EXECUTIVE SUMMARY Government-mandated health warnings have been required to appear on packages for more than 50 years and on cigarette advertisements for more than 45 years. Today, cigarette is at an all-time low. PM USA and believe that adult smokers should have the right to make informed choices about the use of products and that there should be a single, consistent public health message about the risks of smoking that should help guide these choices. While we recognize the need for health warnings on cigarette packages and advertisements, FDA’s second attempt at a regulation requiring graphic health warnings suffers from a number of fatal flaws. Our comments are organized into four sections. The first section urges FDA to exempt HeatSticks®, which are unique non-combustible “” designed for use with the IQOS® tobacco heating system, from this rule. Although those products do not eliminate all health risks associated with tobacco and use, they are materially different from traditional combustible cigarettes, and, we believe, also significantly less harmful. Indeed, this Agency has already determined that the marketing of this product is “appropriate for the protection of public health.” The second section raises practical concerns impacting the implementation of the Proposed Rule. The third section explains that without substantial changes, the Proposed Rule would violate the Administrative Procedure Act (“APA”) on multiple grounds. The fourth, and final, section demonstrates that the text and graphic warnings proposed by this rule would violate free speech protections guaranteed by the First Amendment of the U.S. Constitution. To guide the Agency in the materials that follow, we provide an overview of the four sections listed above. Any Final Rule Should Exempt HeatSticks® from the Proposed Graphic-Warnings Requirements IQOS® is an electronic device that heats tobacco in HeatSticks® to generate nicotine-containing without , fire, ash, or . After a thorough science-based review, FDA granted marketing authorization for IQOS® and HeatSticks® in April 2019, concluding that marketing those products would be “appropriate for the protection of public health.”1 FDA based this conclusion on its rigorous assessment of the critical differences between HeatSticks® and traditional, combustible cigarettes. FDA noted several material differences—among them, that HeatSticks® do not combust or produce smoke; that HeatSticks® reduce exposures to harmful and potentially harmful constituents (“HPHCs”); that HeatSticks® are “likely to lead to less risk of tobacco-related diseases;” and that HeatSticks® have the potential to benefit current smokers who switch to them.2 Because HeatSticks® qualify as a “cigarette” as that term is defined in the Federal Cigarette Labeling and Advertising Act (“FCLAA”),3 FDA’s Marketing Order requires HeatSticks® packaging and advertising to bear the rotating Surgeon General’s warnings required by that Act. Significantly, however, based on the Agency’s recognition of significant differences

1 Attachment 1, CTP Technical Project Lead Review (TPL) Review of IQOS PMTA: PM0000424, PM0000425, PM0000426, and PM0000479 (April 29, 2019) (“IQOS® TPL Review”) at 12, available at https://www.fda.gov/media/124247/download. 2 Id. at 11-12, 92, 95; see also id. at 64-65. 3 15 U.S.C. § 1332(1).

2 between HeatSticks® and traditional combustible cigarettes, FDA also changed the mandated textual warnings that accompany HeatSticks®, for instance by excluding the Surgeon General’s warning.4 Although IQOS® and HeatSticks® do not eliminate all health risks associated with tobacco and nicotine use, we believe they are significantly less harmful than traditional, combustible cigarettes, and FDA has determined that the marketing of this unique of “cigarette” is “appropriate for the protection of public health.” Notwithstanding the differences between HeatSticks® and combustible cigarettes, the Proposed Rule would apply the graphic-warnings requirements to HeatSticks® even though those graphic warnings would not educate consumers about the health risks of using HeatSticks®. To the contrary, those warnings would misinform consumers by suggesting, inaccurately and without evidence, that HeatSticks® have the exact same health risks as traditional combustible cigarettes, and by describing HeatSticks® as involving “smoking” or “tobacco smoke.” Imposing the graphic-warnings requirements on HeatSticks®, in short, would undermine FDA’s stated policy goal of promoting greater public understanding of the negative health consequences of cigarette smoking. It would also frustrate FDA’s strategy of adopting policies to move smokers to non-combusted tobacco products lower on the continuum of risk. And, in addition to undermining important health policy objectives, imposing the graphic-warnings requirements on HeatSticks® would violate the APA and the U.S. Constitution on multiple grounds. For these reasons, we respectfully urge FDA to exempt HeatSticks® from the scope of any final rule. The Proposed Rule Raises Practical Concerns Impacting the Implementation of the Required Warnings on Cigarette Packages and Advertisements The Proposed Rule, if finalized in its current form, would create significant implementation problems for both regulated industry and FDA that would substantially hinder the ability of the regulated industry to comply with the requirements of the rule. FDA should modify the Proposed Rule to address each of these problems. First, as an initial matter, FDA should reduce the number of required graphic warnings from 13 to either 12 or 9. This modification is necessary due to extremely difficult and potentially insurmountable technical challenges with printing and distributing 13 warnings equally and randomly on cigarette packages. Second, FDA should address other issues, addressed in more detail below, impacting the implementation of the required warnings on cigarette packages to provide flexibility to manufacturers and to account for the variety of cigarette package shapes and sizes in the marketplace. Third, FDA should address issues impacting the implementation of the required warnings on cigarette advertisements. These issues include the need for FDA to account for and provide clear direction regarding different formats and sizes.

4 See Attachment 2, Philip Morris Products S.A. Marketing Order, FDA Submission Tracking Numbers (STNs): PM0000424-PM0000426, PM0000479 (April 30, 2019) (“FDA Marketing Order”) at 13 n.9, available at https://www.fda.gov/media/124248/download.

3 Finally, FDA should extend or toll the implementation deadline to account for the complex and interdependent chain of events necessary for manufacturers to implement graphic health warnings. The Proposed Rule, if Finalized Without Substantial Changes, Would Violate the Administrative Procedure Act The Proposed Rule, if adopted without substantial modifications, would violate the APA. Agency action, findings, and conclusions must be set aside where they exceed the Agency’s authority, fail to comply with statutory requirements, or are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.”5 The Proposed Rule cannot survive scrutiny under these standards. First, FDA failed to disclose sufficient information about referenced research studies to allow for meaningful evaluation of their validity, in violation of both the Family Smoking Prevention and Act (“TCA”) and general principles of administrative law. Nor has FDA provided sufficient information for stakeholders to determine why it highlighted certain diseases over others in the graphic warnings; why it selected the particular graphics it chose; or how or why it determined that the proposed warning statements are comprehensible to consumers. Second, FDA relies upon unreliable evidence that does not adequately support the Proposed Rule. FDA’s core evidence consists of commissioned studies that are seriously flawed in their design. FDA’s other cited studies fare no better, relying on data from non-U.S. populations without justification and in disregard of FDA’s own scientific standards, and suffering from various methodological defects. Individually, and even more so collectively, these flaws call into question the reliability and soundness of FDA’s primary scientific justifications for the Proposed Rule. Third, even if FDA’s studies were reliable, their results do not support FDA’s conclusion that the proposed graphic warnings would be effective in improving public understanding of the health consequences of smoking. FDA’s own commissioned studies are the only scientific evidence invoked by the Proposed Rule that even purport to test the replicability of the actual proposed warnings. And their results do not support the Proposed Rule. Fourth, in violation of the APA, FDA does not adequately address science and evidence that runs contrary to its proposal. FDA either ignores or fails to adequately explain a litany of evidence that undermines virtually every premise of the Proposed Rule. This includes: (i) evidence that graphic warnings do not meaningfully influence consumer knowledge regarding the consequences of smoking; (ii) evidence that graphic warnings may backfire, prompting consumers to avoid or reject them in fear or disgust; (iii) evidence that the proposed size requirements for packaging and advertisements do not advance consumer understanding; (iv) evidence indicating that consumers already understand the health consequences of smoking; and (v) evidence regarding the differential effect of these graphic warnings on particularly critical subpopulations. Fifth, FDA fails to assess alternatives to the proposed warnings, even though the APA requires it to do so. The Proposed Rule does not adequately evaluate, or justify the rejection of, other

5 5 U.S.C. § 706(2).

4 means of communicating information that would satisfy FDA’s asserted goal of educating the public about the health consequences of cigarette smoking without burdening manufacturers’ speech. Sixth, FDA adduces no evidence whatsoever in support of its proposal to include two different graphic warnings as accompaniments to its textual warning concerning COPD. The Proposed Warnings, if Finalized, Would Violate the First Amendment Forcing manufacturers to place the proposed warnings on cigarette packages and advertisements would violate the First Amendment. The “freedom of speech” guaranteed by the First Amendment “includes both the right to speak freely and the right to refrain from speaking at all.”6 Before FDA can restrict manufacturers’ speech, FDA must demonstrate that it has met legal requirements put in place to protect manufacturers’ First Amendment rights. FDA cannot meet that burden. The Proposed Rule falls far short of satisfying any applicable legal standard, including those cited by FDA, and therefore violates the First Amendment, for at least three reasons. First, the size and placement requirements of the Proposed Rule unconstitutionally burden speech. The Proposed Rule requires cigarette manufacturers to dedicate 50% of their packages and 20% of their advertisements to promoting the government’s message through large, prominently placed warning statements. By commandeering such a large amount of premium space, FDA “drowns out the [manufacturer’s] own message” and significantly restricts manufacturer speech.7 FDA has failed to carry its burden to show that those size and placement requirements are not “unjustified or unduly burdensome” means to further its asserted interest in improving consumer understanding of smoking-related health risks, much less that the requirements are no more extensive than necessary.8 FDA has proffered no evidence to show that smaller warnings – such as those covering 30% of a package or 15% of advertising space, for example – would be insufficient to further the Agency’s stated consumer-education interest. Second, the proposed graphic warnings are impermissible speaker-, content- and viewpoint- based regulations of speech. These graphics would turn cigarette packages and advertisements into mini-billboards that affirmatively dissuade consumers from purchasing the product and appeal to emotion to further the government’s anti-smoking position. Indeed, the Proposed Rule suffers from many of the same fatal flaws that led the D.C. Circuit to invalidate FDA’s 2011 graphic warning rule—among them, that the graphics do not convey any new and “uncontroversial” “warning information” beyond the text, are “primarily intended to evoke an emotional response, or, at most, shock the viewer into retaining the information in the text warning,” and constitute an effort to “browbeat consumers into quitting” smoking rather than make “an accurate statement about cigarettes.”9 Moreover, FDA again asserts a problematic interest in “effective” communication of health information that the D.C. Circuit has considered and rejected. Under any potentially applicable First Amendment standard, FDA has failed to

6 Janus v. Am. Fed’n of State, Cty., & Mun. Employees, Council 31, 138 S. Ct. 2448, 2463 (2018) (internal quotation marks omitted). 7 Nat’l Inst. of Fam. & Life Advocates v. Becerra, 138 S. Ct. 2361, 2378 (2018) (“NIFLA”). 8 Id. at 2372 (internal quotation marks omitted). 9 R.J. Reynolds Tobacco Co. v. FDA, 696 F.3d 1205, 1216-17 (D.C. Cir. 2012), overruled on other grounds by Am. Meat Inst. v. U.S. Dep’t of Agric., 760 F.3d 18 (D.C. Cir. 2014) (en banc).

5 satisfy its heavy burden to justify the speaker-, content-, and viewpoint-based speech restrictions that the proposed graphic warnings would impose. Third, the proposed graphic and textual warnings are all but certain to mislead consumers regarding the health risks associated with cigarette smoking. These warnings mislead by emphasizing extreme and unlikely conditions and treat conditions of differing severity as if they were equal. By grouping together conditions with different prevalence and severity (e.g., diseases with high mortality versus diseases that are not fatal), the proposed warnings make certain diseases appear to be equivalent or more likely than they actually are. The fact that the proposed warnings would likely mislead consumers renders them invalid. Each of the proposed warnings is constitutionally deficient. This submission includes a table, see infra Section IV.D, reviewing the constitutional deficiencies of the content of each warning.

