Eli Lilly and Company v. Arla Foods USA, Inc.

• Court: Court of Appeals for the Seventh Circuit
• Decided: June 15, 2018
• Docket: 17-2252
• Status: Published

Opinion

In the

United States Court of Appeals For the Seventh Circuit ____________________ No. 17-2252 ELI LILLY AND COMPANY and ELANCO US, INC., Plaintiffs-Appellees,

v.

ARLA FOODS, INC., and ARLA FOODS PRODUCTION LLC, Defendants-Appellants. ____________________

Appeal from the United States District Court for the Eastern District of Wisconsin. No. 17-C-703 — William C. Griesbach, Chief Judge. ____________________

ARGUED SEPTEMBER 13, 2017 — DECIDED JUNE 15, 2018 ____________________

Before BAUER, ROVNER, and SYKES, Circuit Judges. SYKES, Circuit Judge. In 2017 Arla Foods, a global dairy conglomerate based in Denmark, launched a $30 million advertising campaign aimed at expanding its cheese sales in the United States. Branded “Live UnprocessedTM,” the campaign covers all major media platforms and targets the growing market for all-natural foods. To that end, the ads 2 No. 17-2252

assure consumers that Arla cheese contains no “weird stuff” or “ingredients that you can’t pronounce”—in particular, no milk from cows treated with recombinant bovine somato- tropin (“rbST”), an artificial growth hormone. The flagship ad in the campaign features a vivid rhetorical flourish implying that milk from rbST-treated cows is unwholesome. Narrated by a seven-year-old girl, the ad depicts rbST as (among other things) a cartoon monster with razor sharp horns and electric fur. Enter Eli Lilly & Company and its subsidiary, Elanco US, Inc. (collectively, “Elanco”). Elanco makes the only FDA- approved rbST supplement and markets it under the brand name Posilac®. Soon after the Arla ad campaign debuted, Elanco filed suit alleging that the ads contain false and misleading statements in violation of the Lanham Act. Elanco simultaneously moved for a preliminary injunction and supported the motion with copies of the ads, scientific literature documenting rbST’s safety, and evidence that a major cheese producer had decreased its demand for rbST in response to the ad campaign. The district judge concluded that Elanco has a reasonable likelihood of success on the merits and issued the requested injunction. The judge later modified his order to cure technical deficiencies. Arla appeals, arguing that Elanco (1) failed to produce consumer surveys or other reliable evidence of actual con- sumer confusion; and (2) did not submit adequate evidence linking the ad campaign to decreased demand for its rbST. Arla also challenges the modified injunction as vague and overbroad and lacking adequate factual findings. We affirm. Consumer surveys or other “hard” evidence of actual consumer confusion are unnecessary at the No. 17-2252 3

preliminary-injunction stage. And the evidence of causation is sufficient at this stage of the proceedings: the harm is easily traced because Elanco manufactures the only FDA- approved rbST supplement on the market. Finally, the modified injunction is sufficiently definite and adequately supported by the record and the judge’s findings. I. Background On April 25, 2017, Arla rolled out a $30 million advertis- ing campaign designed to expand its presence in the United States. Dubbed “Live UnprocessedTM,” the campaign centers on “Americans’ increasingly voracious desire to know more about the products they’re eating and feeding their fami- lies.” The campaign spans all platforms: television commer- cials, YouTube advertisements, social-media outreach, in- store advertising, and a website. Two 30-second television commercials form the center- piece of Arla’s campaign. One is specifically at issue here. The commercial opens with this caption: “Arla Cheese Asked Kids: What is r[b]ST?” As the audience watches a cartoon of a six-eyed monster and a fisherman, a seven-year- old girl named Leah narrates: “RbST has razor sharp horns. It’s so tall that it could eat clouds. You may want to pet it but the fur is electric.” The commercial then cuts to a scene of Leah enjoying a cheese sandwich, and the narrator’s voice switches to that of an adult woman: “Actually, rbST is an artificial growth hormone given to some cows, but not the cows that make Arla cheese. No added hormones. No weird stuff. Arla, live unprocessed.” A small written disclaimer appears for a few seconds toward the end of the commercial: “Made with milk from cows not treated with r[b]ST. No 4 No. 17-2252

