola and Johnson & Johnson regarding speculative injury do not arise. A misleading comparison to a specific competing product necessarily diminishes that product’s value in the minds of the consumer. By falsely implying that Advil is as safe as Tylenol in all respects, AHP deprived McNeil of a legitimate competitive advantage and reduced consumers’ incentive to select Tylenol rather than Advil. This is analogous to a Lanham Act trademark dispute. An infringing mark, by its nature, detracts from the value of the mark with which it is confused. In that context, we recently confirmed that irreparable harm will be presumed. See Home Box Office, Inc. v. Showtime/The Movie Channel Inc., 832 F.2d 1311, 1314 (2d Cir. 1987); Charles of the Ritz Group, 832 F.2d at 1321; Standard & Poor’s Corp. v. Commodity Exchange, Inc., 683 F.2d 704, 708 (2d Cir.1982). Consequently, the district court did not err in presuming harm from a finding of false or misleading advertising.

*39 The decision of the district court is affirmed.