I. THE PROPOSED RULE SHOULD EXEMPT HEATSTICKS® AND ACCOUNT FOR IMPLEMENTATION PROBLEMS IQOS® is an electronic device that heats tobacco in HeatSticks® and produces a nicotine- containing tobacco aerosol without combustion. Under FDA’s Proposed Rule, HeatSticks®, a roll of tobacco wrapped in paper, would be subject to the same graphic-warnings requirements applicable to traditional, combustible cigarettes. FDA has previously acknowledged that HeatSticks® heated with the IQOS® device are materially different from traditional cigarettes, are likely to present lower health risk, and have the potential to benefit current smokers who switch to these products. We urge FDA not to apply the graphic health warnings requirements to HeatSticks®. Requiring HeatSticks® to bear the same graphic health warnings as those appearing on traditional, combustible cigarettes, despite the differences that FDA itself has acknowledged, would undermine the Agency’s stated policy goal of promoting greater public understanding of the negative health consequences of cigarette smoking. It would also impede FDA’s strategy to adopt policies that help move current smokers down the continuum of risk. In short, FDA should exempt HeatSticks® from the final rule as a matter of sound public health policy. Further, applying any final rule to HeatSticks® would violate the APA and the Constitution. A. FDA should exempt HeatSticks® from any final rule addressing the health risks presented by traditional, combustible cigarettes In April 2019, FDA granted marketing authorization for IQOS® and HeatSticks®, which generate a nicotine-containing aerosol without combustion, fire, ash, or smoke. In its technical evaluation of the products, FDA concluded that the marketing of IQOS® and HeatSticks® would be “appropriate for the protection of public health,” due in part to FDA’s recognition that the products are likely to present less risk than traditional combustible cigarettes and have the potential to benefit current smokers who switch to them. Notwithstanding such findings, the Proposed Rule applies the same graphic-warnings requirements to HeatSticks®. Subjecting HeatSticks® to the Proposed Rule’s graphic-warnings requirements would misinform consumers and undermine significant public health objectives. FDA should accordingly make clear that the scope of any final rule does not extend to these products.

6 B. Applying the Proposed Rule to HeatSticks® would hamper FDA’s stated public health policy goals FDA’s determinations regarding the risk profile for IQOS® and HeatSticks® as compared to traditional, combustible cigarettes illustrate that applying graphic-warnings requirements to HeatSticks® would undercut important health objectives. On April 30, 2019, following a nearly two-and-a-half-year period including review by experts from 13 disciplines, FDA granted marketing authorization for the IQOS® System Holder and Charger, HeatSticks®, Marlboro Smooth Menthol HeatSticks®, and Marlboro Fresh Menthol HeatSticks®.10 In the Marketing Order, FDA “find[s] that the marketing of the new tobacco products . . . is appropriate for the protection of public health, and that you have met the other requirements of section 910(c) of the FD&C Act.”11 FDA expressly recognized that use of these products has “the potential for a relative benefit” for smokers who switch completely to IQOS® tobacco heating system as compared to traditional cigarettes and is “likely to lead to less risk of tobacco-related diseases.”12 FDA based its findings on its analysis of scientific evidence showing how HeatSticks® used with the IQOS® device materially differ from combustible cigarettes. FDA’s Technical Project Lead Review for IQOS® categorized HeatSticks® as a “filtered non-combusted cigarette.”13 FDA’s technical review concluded that “many of the known HPHCs found in [combustible cigarette] smoke are very low or undetectable in Heatstick .”14 FDA further found that switching completely to these products would reduce exposure to harmful and potentially harmful constituents.15 For example, FDA determined that “although the use of HeatSticks in the IQOS device does produce [carbon monoxide], the exposure to [carbon monoxide] from IQOS use is comparable to environmental exposure,” further differentiating HeatSticks® from traditional cigarettes.16 FDA then clarified in its Marketing Order that its findings necessitated different warnings for HeatSticks® than for traditional cigarettes. In light of these recognized differences, applying the Proposed Rule to HeatSticks® products, so as to convey to consumers that these products present the same health risks as traditional combustible cigarettes, would mislead and misinform the public. Indeed, given that HeatSticks® must also bear an additional nicotine warning not required for traditional cigarettes,17 consumers might even infer that HeatSticks® used with the IQOS® device are more harmful than traditional cigarettes. Beyond that, HeatSticks® do not combust and do not create smoke and, yet, each of the required textual warnings include the words “tobacco smoke” or “smoking”—further misleading consumers about what these products contain and how they work. Requiring

10 Representatives from the following disciplines participated in the review of the IQOS® premarket application: Behavioral and Clinical Pharmacology, Chemistry, Environmental Science, Engineering, Epidemiology, Medical, Microbiology, Social Science, Statistics, Toxicology, Office of Compliance and Enforcement: DEM, DPAL, and BIMO. FDA’s review included premarket deliberation by the Office of Science, as well as postmarket deliberation by the Office of Compliance and Enforcement. 11 Attachment 2, FDA Marketing Order at 1. 12 Attachment 1, IQOS® TPL Review at 42, 11. 13 Id. at 15. 14 Id. at 91-92. 15 Id. at 11, 91-92. 16 Id. at 90. 17 See infra Section I.C.2.

7 HeatSticks® packaging and advertising to bear graphic warnings designed for traditional cigarettes, in short, would undermine the stated objective of FDA’s rulemaking, i.e., to accurately inform consumers about health risks.18 A final rule that misleads consumers in this manner would also undermine FDA’s regulatory strategy of encouraging smokers to move to products that are lower on the continuum of risk. In July 2017, FDA announced a new comprehensive regulatory plan for nicotine and tobacco that formally acknowledged the continuum of risk between various nicotine-containing products.19 FDA announced that a regulatory framework “where adults who still need or want nicotine could get it from alternative and less harmful sources, needs to be the cornerstone of our efforts.”20 At the time, FDA decided to extend premarket submission deadlines for innovative, potentially less harmful tobacco products, so that manufacturers would have more time to prepare product applications and FDA would have more time to establish foundational rules.21 FDA has since authorized IQOS® and three HeatSticks® products under its Premarket Tobacco Product Application (“PMTA”) pathway, concluding that the products are “likely to lead to less risk of tobacco-related diseases” and have “the potential for a relative benefit” for smokers as compared to traditional cigarettes.22 By encouraging smokers to get nicotine from products that are “likely to lead to less risk of tobacco-related diseases,”23 the availability of IQOS® and HeatSticks® furthers FDA’s goal of shifting smokers down the “continuum of risk.” Applying the Proposed Rule to HeatSticks® products would counteract that goal by misleading consumers about those products’ place on the continuum of risk. This would be particularly detrimental to public education and health because, despite overwhelming scientific evidence that non-combustible tobacco products contain less harmful and potentially harmful constituents than combustible tobacco products, many tobacco users still misjudge the risk.24 Rather than contributing to such misperceptions, FDA should accurately inform consumers about the relative risks attributable to different tobacco products. We thus urge FDA to exempt HeatSticks® from the final graphic-warnings rule. C. Applying the Proposed Rule to HeatSticks® would be inconsistent with FDA’s legal obligations In addition, any final rule that does not exempt HeatSticks® would violate the APA for at least three independent reasons.

18 See, e.g., 84 Fed. Reg. at 42,785 (articulating FDA’s interest in “promoting greater public understanding of the negative health consequences of smoking”). 19 FDA Press Release, “FDA announces comprehensive regulatory plan to shift trajectory of tobacco-related disease, ” (July 27, 2017), available at: https://www.fda.gov/news-events/press-announcements/fda-announces- comprehensive-regulatory-plan-shift-trajectory-tobacco-related-disease-death. 20 Id. 21 FDA recently issued a proposed PMTA rule on September 25, 2019. See generally Premarket Tobacco Product Applications and Recordkeeping Requirements, 84 Fed. Reg. 50,566. 22 Attachment 1, IQOS® TPL Review at 11, 95. 23 Id. at 11. 24 Data excerpted from “Highlighted Findings from Wave 1 of the Population Assessment of Tobacco and Health (PATH) Study.” Presented by Andrew Hyland Ph.D. at the 2016 Annual Meeting of the Society for Research on Nicotine and Tobacco, March 2016.

8 • First, because FDA’s justifications for the proposed graphic warnings apply only to traditional, combustible cigarettes, FDA has failed to carry its burden of showing a connection between its evidence and its choice to apply the Proposed Rule to HeatSticks®. • Second, applying the Proposed Rule to HeatSticks® would contradict without explanation FDA’s prior conclusions regarding the health risks associated with these products, as well as warning requirements applicable to HeatSticks® under FDA’s Marketing Order. • Third, applying the Proposed Rule to HeatSticks® would violate the First Amendment and raise potential concerns under the Takings Clause of the Fifth Amendment, and would therefore violate the APA for this reason as well. 1. FDA has failed to provide an adequate evidentiary basis for applying the Proposed Rule to HeatSticks® FDA’s notice of proposed rulemaking justifies the graphic warnings in the Proposed Rule only with reference to history, science, and evidence relating to traditional, combustible cigarettes. The Proposed Rule does not mention non-combustible products like HeatSticks®. Nor does it explain why the same health warnings that are appropriate for combustible cigarettes, and that each reference “tobacco smoke” or “smoking,” should also apply to non-combustible tobacco products. That evidentiary disconnect precludes FDA from applying the Proposed Rule to HeatSticks®. Importantly, none of the illnesses or conditions referenced in the graphic health warnings—i.e., “harm” to children from secondhand tobacco smoke; “stunt[ed] fetal growth”; age-related macular degeneration leading to blindness; bladder cancer; reduced blood flow to the limbs; cataracts leading to blindness; COPD; heart disease and strokes; head and neck cancer; type 2 diabetes; fatal lung disease in nonsmokers; or erectile dysfunction—has been causally linked to HeatSticks® used with the IQOS® device. Neither FDA nor the studies upon which FDA relies contain any indication that HeatSticks® causes those conditions; all the evidence concerns traditional cigarettes. FDA explains, for example, that it developed the new graphic warnings based on the existing “scientific literature”—in particular, the 2017 National Survey on Drug Use and Health (“NSDUH”) and the 2014 Surgeon General’s Report—and that this evidence shows that all conditions described in the warnings “have been established to have sufficient evidence to infer a causal link to cigarette smoking.”25 All of that only underscores that FDA’s science and evidence are specific to combustible cigarettes. FDA has provided no relevant data, evidence, or explanation that would justify extrapolating this evidence to non-combustible products like HeatSticks®. Nor has FDA examined the effect of the proposed graphic warnings in the context of products, such as HeatSticks®, that FDA determined should carry different government-mandated warnings.26 Those deficiencies not only counsel against applying the graphic-warnings requirements to HeatSticks® as a policy matter; they also would be fatal under the APA to any final rule that includes HeatSticks® within its scope.

25 84 Fed. Reg. at 42,754, 42,756; see also id. at 42,758. 26 See infra Section I.C.2.