significant difference has been shown between milk derived from r[b]ST-treated and non r[b]ST-treated cows.” Other parts of the ad campaign reiterate the claim that Arla cheese contains “no weird stuff.” Arla defines “weird stuff” on its website: No artificial additives. No ingredients that you can’t pronounce. No ingredients that sound confusing or in any way like a made-up word. No ingredients with names that sound like they may be aliens with nine arms, beasts with electric fur, gigantic robots[,] or bears in dis- guise. No artificial growth hormones like r[b]ST.* … Nor anything else artificial[] be- cause our cheese has always been made with simple ingredients and never anything weird. The asterisk directs readers to another part of the website containing the same disclaimer that appears in small print in the television commercial. Elanco produces and sells the only FDA-approved rbST supplement under the brand name Posilac®. On May 19, 2017, Elanco filed suit against Arla and simultaneously moved for a preliminary injunction. The suit alleges that Arla’s ad campaign makes false and misleading statements concerning the composition, health, and safety of dairy products made from milk from rbST-treated cows. The complaint asserts claims for violation of the Lanham Act, 15 U.S.C. § 1125(a), and the Wisconsin Deceptive Trade Practices Act, WIS. STAT. § 100.20(1). As factual support for a preliminary injunction, Elanco submitted copies of the advertisements themselves, scientific No. 17-2252 5

studies and expert testimony about the safety of rbST, the FDA’s regulatory guidance regarding rbST-related adver- tisements, and confidential information that a major cheese producer chose to terminate its use of rbST partially in response to Arla’s advertisements. The judge held an evidentiary hearing and thereafter granted Elanco’s motion for preliminary injunctive relief. After reviewing the scientific evidence submitted by the parties, the judge found that milk from rbST-treated cows is equally safe and healthy for human consumption as other milk. He then concluded that Arla’s ads contain a “mislead- ing message that cheese from cows treated with rbST is dangerous, unhealthy, and something that you should not feel good about feeding to your family.” The judge prelimi- narily enjoined Arla from disseminating the ads and any others “substantially similar thereto.” The order also prohib- ited Arla from making claims disparaging rbST or Posilac®. Arla sought interlocutory review as permitted under 28 U.S.C. § 1292(a)(1). In its opening brief, Arla identified two technical defects in the judge’s order: it impermissibly incorporated documents by reference, and it failed to specify which advertisements or specific promotional claims were enjoined. In response Elanco returned to the district court and asked the judge to cure the technical problems in a modified order. The judge did so, issuing a modified injunc- tion pursuant to Rule 62(c) of the Federal Rules of Civil Procedure. II. Discussion To win a preliminary injunction, the moving party must establish that (1) without preliminary relief, it will suffer 6 No. 17-2252

irreparable harm before final resolution of its claims; (2) legal remedies are inadequate; and (3) its claim has some likelihood of success on the merits. BBL, Inc. v. City of Angola, 809 F.3d 317, 323–24 (7th Cir. 2015). If the moving party makes this showing, the court balances the harms to the moving party, other parties, and the public. Id. For purposes of this appeal, Arla concedes that rbST- derived dairy products are of the same quality, nutrition, and safety as other dairy products. 1 Arla instead focuses on whether Elanco produced sufficient evidence to establish a

1 Arla would have trouble contesting the point anyway. The FDA has twice confirmed the safety of rbST-derived dairy products. It initially approved Posilac® in 1993 after determining that rbST “is safe and effective for dairy cows, that milk from rbST-treated cows is safe for human consumption, and that production and use of the product do not have a significant impact on the environment.” Interim Guidance on the Voluntary Labeling of Milk and Milk Products from Cows that Have Not Been Treated with Recombinant Bovine Somatotropin, 59 Fed. Reg. 6279, 6279–80 (Feb. 17, 1994). The FDA reexamined the possible health risks of rbST in 2016, confirming that rbST is “safe and effective for its intended uses and that there is no significant difference between milk from cows treated with [rbST] and untreated cows.” U.S. FOOD & DRUG ADMIN., U.S. DEP’T OF HEALTH & HUMAN SERVS., CITIZEN PETITION DENIAL RESPONSE FROM FDA CDER TO THE CANCER PREVENTION COALITION, ET AL., 16 (Mar. 21, 2016), https://www.regulations.gov/document?D=FDA- 2007-P-0119-0007. In 2014 a joint panel of the United Nations and World Health Organ- ization also found “no evidence to suggest that the use of rbSTs would result in a higher risk to human health.” World Health Organization, Food and Agriculture Organization of the United Nations, Joint FAO/WHO Expert Committee on Food Additives, Evaluation of Certain Veterinary Drug Residues in Food, WHO TECHNICAL REPORT SERIES 988, 78 (2014), http://apps.who.int/iris/bitstream/10665/127845/1/9789241209885_ eng.pdf?ua=1. No. 17-2252 7