9 2. Applying the Proposed Rule to HeatSticks® would contradict FDA’s findings and warning requirements in its Marketing Order FDA has previously concluded that HeatSticks® used with the IQOS® device are likely to present less health risk than traditional cigarettes, and that HeatSticks® should comply with different warning requirements. FDA has not explained the basis for departing from these prior findings in the Proposed Rule, and its failure to do so would render the application of the Proposed Rule to HeatSticks® invalid under the APA. The Marketing Order for IQOS® and HeatSticks® also sets forth other specific requirements for the warning labels on packaging and advertisements for HeatSticks®. For instance, all “HeatSticks packs” and “all kits containing HeatSticks packs” must include a nicotine warning: “WARNING: This product contains nicotine. Nicotine is an addictive chemical.”27 The required nicotine warning must “[b]e located in a conspicuous and prominent place on the two principal display panels of the package” and “comprise at least 30 percent of each of the principal display panels.”28 In addition, the warning must be printed “in a manner that contrasts by typography, layout, or color, with all other printed material on the package” and “centered in the warning area in which the text is required to be printed.”29 And all HeatSticks® advertisements must include a nicotine warning that “appear[s] in the upper portion of the area of the advertisement” and occupies “at least 20 percent of the area of the advertisement.”30 HeatSticks® are not, however, required to carry the carbon monoxide warning currently required for combustible cigarettes because, as FDA concluded, adding such a warning would be “misleading with respect to IQOS products.”31 FDA previously acknowledged that applying graphic warnings to these products would conflict with the terms of the Marketing Order, stating that “[w]hen FDA promulgates a final rule with respect to health warnings for cigarettes, FDA will reevaluate the conditions of marketing with respect to warnings for the products subject to this order.”32 Yet the Proposed Rule does not engage in any such “reevaluation.” Despite the differences outlined above, the Proposed Rule, if finalized, would subject HeatSticks® packaging and advertising to the same graphic warnings required of traditional cigarettes. FDA’s failure to explain the apparent conflict between its previous treatment of HeatSticks® and its proposed application of graphic warnings to those products creates two independent problems under the APA. First, FDA has not yet addressed its previous conclusions regarding the health risks presented by HeatSticks® used with the IQOS® device. The Proposed Rule notably does not mention FDA’s marketing authorization for IQOS® and HeatSticks®. To the extent FDA’s omission reflects the fact that FDA never intended to apply the proposed graphic warnings to HeatSticks®, FDA should clarify in any final rule that HeatSticks® are exempt from the warning requirements. But if FDA indeed intends to require HeatSticks® to bear the same graphic warnings as traditional

27 Attachment 2, FDA Marketing Order at 13. 28 Id. 29 Id. 30 Id. 31 Attachment 1, IQOS® TPL Review at 90; see Attachment 2, FDA Marketing Order at 14. The carbon monoxide warning for combustible cigarettes states “SURGEON GENERAL’S WARNING: Cigarette Smoke Contains Carbon Monoxide.” 32 Attachment 2, FDA Marketing Order at 13 n.9.

10 cigarettes, FDA must acknowledge its changed position, provide a detailed explanation for why the evidence now supports treating these products as if they have the same health risks as combustible cigarettes and why the evidence FDA just assessed through a rigorous multiyear process is no longer valid, and take into account the substantial reliance interests engendered by its marketing authorization for IQOS® and HeatSticks®.33 Without such explanation and evidence, subjecting HeatSticks® to any graphic-warnings requirements would violate the APA. Second, FDA’s proposal to apply the graphic-warnings requirements to HeatSticks® does not acknowledge an important aspect of the graphic-warnings issue as applied to HeatSticks®— namely, that subjecting HeatSticks® to the final rule would create serious inconsistencies between the graphic-warnings requirements and FDA’s Marketing Order for IQOS® and HeatSticks®. Under the Proposed Rule—which applies to “any roll of tobacco wrapped in paper or in any substance not containing tobacco”34—each HeatSticks® package would be required to include a graphic health warning comprising “the top 50 percent of the front and rear panels,” a location FDA characterizes as “prominent[]” in the Proposed Rule.35 Yet FDA has not explained how that requirement interacts with the Marketing Order, which provides that all packages of HeatSticks® must contain a nicotine warning located “in a prominent place” covering at least 30 percent of each of the two principal display panels.36 Nor has FDA addressed the fact that both the proposed graphic health warnings and the nicotine warning would be required to be presented “in a manner that contrasts . . . with all other printed material on the package.”37 With respect to HeatSticks® advertisements, moreover, FDA has not advised manufacturers on how to navigate the dueling requirements that the proposed graphic-health warnings and required nicotine warning, respectively, “comprise at least 20 percent of the area . . . at the top of the advertisement” and “at least 20 percent” of the “upper portion of the area of the advertisement.” FDA’s failure to explain whether or how the graphic-warnings requirements apply in light of conflicting warnings requirements in the Marketing Order would also generate implementation problems. PM USA could be put to the choice of complying either with the IQOS® and HeatSticks® Marketing Order or the graphic warnings requirements. Further, under the Proposed Rule, cigarette manufacturers must submit proposed rotational warning plans to the FDA no later than 5 months after FDA issues any final rule; manufacturers would have to implement the proposed warnings within 15 months of a final rule.38 PM USA, however, cannot prepare, submit, and obtain approval of a rotational warning plan for HeatSticks® until the FDA decides whether to “reevaluate” or change the warning requirements for HeatSticks®, which may require a significant amount of time. This puts the company at significant risk of missing the implementation deadline and, consequently, being unable to market products that FDA has determined to be appropriate for the protection of public health. To guard against this unfair result and comply with its duties under the APA, FDA should acknowledge and address this important problem of conflicting warning requirements.

33 See FCC v. Fox Television Stations, Inc., 556 U.S. 502, 515 (2009). 34 15 U.S.C. § 1332(1)(A); see 84 Fed. Reg. at 42,796. 35 84 Fed. Reg. at 42,755. 36 Attachment 2, FDA Marketing Order at 9. 37 Compare id., with 84 Fed. Reg. at 42,781. 38 See 84 Fed. Reg. at 42,784 (proposing 15-month effective date).

11 3. Applying the Proposed Rule to HeatSticks® would violate the First Amendment and raise Takings Clause concerns Any final rule that applies the proposed warnings to HeatSticks® would also violate the First Amendment and raise serious concerns under the Takings Clause of the Fifth Amendment. First, while the entire Proposed Rule contravenes the First Amendment, see infra Section IV, requiring HeatSticks® packaging and advertisements to include the proposed text and graphic warnings would present an especially clear First Amendment violation. The First Amendment’s guarantee of free speech includes the right not to speak—a right that extends to individuals and commercial entities alike. The government may compel commercial disclosures only under two circumstances. Under the so-called Zauderer framework, the government must show that the disclosure (i) contains “purely factual and uncontroversial information” and (ii) is not “unjustified or unduly burdensome.”39 If a disclosure requirement does not satisfy those requirements, then the government must demonstrate, at the very least, that the requirement “directly and materially advances a substantial state interest in a manner no more extensive than necessary to serve that interest”—in other words, the government must satisfy the intermediate- scrutiny standard of review.40 FDA faces particular hurdles in satisfying either of these tests so as to apply the proposed graphic warnings to HeatSticks®. Second, applying the Proposed Rule to HeatSticks® could severely diminish the value of the product and its trademarks and trade dress. That result would raise concerns under the Takings Clause of the Fifth Amendment. a. The Proposed Rule violates the First Amendment as applied to HeatSticks® Applying the rule to HeatSticks® would raise two particularly severe First Amendment problems that FDA’s Proposed Rule overlooks. First, the requirements that 50% of both the front and rear panels of every HeatSticks® package and 20% of every HeatSticks® advertisement feature “prominently” placed graphic warnings, when considered with the nicotine warning requirements that FDA already mandated, restrict too much speech to pass constitutional muster. Under the Marketing Order (which reflects federal law), 30% of the front and rear panels of every HeatSticks® package and 20% of every HeatSticks® advertisement must already convey a nicotine warning.41 By adding the graphic- warnings requirements, the Proposed Rule would dedicate a combined 80% of the front and rear panels of HeatSticks® packaging and 40% of HeatSticks® advertisements to the government’s anti-smoking message. That does not even account for other necessary elements of HeatSticks® packaging, such as UPC codes. FDA acknowledged the potentially problematic interaction of the warnings in the Marketing Order, and promised to “reevaluate the conditions of marketing with respect to warnings” for these products.42 But the Proposed Rule neither “reevaluate[s]” the HeatSticks® marketing requirements nor acknowledges the extraordinary limitation on PM USA’s ability to “call

39 NIFLA, 138 S. Ct. at 2372 (quoting Zauderer v. Office of Disciplinary Counsel of Sup. Ct. of Ohio, 471 U.S. 626, 651 (1985)). 40 Ibanez v. Fla. Dep’t of Bus. & Prof’l Regulation, Bd. of Accountancy, 512 U.S. 136, 142 (1994). 41 See Attachment 2, FDA Marketing Order at 13. 42 Id. at 13 n.9.

12 attention to . . . its own message” that would result from confiscating 80% of HeatSticks® packaging and 40% of HeatSticks® advertisements.43 Courts have repeatedly struck down less- onerous disclosure requirements as “unduly burdensome” of commercial speech.44 As with those requirements, the Proposed Rule’s unprecedented restrictions on the message PM USA could convey via HeatSticks® packaging and advertisements would be unduly burdensome and lack any adequate justification, and thus invalid under the applicable First Amendment standards. Second, FDA’s asserted interest in “promoting greater public understanding of the negative health consequences of smoking” does not justify applying any of the textual or graphic warnings to HeatSticks®, which do not combust or produce smoke. As noted, requiring graphic warnings for health risks related to smoking combustible cigarettes, and requiring text that expressly refers to “tobacco smoke” or “smoking,” would plainly mislead consumers as to the health risks of these products. As such, the proposed graphic warnings would not be “factual and accurate”45 as applied to HeatSticks®. For that reason, the warnings cannot be characterized as allowable disclosure requirements for HeatSticks® under the Zauderer framework. Nor could FDA justify applying the Proposed Rule to HeatSticks® under more rigorous standards of scrutiny, as the government does not have a legitimate interest in forcing commercial entities to engage in misleading speech. b. Applying the Proposed Rule to HeatSticks® would raise Takings Clause concerns By reserving 80% of the front and rear panels of HeatSticks® and nearly half of each HeatSticks® advertisement for government warnings, applying the Proposed Rule to HeatSticks® would also raise concerns under the Takings Clause of the Fifth Amendment. The Takings Clause restricts the government’s ability to promulgate “burdensome regulations” that “impede[] the use of property” even “without depriving the owner of all economically beneficial use.”46 That inquiry looks, for instance, into whether a regulation imposes a severe economic impact on a regulated party and the extent to which the regulation interferes with a party’s reasonable, investment- backed expectations regarding the value of the property.47 Those considerations underscore the harsh consequences of applying the Proposed Rule to HeatSticks®. The rule would impede PM USA’s ability to make use of its HeatSticks® trademarks and trade dress by “press[ing] into . . . public service” the vast majority of the packaging area on which PM USA could otherwise advertise to consumers.48 Further, requiring packages and advertisements for HeatSticks® to contain the mandated warnings would contradict FDA’s prior findings and important health-policy goals. FDA’s analysis of the science and evidence found that switching to the IQOS® device and HeatSticks® is “likely to lead to less risk of tobacco-related diseases” compared to traditional cigarettes.49 Yet subjecting HeatSticks® to the same or greater warning regime would signal that HeatSticks® used with the IQOS® device

43 NIFLA, 138 S. Ct. at 2378. 44 See, e.g., id. (invalidating statute requiring pregnancy centers to include a 29-word, “government-drafted statement” on “all print and digital advertising materials”); Am. Beverage Ass’n v. City & Cty. of San Francisco, 916 F.3d 749, 756-57 (9th Cir. 2019) (en banc) (invalidating warning that would have occupied 20% of beverage labels). 45 84 Fed. Reg. at 42,757. 46 Murr v. Wisconsin, 137 S. Ct. 1933, 1939, 1943 (2017). 47 Id. at 1943. 48 Lucas v. South Carolina Coastal Council, 505 U.S. 1003, 1018 (1992). 49 Attachment 1, IQOS® TPL Review at 11.

13 present the same or greater risks. Applying the graphic-warnings requirements to HeatSticks® would thus threaten to unfairly deprive PM USA of the significant and reasonable (in light of FDA’s Marketing Order and stated policy of transitioning current smokers to products lower on the continuum of risk) investments it has made to bring them to market. In sum, based on the weighty public health concerns and legal issues outlined above, we respectfully request that FDA exempt HeatSticks® from any final graphic-warnings rule the Agency issues.

II. FDA’S PROPOSED RULE RAISES A NUMBER OF PRACTICAL CONCERNS IMPACTING THE IMPLEMENTATION OF THE REQUIRED WARNINGS ON CIGARETTE PACKAGES AND ADVERTISEMENTS Finalizing the Proposed Rule in its current form would create significant implementation problems for regulated industry and FDA that, if left unaddressed, would hinder the ability of the regulated industry to faithfully and efficiently comply with the requirements of the rule. In its prior rulemaking regarding required warnings for cigarettes,50 FDA responded to many practical implementation concerns raised by industry. We encourage FDA to do no less here. The Agency should begin engaging with industry now, so that practical implementation issues with the Proposed Rule can be resolved in a time and manner that will help industry meet its deadline to submit rotational warning plans and achieve compliance by any final rule’s effective date. FDA should alter the Proposed Rule in several ways, as explained below. • First, as a threshold matter, FDA should reduce the number of required graphic warnings from 13 to either 12 or 9. This modification is necessary due to extremely difficult and potentially insurmountable technical challenges with printing and distributing 13 warnings equally and randomly on cigarette packages.

• Second, FDA should address other issues impacting the implementation of the required warnings on cigarette packages, such as allowing manufacturers to include two different warnings on the front and rear panels of the package, permitting them to adapt the warnings to fit packages of varying sizes and shapes, clarifying how to incorporate the warnings on hard pack and soft pack configurations, and providing additional information to facilitate the implementation of, and meaningful comment on, the required warnings.

• Third, FDA should address issues impacting the implementation of the required warnings on cigarette advertisements, including the need for clear direction regarding digital formats and either exemptions or alternative fonts and images for small-sized advertisements.

• Finally, to account for the interdependent chain of events necessary to implement graphic health warnings, FDA should extend or toll the implementation deadline.

50 Required Warnings for Cigarette Packages and Advertisements, 76 Fed. Reg. 36,628, 36,690-94. (June 22, 2011).

14 A. FDA should reduce the number of required graphic warnings to 12 or 9 The Proposed Rule requires manufacturers to randomly display 13 graphic images, each associated with one of 12 textual warning label statements, on their cigarette packages and advertisements.51 FDA invites comment on “whether and how the number of final required warnings selected would affect the random and equal display and distribution of the required warnings on packages and the quarterly rotation of the required warnings in advertisements.”52 As explained below, manufacturers cannot satisfy the “random and equal” requirement for 13 different warnings without drastic and uncertain changes to packaging production.53 Thirteen is both an odd and prime number. As such, it is fundamentally incompatible with industry-wide printing practices and the random and equal distribution requirements of proposed § 1141.10(g) for cigarette packages. We urge FDA to reduce the number of graphic warnings to 12 or 9. 1. Manufacturers cannot randomly and equally display and distribute 13 warnings on cigarette packages. Standard industry practices print graphics in grid layouts. When package designers plan a print run, they arrange the desired package graphics in a grid based on the package style and type of printing press used.54 Each rotation of the press cylinder prints the grid pattern onto packaging materials.55 The grid pattern depends on the technical capabilities of the printing press, which are fixed.56 We use a handful of grid layouts, depending on type of press and package being produced.57 Standard grid layouts cannot accommodate a requirement to print equal quantities of 13 warnings. Our 32-inch press for hard packs illustrates this conundrum.58 This press prints in an 8x3 layout, resulting in 24 total impressions (Figure 1). Figure 1: 32-Inch Layout with 13 Proposed Warnings Cylinder Layout– 8 lanes 3 Around (Number represents Proposed Warning #) 1 4 7 10 13 3 6 9

2 5 8 11 1 4 7 10

3 6 9 12 2 5 8 11

51 84 Fed. Reg. at 42,797. 52 Id. at 42,784 (emphasis added). 53 See Attachment 3, Declaration of Yvonne DeVerry (“DeVerry Decl.”) ¶ 13. 54 Id. ¶ 14. 55 Id. 56 Id. ¶ 15. 57 PM USA uses two layouts for hard pack printing presses. The 26-inch presses are arranged with six lanes and 3 impressions per lane (6x3). The 32-inch presses print 8 lanes with 3 impressions per lane (8x3). We currently use a 2x 2 and a 1x2 layout for cartons. We currently use 9x5 and 8x5 layouts for soft pack presses. 58 Similar outcomes result from all layouts used. The 8x3 layout is being used as an illustrative example. Attachment 3, DeVerry Decl. ¶ 18.

15 In this layout, no arrangement of 13 warnings would result in printing an equal number of each warning. The first 11 warnings appear twice, but proposed warnings #12 and #13 appear only once.59 On a commercial scale, this difference would result in significantly unequal numbers of warnings. In fact, no industry standard grid formation has the capability of printing 13 warnings in equal quantities.60 Thirteen warnings simply cannot be arranged in a 6x3, 8x3, 1x2, 2x2, 9x5, 8x5 or 8x6 grid. We are exploring alternate ways of satisfying a 13-warning requirement at different stages of the packaging production process. Any of these alternatives, even assuming they could be implemented, would be orders of magnitude more complex and expensive than randomly and equally distributing 12 or 9 warnings.61 For example, we are exploring the feasibility of a process to print blanks in the layout, cut them out, and discard them. This approach, even if feasible, would substantially increase waste and slow production. We are also considering if it would be feasible to manually insert a stack of printed packages bearing only the 13th warning on top of a pallet of printed packages bearing warnings 1 through 12. Critically, both of these techniques are unproven, especially at the commercial scale.62 Whether or not these or other untested production changes are technically reliable, commercially scalable, and could ultimately lead to random and equal distribution throughout the United States over a 12-month period has not been determined.63 Moreover, these approaches would substantially increase costs to manufacturers. We estimate that the costs of materials alone could increase by a minimum of 40 percent and possibly exceed 165 percent, if FDA requires 13 warnings instead of a number of warnings that does not create these problems.64 2. Manufacturers can randomly and equally display and distribute 12 or 9 warnings on cigarette packages. Manufacturers can more easily meet the “random and equal” requirement for either 12 or 9 warnings. The process for randomly and equally distributing 12 or 9 warnings is more straightforward using standard printing practices, which will increase the ability of manufacturers to comply.65 For example, using the same 8x3 layout described above, packaging designers would arrange 12 warnings as shown in the following layout (Figure 2).

59 The same challenges associated with implementing 13 warnings would also apply to 5, 7, or 11 warnings. Like 13, 5, 7 and 11 are also prime numbers that cannot be arranged in a grid pattern that could be produced in equal numbers without potentially drastic changes to printing and manufacturing processes. 60 Attachment 3, DeVerry Decl. ¶ 13. 61 Id. ¶¶12-13. 62 Id. ¶¶ 20-22. 63 Id. ¶ 21. 64 Id. ¶ 22. 65 Id. ¶ 24.

16 Figure 2: 32-Inch Layout with 12 Proposed Warnings Cylinder Layout– 8 lanes 3 Around (Number represents Proposed Warning #) 1 4 7 10 1 4 7 10

2 5 8 11 2 5 8 11

3 6 9 12 3 6 9 12

Here, each rotation of the cylinder would print 2 impressions of each warning, producing an equal number of all 12 health warnings. This example is illustrative of the similar capabilities of our 6x3, 2x2 and 1x2 presses. Additionally, compliance for soft pack presses could be achieved by using an 8x6 layout. The process for randomly and equally distributing 9 warnings is similarly straightforward using standard printing practices. This process would rely on 3 cylinders. Using the same 8x3 layout as an example, packaging designers would arrange the warnings as shown in the following layouts using 3 cylinders (Figure 3). Figure 3: Layouts Using 3 Cylinders Cylinder Layout– 8 lanes 3 Around Set 1 (Number represents Proposed Warning #) 1 4 7 1 4 7 1 4

2 5 8 2 5 8 2 5

3 6 9 3 6 9 3 6

Cylinder Layout– 8 lanes 3 Around Set 2

1 4 7 1 4 7 1 7

2 5 8 2 5 8 2 8

3 6 9 3 6 9 3 9

Cylinder Layout– 8 lanes 3 Around Set 3

1 4 7 1 4 7 4 7

2 5 8 2 5 8 5 8

3 6 9 3 6 9 6 9

Using multiple cylinders, however, would not provide a solution if FDA requires 13 warnings, because there would always be a warning left off the layout, or blank spaces in the layout that

17 must be somehow cut out and discarded, again resulting in significant waste, delay, and added cost. When the printer can produce an equal number of warnings, as with the 12- and 9-warning layouts described above, randomization of the cigarette warnings occurs through normal manufacturing and distribution practices.66 For example, boxes and cases are mixed through normal distribution to the wholesale warehouses. Carton randomization occurs when the pallets and cases are broken down for orders and shipped to vendors. Pack randomization occurs when packs are removed from cartons and sold in point of sale fixtures. To assure the ability of manufacturers to comply with the “equal and random” warning requirement, FDA should reduce the number of graphic warnings to 12 or 9. 3. If FDA requires 13 warnings, it should clarify the “as is possible” limitation on equal display Proposed §1141.10(g)(1) requires that warnings for packages “be randomly displayed in each 12-month period, in as equal a number of times as is possible on each brand of the product and be randomly distributed in all areas of the United States in which the product is marketed…”67 The phrase “as is possible” is undefined and unclear. As explained above, it would be “possible” for manufacturers to produce and randomly and equally distribute 12 or 9 warnings within a small margin of error. On the other hand, if any final rule were to require 13 warnings, it would not be “possible” for manufacturers to randomly and equally distribute 13 warnings without radical, untested, and costly changes to industry practices. Simply stated, FDA should require either a) random and equal distribution of 12 or 9 warnings or b) random but unequal distribution of 13 warnings. The first option would be much easier, for both FDA and manufacturers. If, on the other hand, FDA chooses the second option to require 13 warnings, it should clearly define the phrase “as is possible,” so that manufacturers will know what is expected of them. FDA’s definition should allow a substantial margin of tolerance to recognize that some warnings are likely to appear significantly more frequently than others throughout the U.S. during the 12-month period.68 It will also be important for FDA to apply this margin of tolerance consistently and equitably across regulated industry, which will compound the Agency’s administrative and enforcement burden. 4. Any final rule should make accommodations for low volume products Cigarette products with very low annual sales volume may be unable to meet the Proposed Rule’s “random and equal distribution” requirement over a 12-month period. Achieving random and equal distribution may take several years for such products. For example, some of our soft pack brands sell approximately 2 million cigarettes per year. The packaging for these cigarettes would require approximately 5 rolls of printed materials. However, it takes 4 rolls of packaging to produce an equal number of 12 warnings, and 3 rolls to

66 Id. ¶ 23. 67 84 Fed. Reg. at 42,797. 68 Attachment 3, DeVerry Decl. ¶ 19.

18 produce an equal number of 9 warnings. The additional 1 or 2 rolls needed to package these cigarettes would cause unequal distribution of warnings within a 12-month period. The situation would be even more pronounced for 13 warnings in low volume soft packs. As an example, it would take 13 rolls to produce an equal number of 13 health warnings of soft packs, with 1 warning per roll and no random distribution. Under this example, which would not satisfy the random and equal requirement of the Proposed Rule, it would take 2.5 years for all 13 warnings to enter the marketplace in equal numbers. Any final rule should make accommodations for low volume items. One hard pack pallet is enough packing for approximately 5 million cigarettes. We estimate that it would take 3 pallets worth of hard pack packaging, or about 15 million cigarettes, to achieve random and equal distribution for 12 or 9 warnings. For soft pack, we estimate it would take about 4 rolls to achieve random and equal distribution for 12 or 9 warnings, or again approximately 15 million cigarettes. For the reasons described above, 15 million cigarettes and below should be considered low volume products not subject to the random and equal requirements of the Proposed Rule. B. FDA should address other issues impacting the implementation of the required warnings on cigarette packages 1. FDA should clarify in any final rule that each package may have two different graphic health warnings on the front and rear package panels The Proposed Rule is silent on whether manufacturers can include one graphic warning on the front of a cigarette package and a different graphic warning on the back. Allowing manufacturers to place a different graphic health warning on the front and back of each package has public health policy benefits, such as increasing distribution of 12 warnings and increasing the variety of warnings at retail at a given time. This arrangement would decrease production costs and facilitate equal production,69 though it would not overcome the problems presented by requiring 13 warnings.70 2. FDA should allow manufacturers to make minor graphic and design adaptations, so that warnings are producible on a variety of package shapes and sizes Under the Proposed Rule, manufacturers must ensure that the required warnings “are not distorted, obscured, or otherwise inaccurately reproduced from the incorporated material when reproduced for use in differing types of media…”71 Because cigarette packages come in a variety of shapes and sizes, the Proposed Rule should allow manufacturers to adapt the required warnings to fit packages of all sizes and shapes. Reference 1872 of the Proposed Rule provides illustrations of each warning statement, depicting how it would appear on a single “Cigarette Package.” These illustrations do not account for the

69 For example, on a 6x3 grid with 18 impressions, if manufacturers could print 2 warnings per pack, that would mean 36 total impressions (18+18). Because 36 is a multiple of 12, manufacturers would print an equal number of warnings with each run. 70 Attachment 3, DeVerry Decl. ¶ 25. 71 84 Fed. Reg. at 42,781. 72 Reference or Ref. throughout these comments refers to the Agency’s References. See 84 Fed. Reg. at 42,789- 42,796.