likelihood of success on the merits of its Lanham Act claim. We review an order granting a preliminary injunction for abuse of discretion. Id. at 324. Arla also challenges the scope of the injunction, arguing that it is vague and overbroad and does not meet various formal requirements of Rule 65(d) of the Federal Rules of Civil Procedure. We review these ques- tions of law de novo. Lineback v. Spurlino Materials, LLC, 546 F.3d 491, 504 (7th Cir. 2008). A. Likelihood of Success on the Merits To prevail on a deceptive-advertising claim under the Lanham Act, a plaintiff must establish that (1) the defendant made a material false statement of fact in a commercial advertisement; (2) the false statement actually deceived or had the tendency to deceive a substantial segment of its audience; and (3) the plaintiff has been or is likely to be injured as a result of the false statement. 2 Hot Wax, Inc. v. Turtle Wax, Inc., 191 F.3d 813, 819 (7th Cir. 1999). We have recognized two types of actionable statements under the Lanham Act: those that are literally false and those that are literally true but misleading. Id. at 820. The evidence re- quired to satisfy the first two elements of the claim varies according to the type of statement at issue. A literally false statement will necessarily deceive con- sumers, so extrinsic evidence of actual consumer confusion is not required. Id. We have characterized statements in this category as “bald-faced, egregious, undeniable, [and] over the top.” Schering-Plough Healthcare Prods., Inc. v. Schwarz

2 Elanco’s Lanham Act claim is the main event at this stage of the proceedings. The state-law claim is largely duplicative, so the judge didn’t need to say much about it and neither do we. 8 No. 17-2252

Pharma, Inc., 586 F.3d 500, 513 (7th Cir. 2009). The inquiry asks whether the defendant made an explicit representation of fact that on its face conflicts with reality. See BASF Corp. v. Old World Trading Co., 41 F.3d 1081, 1091 (7th Cir. 1994). For example, if Arla’s television commercial had said, “drinking milk from cows treated with rbST gravely endangers your health,” Elanco would need no additional evidence of consumer confusion to prevail on its claim. Arla’s ads make no explicit false claims about the compo- sition of or dangers posed by milk from rbST-treated cows. Indeed, the explicit statements about rbST are factually accurate: RbST is an artificial growth hormone given to some cows, and Arla does not use milk from those cows. We therefore leave to one side Elanco’s contention that Arla’s ad campaign contains literally false statements. For the other category of actionable statements—those that are literally true but misleading—the plaintiff ordinarily must produce evidence of actual consumer confusion in order to carry its burden to show that the challenged state- ment has “the tendency to deceive a substantial segment of its audience.” Hot Wax, 191 F.3d at 819–20. The parties agree that at trial this evidence typically comes in the form of consumer surveys. They dispute, however, whether Elanco was required to conduct consumer surveys or produce other hard evidence of actual consumer confusion in order to win a preliminary injunction. We’ve held that “such proofs are not required at the pre- liminary injunction stage.” Abbott Labs. v. Mead Johnson & Co., 971 F.2d 6, 15 (7th Cir. 1992). It’s not feasible to require a Lanham Act plaintiff to conduct full-blown consumer sur- veys in the truncated timeframe between filing suit and No. 17-2252 9

seeking a preliminary injunction. Here the judge properly analyzed whether the evidence Elanco submitted at this stage—the ads themselves, the regulatory guidance, and the evidence of decreased demand—established a likelihood of success on the merits. Consumer surveys were unnecessary. The judge permissibly scrutinized the content of Arla’s ads to determine whether they convey a misleading mes- sage. Some of that content is mundane and noncontrover- sial—for example, the characterization of Arla cheese as a food “[y]ou can feel good serving … to the whole family.” But the ad campaign centers on disparaging dairy products made from milk supplied by rbST-treated cows. The ads draw a clear contrast between Arla cheese (high quality, nutritious) and cheese made from rbST-treated cows (im- pure, unwholesome). The use of monster imagery, “weird stuff” language, and child actors combine to colorfully communicate the message that responsible consumers should be concerned about rbST-derived dairy products. The judge reasonably concluded that these ads are likely to mislead consumers about the wholesomeness of products made from milk supplied by rbST-treated cows. Regulatory guidance supports the judge’s conclusion. The FDA has warned that ads concerning rbST may be misleading if not placed “in proper context.” Interim Guid- ance on the Voluntary Labeling of Milk and Milk Products from Cows that Have Not Been Treated with Recombinant Bovine Somatotropin, 59 Fed. Reg. 6279, 6280 (Feb. 17, 1994). Specifically, when a food advertisement states that the product is made “from cows not treated with rbST,” the FDA recommends that the ad also include the following disclaimer: “No significant difference has been shown 10 No. 17-2252