19 variety of packages in the market.73 PM USA currently markets 14 pack and 14 carton sizes. Nat Sherman produces 3 pack and 3 carton sizes. Some packages are wide and short, others are tall and narrow, as shown in Attachment 3, DeVerry Declaration, Exhibit A. To compress or expand a warning to cover the required 50 percent of each of these packages, PM USA and Nat Sherman would need to adjust the width-to-height ratio (i.e., aspect ratio), which would stretch or elongate the required textual and graphic warning.74 The Proposed Rule, however, prohibits distorting the warnings.75 FDA should resolve this issue in one of the following four ways. First, FDA could provide high-resolution art and vector elements with instructions for use for how to re-crop each warning to fit the top 50 percent of each package size and shape. Here, FDA should provide a variety of aspect ratios including 1:1, 1.5:1, 1:2, 2:1, and 2.5:1. Second, FDA could authorize manufacturers to maintain the aspect ratio of the warning while using border area to reach the 50 percent area requirement. FDA could provide warnings with enough continuous tone graphic images in the background so that the graphics could be cropped rather than stretched, as shown in Attachment 3, DeVerry Declaration, Exhibit C. Third, the Agency could provide differently-shaped warnings specifically designed to cover 50 percent of each package configuration submitted by manufacturers to FDA. Lastly, FDA could allow manufacturers to change the aspect ratios to make them fit the required area. Regardless of the approach taken, FDA should provide specific direction for permissible adjustments to the required textual and graphic warnings and tolerate minor variances in how the warnings appear on a manufacturer’s packaging. 3. FDA should clarify how manufacturers should incorporate the warnings on box cigarette packaging with hinged lids The Proposed Rule requires that the “warning comprise at least the top 50 percent of the front and rear panels of the package.”76 Cigarette packages come in both hard pack and soft pack configurations. Printing on the top 50% of a hard pack, which includes the location of the hinged top that opens and closes to provide consumers access to the contents of the package, presents unique considerations. In hard pack cigarette packaging, content printed on the hinged lid may shift up or down by up to one millimeter at the point where the lid meets the front of the pack due to normal variations in the manufacturing process.77 This minor variation would likely be more noticeable when text is printed across the hinged lid and the pack front.78 To address this issue, FDA could design the warnings so that the text is located entirely above or entirely

73 See FDA Proposed Regulation: Required Warning For Cigarette Packages and Advertising (Reference 18) FDA– 2019–N–3065 (Aug. 2019), available at https://www.regulations.gov/document?D=FDA-2019-N-3065-0020. 74 Attachment 3, DeVerry Decl. ¶¶ 29-30. 75 84 Fed. Reg. at 42,781. 76 Id. 77 Attachment 3, DeVerry Decl. ¶ 32. 78 Id.

20 below the point where the hinged lid meets the pack for each of the various hard pack sizes we produce. Alternatively, FDA could provide font suitcases and instructions for use that allow manufacturers to flow text freely within a designated text area to ensure the text is not interrupted by the package features. This would allow manufacturers to adjust the location of the text to prevent individual words from being split by the package features. For example, in Figure 4 below FDA could allow manufacturers to include the words “a lung disease” or “that can be” on different lines of text depending on the package shape and size. Figure 4: Example of Impermissible Text Splitting

If FDA does not allow manufacturers to make such adjustments, it should at least allow for unavoidable but minor manufacturing variation in box cigarette packaging with hinged lids. Normal variation in the manufacturing processes may cause a small gap in the hinged lid, approximately 1.5-millimeters in size, where the lid meets the front of the package (as shown in Figure 5).79 This effect cannot be overcome through manufacturing practices and could result in variation of warning appearance.80

79 Id. ¶ 33. 80 Id.

21 Figure 5: Example of Hard Pack with Fully Closed Hinged Lid and with Small Gap

4. FDA should allow manufacturers to position warnings below soft pack closures. Soft-pack configurations pose a different consideration. During the manufacturing process, after the cigarettes are inserted into the package, the top of a cigarette soft pack is folded down and held down by an adhesive closure.81 This adhesive closure is applied after the packages have been printed. Without any accommodation, that closure would obstruct a portion of the required warning (see Figure 6). In FDA’s prior graphic health warnings rulemaking, the Agency permitted manufacturers to “adapt the warnings on ‘soft pack’ style packaging by moving the warning below the closure.”82 There, FDA allowed the upper boundary of the display area to be up to 0.375 inches from the top edge of the package. The previously stated boundary of 0.375 inches, however, is too small to ensure that there is enough adhesive for the package to remain closed while also accounting for standard printing variation. In any final rule, FDA should allow the closure to extend up to 0.482 inches from the top edge of the package (see Figure 6). This accommodation will help ensure that manufacturers with soft pack packaging are able to comply with any final rule.83

81 Id. ¶ 35. 82 76 Fed. Reg. at 36,691. 83 Attachment 3, DeVerry Decl. ¶ 36.

22 Figure 6: Target Label and Closure Placement with Variation84

5. FDA should provide “Printer’s Proofs” for each warning to facilitate consistent reproduction of the color images Warnings should have a consistent appearance throughout the marketplace. Manufacturers, however, do not calibrate their equipment with each other and use different ink application techniques and substrates.85 This could result in a vast array of packaging appearances from manufacturer to manufacturer with regards to the colors used for graphic images (see Figure 7). FDA should provide “printer’s proofs” to manufacturers at the time any final rule is issued.86 This will help ensure that each manufacturer can properly calibrate its equipment and materials to accurately reproduce the required warnings and promote consistency across manufacturers.

84 Based on description of soft pack implementation with closure in Required Warnings for Cigarette Packages and Advertisements, 76 Fed. Reg. at 36,628 (June 22, 2011). 85 Attachment 3, DeVerry Decl. ¶ 38. 86 Id. ¶ 37.

23 Figure 7: Examples of Potential Inconsistencies in Appearance of Proposed Warnings

6. FDA should disclose additional information to facilitate meaningful comment on FDA’s proposal The Proposed Rule includes a PDF containing images of the proposed warnings. The Proposed Rule promises that “[a]ny final rule would provide information on how to obtain the final electronic, layered design files for each required warning, as well as technical specifications to help regulated entities appropriately select, crop, and scale the warnings to ensure the required warnings are accurately reproduced across various sizes and shapes of cigarette packages and cigarette advertisements.”87 FDA should immediately provide these materials to interested parties so they may quickly identify any potential implementation issues prior to finalization of the rule. We submitted a formal request to FDA explaining the importance of FDA releasing these files for public review at the Proposed Rule stage.88 C. FDA should address issues impacting the implementation of the required warnings on cigarette advertisements 1. FDA should establish clear direction for the display of warnings in digital advertisements Section 1141.10(d)(1) of the Proposed Rule requires that “for print advertisements and other advertisements with a visual component (including, for example, advertisements on signs, retail displays, internet web pages, digital platforms, mobile applications, and email correspondence), the required warning must appear directly on the advertisement.”89 Without further clarification, compliance with the requirements for digital advertisements will be an exercise in guesswork. Consumers may access and view the same advertisement across a range of electronic screens (e.g., computers, cellular phones, tablets, etc.) with different sizes and formats. These screens may display the same advertisement differently based on the device or software. The Proposed Rule does not address this reality at all. To prevent unnecessary complications with compliance, FDA should evaluate existing platforms and provide specific direction on how to display the required warnings based on specific devices and software prior to finalizing this rule.

87 84 Fed. Reg. at 42,757. 88 Docket No. FDA-2019-N-3065-0333 (84 Fed. Reg. 42,754, Aug. 16, 2019), Notice of Proposed Rulemaking on Tobacco Products; Required Warnings for Cigarette Packages and Advertisements - Request for Extension of Public Comment Period from Client Services LLC (Sept. 5, 2019), available at https://www.regulations.gov/document?D=FDA-2019-N-3065-0333 (“Sept. 5, 2019. (“Sept. 5, 2019 Request for Extension”). 89 84 Fed. Reg. at 42,797.

24 In particular, FDA should establish clear direction regarding how to comply with Section 1141.10(d)(2)’s “prominent and conspicuous” requirements in the context of digital advertisements (e.g., websites, e-mail, digital banner ads). When FDA finalizes the rule, FDA should allow manufacturers to comply with the “prominent and conspicuous” requirement by making the required warning initially visible on the screen in its entirety and permitting the consumer to scroll away from the warning to access advertising content. Digital advertising requires image-based code deployment, which can present challenges such as poor site load, wi-fi access, or cellular service that could cause a delay in the loading of images. To help address these challenges, FDA should provide the text portion of the warning within the image files it provides to manufacturers to help with implementation of the graphic warnings in digital advertising. In addition, any final rule should provide manufacturers with the flexibility to customize their advertisements based on the type of technology being used and to help them adapt to constantly evolving technologies. FDA should, therefore, provide appropriately sized digital images for use with multiple devices and software system combinations. 2. FDA should exempt or establish alternative requirements for small-sized advertisements FDA should address required warnings for small-sized advertisements (such as small point-of- sale materials and certain digital media). Some point-of-sale materials, for example, may be less than 36 square inches, which could result in warnings that are illegible or distorted. FDA should exempt small-sized advertisements from the graphic warning requirements, or in the alternative, revise the minimum font requirement and use an appropriate graphic image specifically designed for small formats. D. FDA should extend or toll the 15-month implementation period to allocate a reasonable time for manufacturers to comply with any final rule The Proposed Rule requires each cigarette manufacturer to submit a rotational warning plan to FDA “no later than 5 months from the date of the publication of any final rule.”90 FDA’s Proposed Rule fails to consider many of the logistical and practical challenges that manufacturers will face following issuance of any final rule. The implementation of graphic health warnings is the end result of a chain of events that must occur in sequence. • First, FDA must issue a final rule so manufacturers can know the relevant requirements; • Second, manufacturers must develop packaging and advertising that complies with the requirements of any final rule; • Third, manufacturers must prepare and submit rotational warning plans; • Fourth, FDA must review and approve rotational warning plans; • Fifth, manufacturers must procure packaging materials printed with the warnings as per the approved warning plans; and

90 84 Fed. Reg. at 42,784.