between milk derived from rbST-treated and non-rbST- treated cows.” Id. Arla’s ads do not provide this needed context. True, the FDA’s recommended disclaimer appears in the ad cam- paign, but only in tiny print in the television commercial and in an obscure location on the webpage. Neither disclaimer dispels the central message of these advertisements: that cheese made from milk supplied by rbST-treated cows is unwholesome. Elanco also introduced confidential evidence that a major cheese producer decided to cease using milk from rbST- treated cows based in part on Arla’s ads. To be sure, this evidence of decreased demand does not reflect actual con- sumer confusion. But given the cheese producer’s economic incentive to accurately predict consumer demand, its con- cern about the ad campaign’s impact on consumers supports the judge’s conclusion that the ads convey a misleading message about rbST. Given Arla’s concession that rbST- derived dairy products are no different than other dairy products, all the available evidence at this stage—the ads themselves, the FDA’s regulatory guidance, and the evi- dence of decreased demand—points in the same direction: Elanco has a reasonable likelihood of success on the merits of its Lanham Act claim. 3

3 Elanco argues in the alternative that Arla’s ads are false by necessary implication. This Lanham Act doctrine allows a plaintiff to bypass presenting evidence of consumer confusion altogether “[i]f the words or images, considered in context, necessarily imply a false message.” Time Warner Cable, Inc. v. DIRECTV, Inc., 497 F.3d 144, 158 (2d Cir. 2007). At least five other circuits use this “misleading per se” doctrine for Lanham Act claims. See id.; Scotts Co. v. United Indus. Corp., 315 F.3d 264, 274 (4th No. 17-2252 11

Arla next attacks the sufficiency of the evidence on the element of causation—that is, whether Elanco established a likelihood of success that the ads actually drove away cus- tomers. Causation requires proof of an injury to a commer- cial interest in sales or business reputation proximately caused by the defendant’s misrepresentations. Lexmark Int’l, Inc. v. Static Control Components, Inc., 134 S. Ct. 1377, 1395 (2014). As an initial matter, Arla contends that the judge im- properly skipped the causation question in his preliminary- injunction analysis. Not so. After concluding that Arla’s ads are misleading, the judge specifically addressed Elanco’s evidence that a major cheese producer stopped using milk from rbST-treated cows based in part on the ad campaign. A more extended treatment of the causation question was largely unnecessary given how easy it is to trace Elanco’s harm. Elanco sells the only FDA-approved rbST supplement in the United States, so any false or misleading advertising regarding rbST that decreases demand for the supplement will necessarily harm Elanco. And Elanco’s evidence of a large cheese producer’s response to Arla’s ad campaign is sufficient factual support for the judge’s decision to issue a preliminary injunction. Nothing more was needed at this stage.

Cir. 2002); Clorox Co. P.R. v. Proctor & Gamble Commercial Co., 228 F.3d 24, 34–35 (1st Cir. 2000); Southland Sod Farms v. Stover Seed Co., 108 F.3d 1134, 1139 (9th Cir. 1997); Castrol Inc. v. Pennzoil Co., 987 F.2d 939, 946–47 (3d Cir. 1993). Because we agree with the judge that Elanco presented sufficient evidence of consumer confusion at this stage of the proceed- ings, we do not need to consider the alternative theory that the ads are false by necessary implication. 12 No. 17-2252

B. Challenges to the Form of the Injunction As we’ve noted, in its opening brief, Arla raised several objections to the form of the original injunction. At Elanco’s request, the judge issued a modified injunction responding to these objections. That’s a permissible procedure. The civil rules allow the district court to modify an injunction to maintain the status quo pending appeal. FED. R. CIV. P. 62(c). More to the point, the court is authorized to “aid[] in the appeal” by resolving technical objections or clarifying imprecise wording. See Dixon v. Edwards, 290 F.3d 699, 709 n.14 (4th Cir. 2002). The judge did just that when he issued the modified injunction. Arla continues to press some of its objections to the form of the injunction. Most prominently, it contends that the injunction is vague and overbroad, so we’ll start there. Rule 65 requires that every injunction “(A) state the reasons why it issued; (B) state its terms specifically; and (C) describe in reasonable detail—and not by referring to the complaint or other document—the act or acts restrained or required.” FED. R. CIV. P. 65(d). The specificity requirements in the rule “spare[] courts and litigants from struggling over an injunction’s scope and meaning by informing those who are enjoined of the specific conduct regulated by the injunction and subject to con- tempt.” Patriot Homes, Inc. v. Forest River Hous., Inc., 512 F.3d 412, 415 (7th Cir. 2008) (internal quotation marks omitted). Whether a particular injunction is overbroad or vague is necessarily a highly fact-bound inquiry. “[An] injunction must … be broad enough to be effective, and the appropriate scope of the injunction is left to the district court’s sound No. 17-2252 13