25 • Finally, manufacturers must produce compliant cigarette-containing packages and deploy compliant advertisements by the 15-month effective date. While there may be some work that can overlap with other steps in the sequence, most of the work is iterative and requires completion of the prior steps before it can begin. Manufacturers cannot secure all necessary packaging materials, which are typically sourced from third parties, to implement graphic health warnings until after FDA approves their rotational warning plans. Otherwise, manufacturers risk producing noncompliant packaging that must be destroyed, resulting in a significant waste of time, resources, and expense. In addition, given the large number of manufacturers that will submit rotational warning plans, FDA’s review and approval of those plans will likely take several months. This is problematic because the Proposed Rule does not toll the 15-month implementation period pending FDA’s review and approval of the rotational warning plans and does not commit FDA to review a plan within any certain time period. FDA’s review and approval will eat into the implementation period for complying with the warning requirements and potentially does not leave sufficient time for work that must occur after approval but prior to implementation. This intricate process is resource-intensive, from beginning to end. Building a supply line takes time. In the absence of an approved plan, any orders for materials will come with significant risk. Under ordinary circumstances, it takes an average of 12 weeks to obtain such materials. Implementing graphic health warnings, however is not an ordinary circumstance. Manufacturers will have to develop and produce a high volume of compliant packages concurrently and within a compressed time period. Altogether, we estimate that this process will take approximately 67 weeks, exclusive of FDA review and approval. Manufacturers, particularly large manufacturers with hundreds of different products, will need the full 15 months—and possibly longer—to fully implement graphic health warnings on packaging. Indeed, to stand a chance of meeting this deadline, a manufacturer would have to submit its rotational warning plan immediately after any final rule is published. This simply is not realistic. To allow manufacturers a reasonable time to comply, FDA should extend the 15-month implementation period, toll it during FDA’s review and approval of warning plans, or both. The objective of this modification is to account more realistically for the interdependent chain of events that will be necessary for manufacturers to implement graphic health warnings, including but not limited to, waiting for FDA approval of rotational warning plans. Under the Proposed Rule, FDA has no deadline for reviewing warning rotation plans. Indeed, it is under no obligation at all to approve warning rotation plans. As a result, manufacturers who submit plans in a timely manner may not receive FDA approval prior to the implementation deadline, making it impossible for them to produce product packages and advertisements that comply with the law. In the alternative, FDA should exercise its enforcement discretion to allow products to remain on the market while a manufacturer’s rotational warning plan is under its review, as it did when smokeless rotational warning plans were under review.91

91 Regardless of how FDA responds to our comments on the implementation period, any final rule cannot constrain the ability of distributors and retailers to sell through their lawfully acquired cigarettes. Congress imposed the effective date of the new warnings on the date of manufacture, not sale. TCA § 201(b). Yet, FDA states that

26 III. IF IMPLEMENTED AS PROPOSED, THE RULE WOULD VIOLATE THE ADMINISTRATIVE PROCEDURE ACT If FDA proceeds to a final rule without substantial modifications, the rule would violate the APA for numerous reasons. The APA directs courts to set aside agency action, findings, and conclusions that exceed the agency’s authority, fail to comply with statutory requirements, or are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.”92 The APA requires agencies to rest their findings and conclusions on “reasoned analysis,” and requires invalidation of rules that for instance fail to consider key aspects of the problem the agency is trying to solve, fail to support findings with substantial evidence, or ignore alternative approaches or contrary evidence.93 FDA’s evaluation of the scientific evidence includes numerous explanatory and evidentiary gaps. Left unaddressed, these gaps would render any final rule invalid, because the APA requires FDA to support its conclusions with substantial scientific evidence and to connect its determinations to record evidence. • First, FDA has not provided sufficient information regarding its referenced research studies to allow for meaningful evaluation of the validity of those studies, in violation of its statutory obligations under the TCA and general principles of administrative law. Nor has FDA provided sufficient information for stakeholders to determine why it decided to highlight certain diseases over others in graphic warnings; why FDA selected the graphics it did; or how and why FDA determined that the chosen textual language is comprehensible. • Second, FDA relies upon fundamentally unreliable scientific evidence that does not adequately support the Proposed Rule. FDA’s core scientific evidence consists of commissioned studies that are seriously flawed in their design. FDA’s other cited studies are inapt, not least because they rely on data from non-United States populations without justification and in disregard of FDA’s own scientific standards. These studies do not provide meaningful insight into how U.S. consumers would respond to the proposed warnings and contain other methodological defects. • Third, even were FDA’s studies reliable, their results do not support FDA’s conclusion that the proposed graphic warnings would be effective in improving comprehension. • Fourth, FDA does not adequately address the science and evidence that runs contrary to its proposal.

“[w]hile this statutory limitation applies to only manufacturers, FDA believes that keeping products without the required warnings under any final rule on the market for an extended period would not be in the interest of public health.” 84 Fed. Reg. at 42,785. Putting aside the fact that FDA does not define “extended period,” which prevents meaningful comment, Congress did not restrict the ability of wholesalers and retailers to sell through their inventory of cigarettes with the prior FCLAA warnings timely manufactured and distributed by manufacturers. TCA § 201(b). Any such restraint would unfairly deprive them of the benefit of their investment and be contrary to law. 92 5 U.S.C. § 706(2). 93 Motor Vehicle Mfrs. Ass’n of the U.S. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 42-43 (1983); Am. Gas Ass’n v. FERC, 593 F.3d 14, 19-20 (D.C. Cir. 2010). See also, e.g., Chamber of Commerce v. SEC, 412 F.3d 133, 145 (D.C. Cir. 2005); Laclede Gas Co. v. FERC, 873 F.2d 1494, 1498 (D.C. Cir. 1989); Yakima Valley Cablevision, Inc. v. FCC, 794 F.2d 737, 746-47 (D.C. Cir. 1986).

27 • Fifth, FDA fails to consider important aspects of the graphic-warnings issue, including whether alternative means of communicating information would satisfy its asserted consumer-education goal in a way that is less burdensome on manufacturer speech. • Sixth, FDA adduces no evidence whatsoever in support of its proposal to include two different graphic warnings as accompaniments to its textual warning concerning COPD.

A. FDA’s Proposed Rule fails to provide the required and necessary underlying study information to allow stakeholders to evaluate its selection of the proposed warnings The particular graphic health warnings selected by FDA in the Proposed Rule were not published or released prior to the publication of the Proposed Rule. FDA grounds its primary scientific justification for its proposed warnings in just two quantitative research studies. See OMB Control No. 0910-9866; OMB Control No. 0910-0848. Indeed, the only studies examining the actual warnings FDA proposed are FDA’s own two studies. To our knowledge, these studies included an initial survey that screened 24 textual warnings (15 revised textual warnings and 9 warnings proposed in the TCA). FDA then selected 12 of those textual warnings and paired them with 13 different graphics to come up with the proposed warnings in the rule. FDA then tested this final set of 13 graphic health warnings, among others, in one quantitative survey examining their effects on consumer health beliefs.94 There is no support in the literature for FDA’s proposed warnings. Although FDA cites previous work in support of the proposed graphic health warnings, much of that work is inapposite because it studies different graphic warnings and overwhelmingly discusses their effects on non-U.S. populations, as discussed further below.95 See infra Section III.B.3. The work FDA cited that did include U.S. populations only assessed the graphic health warnings proposed by FDA in 2011, which FDA withdrew after the D.C. Circuit deemed the 2011 rule unconstitutional.96 The scientific evidence FDA invokes in the Proposed Rule suffers from a number of serious explanatory gaps. If FDA fails to correct these deficiencies in any final rule, the rule would violate the APA, which inter alia prohibits agencies from failing to support their conclusions with substantial evidence and requires agencies to connect their determinations to record evidence.97 First, FDA has failed to provide the final data sets from the quantitative studies to allow for replication and verification of the statistical analyses or examination of data quality. FDA also cited three qualitative research studies purporting to support development of the initial textual warnings and selection of the graphics, but reports are unavailable.98 According to FDA, these three qualitative research studies guided the development of the textual warnings and the selection of the images for the proposed graphic warnings.99 However, FDA did not provide the reports or transcripts for this initial work, making it impossible for stakeholders to understand

94 See 84 Fed. Reg. at 42,767-69. 95 See, e.g., id. at 42,763. 96 See R.J. Reynolds, 696 F.3d 1205; 84 Fed. Reg. at 42,757. 97 Dickinson v. Zurko, 527 U.S. 150, 164 (1999); Burlington Truck Lines v. United States, 371 U.S. 156, 168 (1962). 98 See 84 Fed. Reg. at 42,767 (“(OMB control number 0910–0674, ‘Qualitative Study on Cigarettes and Smoking: Knowledge, Beliefs, and Misperceptions’)”); id. at 42,770 (“(OMB control number 0910–0796, ‘Qualitative Study of Perceptions and Knowledge of Visually Depicted Health Conditions’)”); id. (“(OMB control number 0910–0796, ‘Qualitative Study on Consumer Perceptions of Cigarettes Health Warning Images’)”). 99 See id. at 42,767, 42,770.

28 how FDA selected the text or graphics. PM USA and Nat Sherman asked FDA to make this information as well as the quantitative data sets available so that they could address it in these comments,100 but FDA did not respond to our request. Second, FDA has failed to explain how it selected the particular potential health consequences of smoking it tested with consumers. The rationale for FDA’s selection of the health conditions initially tested in OMB Control No. 0910-0848 appears to be FDA’s determination “that some of the health conditions newly identified in the 2014 Surgeon General’s Report represented an opportunity to educate the public about the negative health consequences of smoking.”101 The Surgeon General has identified over 50 health consequences potentially associated with smoking, including the 11 conditions the Surgeon General newly identified in 2014.102 The Proposed Rule represents that “FDA determined that having warning statements focused on less-known health consequences of smoking accompanied by photorealistic images can promote greater public understanding of the risks associated with cigarette smoking.”103 FDA did not, however, explained what, if any, criteria it used in deciding which less-known potential health consequences to test with consumers. The Proposed Rule merely states that “revised statements were reviewed by FDA internal epidemiological experts to ensure that the health conditions under consideration were causally linked to cigarette smoking or exposure to secondhand smoke, and that these smoking-attributed conditions were not rare.”104 But the Proposed Rule does not explain whether other important criteria, such as language comprehension, believability, or informativeness, applied to FDA’s decision-making. Without this information, stakeholders cannot determine, or provide meaningful comment on, whether FDA’s selection of conditions was the product of a science-based assessment, or if FDA used some other criteria. This shortcoming is particularly critical because FDA’s explanation for which conditions its warnings pinpoint is inconsistent and incomplete. FDA lists “heart disease and chronic obstructive pulmonary disease (COPD)” alongside “lung cancer” and contrasts those three conditions with “less-known effects of smoking.”105 Yet FDA decided to feature heart disease and COPD but not lung cancer, among the proposed warnings, despite having classified all three as better-known effects of smoking and having emphasized the goal of educating about less- known effects. Compounding these inconsistencies, FDA elsewhere cited scientific evidence suggesting that consumers generally lack understanding of these conditions. FDA’s summary of the literature determined that “[c]onsumers suffer from a pervasive lack of knowledge about and understanding of the negative health consequences of smoking,” including findings that 33% of sampled adult smokers “did not know that cigarettes were a proven cause of cancer” and that “people underestimated the percent of people diagnosed with lung cancer who would die from the condition.”106 These findings fail to explain why FDA did not include warnings about lung cancer, yet focused on heart disease and COPD, and raise questions about FDA’s criteria for selecting particular conditions to highlight.

100 See Sept. 5, 2019 Request for Extension. 101 84 Fed. Reg. at 42,766. 102 Id. at 42,756. 103 Id. at 42,756-57. 104 Id. at 42,767. 105 Id. at 42,756, 42,758 106 Id. at 42,761.

29 In a similar vein, FDA provided no explanation for why it chose to develop warnings about certain less-known conditions as opposed to other less-known conditions with worse survival rates or greater incidences. The below table and figure (Table 1 and Figure 8), show a list of conditions, derived from Surgeon General Reports from 2004 and 2014, the relative risk that a current smoker will contract the condition (compared to a non-smoker), adult prevalence of the condition, and the projected five-year survival rate for the condition. FDA noted that “consumers are largely unaware of the negative health consequences of cigarette smoking not mentioned in current warnings as well as more specific information about the negative health effects and their mechanisms.”107 But FDA’s decision to focus on less-known health consequences does not explain why, even at the earliest stages of FDA’s studies, FDA zeroed in on some types of cancer (like mouth and throat cancer, head and neck cancer, and bladder cancer) but failed to test other types of cancer (like colorectal cancer, pancreatic cancer, cervical cancer, liver cancer, or kidney cancer) that are more prevalent or more fatal.

107 Id. at 42,766.

30 Table 1: Summary Table of Diseases Causally Linked to Smoking by the Surgeon General in 2004 or 2014 Reports with Relative Risks, Prevalence Estimation, and 5-Year Survival

Disease Included as Relative risk (current Adult 5-year 1 2 graphic smokers) prevalence survival % warning? Combined Head and Neck Cancer Yes 12.96 (Freedman, Abnet, 142,277 66% (Women) Leitzmann, Hollenbeck, & (SEER, 2018a) (Peng, Grogan, Schatzkin, 2007) & Wang, 2014) Combined Head and Neck Cancer Yes 5.45 340,105 (Men) (Freedman et al., 2007) (SEER, 2018a)

Trachea, Bronchus, and Lung Cancer No 10.92 520,195 18.1% (Men and Women) (Jayes et al., 2016) (SEER, 2018a) (SEER, 2018b) COPD Yes 4.01 16,800,000 41.5% (Jayes et al., 2016) (BRFSS, (Foucher et al., 2018) 1998) Bladder Cancer Yes 3.14 686,874 77.3% (van Osch, Jochems, van (SEER, 2018a) (SEER, 2018b) Schooten, Bryan, & Zeegers, 2016) Atherosclerotic Peripheral Vascular Yes 2.71 8,500,000 66% Disease (Lu, Mackay, & Pell, 2014) (Allison et al., (Criqui et al., 2007) 1992)

Stroke Yes 2.52 7,000,000 ~55%A (Robbins, Manson, Lee, (Roger et al., (Aparicio, Satterfield, & Hennekens, 2011) 2017) 1994) Coronary Heart Disease Yes 2.42 16,300,000 ~74%B (Hackshaw, Morris, (Roger et al., (Mukamal, Boniface, Tang, & 2011) 2001) Milenković, 2018) Aortic Aneurysm No 2.41 1,340,000 90.3% (Cornuz, Sidoti Pinto, (Hirsch et al., (Mani, Bjorck, Tevaearai, & Egger, 2004) 2006) Lundkvist, & Wanhainen, 2009) Age-related Macular Degeneration Yes 2.35 2,070,000 100% (AMD) (Thornton et al., 2005) (National Eye Institute, 2019) Cataracts (nuclear) Yes 2.24 20,500,000 100% (Christen et al., 1992) (Congdon et al., 2004)

31 Disease Included as Relative risk (current Adult 5-year 1 2 graphic smokers) prevalence survival % warning? Rheumatoid Arthritis No 2.02 66,800,000 100% (Di Giuseppe, Discacciati, (BRFSS, Orsini, & Wolk, 2014) 2018) Pancreas Cancer No 1.9 64,361 8.2% (Ordonez-Mena et al., 2016) (SEER, 2018a) (SEER, 2018b) Periodontitis No 1.85 64,700,000 98%C (Leite, Nascimento, Scheutz, (Eke et al., (Kim, Baker, & López, 2018) 2015) Davarian, & Crimmins, 2013) Cervix Cancer No 1.83 224,655 67.1% (Gandini et al., 2008) (SEER, 2018a) (SEER, 2018b) Asthma No 1.61 24,100,000 90%D (Jayes et al., 2016) (BRFSS, (Ringbaek, 2018) Seersholm, & Viskum, 2005) Stomach Cancer No 1.6 94,587 30.6% (Surgeon General (SG), (SEER, 2018a) (SEER, 2018b) 2004) Liver Cancer No 1.56 66,233 17.6% (Gandini et al., 2008) (SEER, 2018a) (SEER, 2018b) Acute Myeloid Leukemia No 1.52 61,048 26.9% (Colamesta et al., 2016) (SEER, 2019) (SEER, 2018b) Kidney and Ureter Cancer No 1.52 471,919 74.1% & (Gandini et al., 2008) (SEER, 2018a) 46.5%, respectively (SEER, 2018b) Erectile Dysfunction Yes 1.5 18,000,000 100% (Bacon et al., 2006) (Selvin, Burnett, & Platz, 2007) Diabetes (type 2) Yes 1.37 27,700,000 ~72%B (SG Report, 2014) (BRFSS, (Mukamal, 2018) 2001)

Colorectal Cancer No 1.2 1,300,518 64.9% (Ordonez-Mena et al., 2016) (SEER, 2018a) (SEER, 2018b)

Notes: 1Cancer prevalence is defined as first per site in the past 39 years. BRFSS estimates for adults assume an adult (18+) population of 253,900,000 and older adult (65+) population of 52,300,000 based on 2018 mid-year intercensal estimates. 2Cancer survival is 5-year relative survival percent. Relative survival is the percent of the diseased population that are alive divided by the general population alive after 5 years. Some chronic diseases have absolute survival reported when a relative survival estimate was not attained. Some conditions (cataracts, age-related macular degeneration, erectile dysfunction, and rheumatoid arthritis) are assumed to not contribute to mortality and have 100% relative survival. AAbsolute survival. BAbsolute survival. CAbsolute survival. DMen, absolute survival (Denmark). No applicable U.S. percentage found.

32 Figure 8: Data Visualization of Conditions Caused by Smoking, by Relative Risk and Five- Year Survival

Source: See Table 1. Third, FDA provided no evidence to support why it selected particular graphics to illustrate the textual warnings. Nor did FDA explain whether it considered alternative graphics to illuminate the same concepts or why it chose the selected graphics over others that could have illustrated the same text. The Proposed Rule states that FDA conducted a series of focus groups to examine “what factual information the images conveyed to participants about the negative health consequences of cigarette smoking in the absence of paired textual warning statements, as well as how concordant participants considered the images to be when paired with potential textual warning statements.”108 FDA used results from this qualitative research to select graphics used

108 Id. at 42,770.

33 in the later quantitative research study. However, FDA has not provided a report of findings or transcripts from these focus groups. Were the Proposed Rule to become final, these shortcomings would render it invalid under the APA for at least two independent reasons. FDA’s failure to provide the underlying data, analyses, and criteria underlying its studies deviates from the relevant procedural requirements. In particular, FDA’s failure to provide the information underlying its studies would violate the Tobacco Control Act, which clearly states that in “[e]ach notice of proposed rulemaking,” FDA “shall set forth . . . the manner in which interested persons may examine data and other information on which the notice or findings is based.” 21 U.S.C. § 387f(b)(1)). That requirement presupposes that FDA will disclose “data and other information on which” its rule “is based.” Congress took this requirement verbatim from the Medical Device Amendments to the FDCA “to provide a rulemaking procedure…in which all participants would have a full opportunity to present their views and analyses of the data underlying the proposed regulation.”109 FDA’s failure to provide critical underpinnings of its Proposed Rule also would render any final rule “not in accordance with law” and, thus, invalid under the APA. Moreover, it would run afoul of the APA principle that “public evaluation . . . [of] the technical studies and data upon which the agency relies [in its rulemaking]” is necessary to give stakeholders fair notice and a meaningful opportunity for comment.110 Further, FDA’s failure to provide the data, analyses, and criteria underlying the Proposed Rule also leaves the proposed graphic warnings without an adequate justification in evidence. FDA’s failure to explain how it developed the new textual warnings or chose the graphics prevents stakeholders from assessing whether FDA’s conclusions are valid because they cannot replicate or verify the results of the statistical analysis, let alone conduct additional analyses or ascertain data quality. Not only that, the two quantitative studies that FDA did conduct, and their manuscript-style reports, did not go through the peer review process, limiting credibility of the evidence provided.

B. FDA’s cited research is flawed and does not provide adequate support for the proposed graphic warnings Even setting aside the gaps identified above, FDA’s cited research is inadequate on its face, and falls far short of the substantial evidentiary threshold FDA would need to clear to justify the Proposed Rule under the APA. The following deficiencies render FDA’s proffered evidence inadequate. 1. FDA’s own quantitative study designs are inherently biased and fail to account for potential confounding variables As noted, FDA’s commissioned quantitative studies (OMB Control No. 0910-0848 [Study 1] and OMB Control No. 0910-0866 [Study 2]) are the centerpiece of its rule. Without those studies, any final rule would fail because no other evidence pertains to the warnings FDA has actually

109 21 U.S.C. § 360j(d); see Becton, Dickinson and Co. v. FDA, 589 F.2d 1175, 1181 (2d Cir. 1978). 110 Am. Radio Relay League, Inc. v. FCC, 524 F.3d 227, 236 (D.C. Cir. 2008); see also Owner-Operator Indep. Drivers Ass’n, Inc. v. Fed. Motor Carrier Safety Admin., 494 F.3d 188, 199 (D.C. Cir. 2007).

34 proposed. But these studies are inherently biased and suffer other serious design flaws, and thus fail to supply a sufficient evidentiary basis for FDA to justify the Proposed Rule. First, both studies fail to isolate critical variables and thus do not reliably test whether study participants actually learn anything new from the proposed warnings. Both studies compare new, more specific information in the proposed warnings to the more general statements contained in the 9 TCA textual warnings and in the 4 Surgeon General warnings currently required for cigarettes (e.g., mouth and throat cancer vs. unspecified cancer). In other words, both studies compare highly detailed to more general statements but fail to control for the fact that survey respondents may artificially inflate self-reported measures of “learning” or “new information” by conflating specificity and length of the new statements with new knowledge. In Study 1, comparisons of new statements to TCA statements when assessing “new knowledge” may skew results for certain comparisons. For example, the “new knowledge” ratings for TCA statements such as “cigarettes are addictive,” “cigarettes cause cancer,” or “smoking can kill you” were likely particularly low. The study was designed such that 8 of the 15 new, less general, statements (treatment) were paired with more generalized existing TCA statements on a similar health topic (control), leaving the remaining 7 new, highly specific statements matched with one of eight randomly selected TCA statements to serve as their controls. Differences in ratings of “new knowledge” may therefore be overinflated. Study 2 suffers from the same flaw, as well as the additional problems of failing to test whether the proposed graphics add any new information above and beyond the new text and failing to control for the effect of altering the warnings’ size and location. The purpose of Study 2 was to test whether the proposed warnings increased public understanding of the health consequences of smoking relative to existing warnings. To do so, Study 2 compared the four existing, less specific Surgeon General statements in their current size and location to 16 graphic health warnings occupying 50% of a mock pack or 20% of a mock advertisement. Critically, the study only tested the combined effects of the revised text and graphics. By testing the two together, the study failed to assess the effects of the revised text alone. As a result, it did not test for whether the graphic warnings impart any new information above and beyond the new text. Similarly, the study fails to isolate the effect, if any, of the size and location of the warnings. Outcomes of Study 2 included exposure to new information, thinking about risks, self-reported learning, perceived factualness, perceived informativeness, perceived understandability, changes in health beliefs, and recall. Those outcomes are inherently biased, because the health belief items in Study 2 were derived directly from the text of the proposed graphic health warnings. As participants responded to these items three times (pre-exposure, after the second exposure at day 2-3, and at day 16-17 before the recall question), the measure of recall for the graphic health warnings relative to the control could be inflated. Second, Study 1 and Study 2 outcomes did not include a formal test of comprehension or understanding of the revised textual statements or graphic warnings. Instead, FDA stated that the scientific literature demonstrates that the outcomes of “new information” and “self-reported learning” can predict public understanding of the risks associated with cigarette smoking as compared to a TCA statement. But FDA did not provide any evidence to support these conclusions. Inferring that higher self-report ratings of “new knowledge” and “learning” among non-representative samples with biased study designs are indicative of public understanding is a gross over-interpretation of these data. Without such information, it is impossible to determine