discretion.” Russian Media Grp., LLC v. Cable Am., Inc., 598 F.3d 302, 307 (7th Cir. 2010). Section 1 of the modified injunction prohibits Arla from disseminating the advertisement attached as Exhibit 1 to the order (we’ve described it above) and “any other advertise- ment substantially similar thereto” that “claims, either directly or by implication,” that rbST or Posilac® is “a massive beast, animal, or monster” with “razor sharp horns,” or “is so tall it could eat clouds,” or has “electric fur,” or “electrocutes or harms people.” This section of the injunction also prohibits Arla from disseminating any adver- tisement that claims or implies that rbST or Posilac® is anything other than an artificial hormone that prolongs the lactation of dairy cows. Arla insists that the “substantially similar” language is vague and overbroad. We disagree. The Lanham Act’s prohibition on implied falsehoods makes the use of some- what inexact language unavoidable. See Scandia Down Corp. v. Euroquilt, Inc., 772 F.2d 1423, 1432 (7th Cir. 1985) (“When the difficulty stems from the inability of words to describe the variousness of experience, the court may prefer brief imprecise standards to prolix imprecise standards.”). Read as a whole, the modified injunction essentially prohibits Arla from portraying rbST as something it’s not. That’s sufficient- ly definite, especially when considered in the context of the rest of the order. See S.C. Johnson & Son, Inc. v. Clorox Co., 241 F.3d 232, 238 (2d Cir. 2001) (finding a bar on “false or misleading descriptions” sufficiently specific when read in the context of the district court’s orders). Arla also objects to section 2(e) of the modified injunc- tion, which prohibits any advertisement that claims, “either 14 No. 17-2252

directly or by implication,” that “consumers should not feel ‘good about eating’ or ‘serving to [their] friends and family’ dairy products made from milk of cows supplemented with rbST or Posilac®.” Arla objects that this part of the injunction forecloses any positive statement that its own products are “something you should feel good about eating.” That fear is unfounded. The prohibited negative inference can arise only if an Arla advertisement specifically mentions rbST or Posilac® in a disparaging way. Nothing in the injunction prohibits Arla from claiming that consumers can feel good about eating its own products. Finally, Arla argues that the modified injunction lacks adequate factual findings. Two rules govern here. Rule 52(a)(2) of the Federal Rules of Civil Procedure requires that the court “state the findings and conclusions that sup- port its action.” Rule 65(d)(1) requires that every injunction “state the reasons why it was issued.” A court “need only make brief, definite, [and] pertinent findings” to support its order. FED. R. CIV. P. 52 advisory committee’s note to 1946 amendment. Arla first complains that section 2(a) of the injunction broadly prohibits any statements claiming that rbST is “dangerous or unsafe,” but the judge’s factual finding in support of this prohibition says only that rbST-derived dairy products are not dangerous or unsafe for human consumption. This is hair-splitting. Rule 52(a)(2) does not require “over- elaboration of detail” or “particularization of facts.” Id. This case concerns Arla’s misleading advertising claims about the safety of rbST-derived dairy products for human consump- tion; it has nothing to do with other uses of rbST-derived dairy products. There is no ambiguity here. No. 17-2252 15

Second, Arla contends that the judge’s factual findings do not adequately support Section 2(e) of the injunction, which again prohibits any direct or implied claim that consumers should not feel “good about eating” cheese made from milk supplied by rbST-treated cows. To the contrary, the order traces the scientific evidence regarding rbST before concluding that milk from rbST-treated cows is just as safe and healthy as other milk. In sum, we find no abuse of discretion, no substantive legal error, and no defect in the form or content of the in- junction. The judge’s order therefore is AFFIRMED. 16 No. 17-2252

ROVNER, Circuit Judge, concurring in part and concurring in the judgment. I join all but one discrete portion of Judge Sykes’ well-reasoned opinion. For purposes of this interlocu- tory appeal, Arla has not challenged the district court’s find- ing that its advertisements are likely to mislead consumers. Apart from arguing (erroneously, I agree) that Elanco was required to submit proof of actual consumer confusion at this stage of the proceedings, Arla has not quarreled with the district court’s preliminary analysis of the content of the ads and the message they convey to consumers. See Arla Reply Br. 10. Consequently, there is no need for us to address that aspect of Elanco’s Lanham Act claim at this time.