35 whether the proposed graphics increase understanding of smoking health risks or simply evoke reactance. For example, an image of a blind person may better improve understanding that smoking can lead to conditions that result in blindness, as opposed to the provocative image FDA selected for advanced macular degeneration, which shows a needle being inserted into an eye. Indeed, previous research with similar images has demonstrated reactions of fear and/or disgust among consumers without necessarily demonstrating improved understanding or comprehension (e.g., Refs. 77, 84, and 201). (Bhat et al., 2018; Cochran, Kydd, Lee, Walker, & Consedine, 2018; Hall, Mendel, Noar, & Brewer, 2018). Third, the applicability of both studies is suspect because they enrolled non-representative participants. To start, the Study 1 population consisted of only 4.2% of adults with less than high school education. According to the Centers for Disease Control website, however, almost 25% of adult cigarette smokers have less than a high school education.111 As FDA’s Proposed Rule acknowledges, participants with lower educational attainment possess inherently different reading comprehension levels.112 FDA’s failure to properly account for such individuals prevents FDA from reliably evaluating whether the text of the new warnings improves consumer understanding, let alone meaningfully assessing what new or additional information individuals might derive from the graphic components of the warnings. Similarly, the ending study population in Study 2 consisted mainly of adults age 25+, a sample that does not accurately represent young adults or adolescents. Under-inclusion of the younger age groups is problematic because previous U.S. research has demonstrated differences in responses to warnings among younger subpopulations (McCloud, Okechukwu, Sorensen, & Viswanath, 2017; Shadel et al., 2019). Underrepresenting these groups could artificially bias results. Fourth, Study 2 is flawed because FDA incorporated none of ALCS’ suggestions regarding demographic and other factors in its November 26, 2018 comments for enhancing “the quality, utility, and clarity of the information to be collected” into the study.113 Potential moderating factors such as , gender and sexual orientation, health literacy, and socioeconomic status (SES) among smokers were not evaluated even though FDA collected most of this information. FDA’s failure to consider these significant aspects concerning the efficacy of the proposed warnings is particularly problematic because previous research has shown that these factors impact the effectiveness of graphic warnings (Bekalu, Ramanadhan, Bigman, Nagler, & Viswanath, 2019; McCloud et al., 2017; Pagano et al., 2017; Quisenberry et al., 2018; Shadel et al., 2019; Tan, Bigman, Nagler, Minsky, & Viswanath, 2017). Relatedly, FDA in Study 2 failed to consider pre- and post- measures of risk perceptions to evaluate whether the proposed warnings meaningfully improved upon incoming risk perceptions. Without such baseline information, FDA and stakeholders cannot reliably evaluate the impact of the proposed warnings on participants’ understanding of the risks associated with cigarette smoking. Again,

111 https://www.cdc.gov/tobacco/data_statistics/fact_sheets/adult_data/cig_smoking/index.htm 112 84 Fed. Reg. at 42,765. 113 See Attachment 4, Docket No. FDA-2018-N-3552 (83 Fed. Reg. 48,625, Sept. 26, 2018) – Comments on “Agency Information Collection Activities; Proposed Collection; Comment Request; Experimental Study of Cigarette Warnings,” (“November 26, 2018 Comments”).

36 these design flaws call into question the fundamental reliability and soundness of FDA’s primary scientific justifications for the Proposed Rule. 2. FDA’s own studies fail to support the proposed graphic health warnings As noted, the only scientific evidence in the Proposed Rule that purports to test the replicability of the actual proposed warnings comes from FDA’s own commissioned studies (Study 1 and Study 2). The results of these studies, however, do not support the Proposed Rule. FDA’s quantitative study results demonstrate low or no impact of several tested statements or graphic- statement pairings on the outcome measures, despite the inherently biased study designs. FDA’s failure to substantiate the efficacy of its proposed warnings is another independent reason why the Proposed Rule—if finalized—would not survive scrutiny under the APA. In Study 1, “participants’ reports of learning new information were significantly higher for revised statements in 12 of 16 comparisons of revised to TCA statements.”114 But 4 of those 12 were less than a one-point difference in learning compared to the standard statement. Similar issues were observed with believability and informativeness. Moreover, the learning, believability and informativeness response option scales ranged from 1-7, labeled only at the anchors (1=not at all, 7=very much). Participants likely did not consistently interpret response options between 2 and 6 in the same way, calling into question the precision and accuracy of the findings generated using this scale, and casting further doubt on the small effect sizes (less than 1 point) observed in that study. Many of the revised statements did not result in statistically significant improvements in the outcomes relative to the TCA statements in Study 1. According to their own statements, it appears that for 11 of the 16 comparisons (over 2/3), respondents were not significantly more likely to say that the revised warning statement made them think about the relevant health risk above and beyond the TCA statements.115 Also, according to FDA’s statements in the Study 1 final report, it appears that respondents did not consider the revised statement to be more informative in 11 of the 16 comparisons of revised statements to TCA statements.116 It also appears that respondents considered seven of the revised statements to be significantly less factual than the TCA statements.117 Respondents also ranked some of the revised statements, such as amputation, diabetes, macular degeneration, erectile dysfunction, and cataracts, significantly less believable than the TCA statements.118 Rather than supporting FDA’s conclusion that its proposed warnings will be effective in improving comprehension, FDA’s own study findings undercut the evidence provided by calling into question the potential impact of many of the revised statements. Study 2 raises similar questions about the efficacy of the proposed warnings, even judged by the study’s own results. Many of the tested graphic health warnings did not result in statistically significant improvements in the outcomes relative to the existing Surgeon General warnings in

114 Experimental Study on Warning Statements for Cigarette Graphic Health Warnings: Study 1 Report (April 2018) (“Study 1 Report”) at 3-5, available at https://www.regulations.gov/document?D=FDA-2019-N-3065-0131. See also 84 Fed. Reg. at 42,768. 115 See Study 1 Report at 3-8. 116 See id. at 3-11. 117 See id. at 3-12. 118 See id. at 3-9.

37 Study 2.119 In Session 1 (i.e., their first exposure to the warnings), respondents considered eight out of 16 tested graphic health warnings to be significantly less factual than controls.120 In Session 2 (i.e., respondents’ second exposure to warnings, 2-3 days later), five out of 16 tested graphic health warnings did not show a positive change in respondents’ health beliefs.121 And in Session 3 (i.e., respondents’ third exposure to warnings, 16-17 days later), 9 out of 16 tested graphic health warnings did not show a positive change in respondents’ health beliefs.122 Moreover, of the 13 currently proposed warnings, 6 did not show a positive change in health beliefs after 16-17 days.123 In sum, FDA’s own studies—the only studies assessing the proposed warnings FDA has actually selected—seriously call into question whether FDA’s proposed warnings would be effective or improve comprehension. This evidence is too weak to carry FDA’s burden under the APA of providing substantial evidence to support its decision-making. 3. FDA’s other evidence improperly rests on non-U.S. data and/or contains numerous additional methodological flaws FDA’s other evidence does not provide the evidentiary support needed to justify the Proposed Rule, because this other evidence relies heavily upon data from other countries that require different warnings. FDA cites studies from several other non-U.S. countries to support the Proposed Rule (e.g., Refs. 102 [in Malaysia]; 103; 105; 106 [in China, India, Bangladesh and Korea]; and 107 [in Australia, Canada and Mexico]; and 203 [in Australia]).124 Much of this other evidence also rests on studies that are inapposite or include methodological flaws that prevent them from adequately illuminating whether FDA’s proposed warnings would be effective in educating U.S. consumers about the health consequences of smoking. First, FDA’s reliance on non-U.S. data is highly questionable because it contravenes, without explanation, FDA’s own scientific standards regarding studies of non-U.S. populations. In the Proposed Rule, FDA emphasizes “numerous non-U.S. studies” that it characterizes as showing an effect of graphic warnings yet discounts recent studies of U.S. populations that failed to find such an effect.125 FDA does not demonstrate that these foreign studies are applicable to the U.S. By contrast, in another recently proposed rule, FDA emphasized that in order to rely on “foreign” studies “conducted outside of the [U.S.],” premarket tobacco product applications “should provide a scientific rationale for why the results of the study can be generalized to other demographic groups that are representative of the U.S. population as [a] whole,” for instance by “discuss[ing] the factors that would be expected to influence study findings and whether they vary significantly across the U.S. population.”126 To avoid arbitrary and inconsistent standards, and for the sake of scientific rigor and sound decision-making, FDA must either apply the same

119 Experimental Study on Warning Statements for Cigarette Graphic Health Warnings: Study 2 Report (April 2018) (“Study 2 Report”), available at https://www.regulations.gov/document?D=FDA-2019-N-3065-0155. 120 See Study 2 Report at 4. 121 See id. at 4-1. 122 See id. at 4-3. 123 See id. at 3-7, 3-8. 124 84 Fed. Reg. at 42,789-796. 125 Id. at 42,763. 126 Premarket Tobacco Product Applications and Recordkeeping Requirements, 84 Fed. Reg. 50,566, 50,600 (Sept. 25, 2019). FDA further stated that applicants may need to address “intrinsic and extrinsic factors in foreign data [cited in] the International Council for Harmonisation (ICH) E5 guidance: Ethnic Factors in the Acceptability of Foreign Clinical Data.” Id. at n.12.

38 scientific standards to itself that it applies to others or provide an adequate explanation for differential treatment. Second, FDA’s reliance on non-U.S. data is especially questionable in the graphic-warnings context, for several reasons. The studies FDA invokes examined warnings in other countries that are substantively different (e.g., different sizes/locations, different text, different images associated with health risks, different health warnings) from FDA’s proposed warnings.127 (Refs. 3, 48, 102, and 103). FDA assumes that the effects of proposed warnings in other countries would be similar in the United States, yet nowhere does FDA account for the unique cultural and other contextual factors present in the U.S. FDA’s conclusions are manifestly inappropriate in light of its failure to consider this critical dimension of the problem. Compounding this problem, FDA fails to acknowledge material evidence showing that, even within regions with more homogenous populations, such as the European Union, the same text warnings have different levels of effectiveness. As Hammond 2011 (Ref. 39) explains: “Levels of effectiveness differ across countries, even for very similar health warnings. Indeed, the same text warnings have been implemented in virtually all EU member states since 2003; yet, smokers and non-smokers in different countries report different levels of effectiveness. Social norms surrounding tobacco use, as well as the strength of other tobacco control measures, likely mediate the impact of warnings. Individual-level differences, such as level of dependence, pre- existing health beliefs, and personal experience with the health effects of smoking may also mediate the impact of health warnings. In addition, not all messages resonate equally well with all individuals or target groups.” (emphases added). That review underscores the difficulty of extrapolating the experiences of one country to another. Each country has different economic, cultural, and governmental situations (e.g., Canada vs. Mexico), which confound any potential comparison across countries. That data is thus of little use in evaluating whether the warnings in the Proposed Rule will be effective or impactful in the U.S. Exacerbating the problem, many studies use the same International Tobacco Control (“ITC”) data to compare across countries (Refs. 3, 48, 49, 72, 82, 83, 86, 87, 89, 103, and 124). By relying on the same survey data set, involving the same period of time and the same underlying flaws, FDA has failed to establish replicability of the effectiveness of graphic health warnings in those countries. Third, FDA’s cited literature carries numerous study design and execution limitations, which prevent adequately informed scientific conclusions regarding the utility of pictorial health warnings in increasing knowledge. For instance, FDA cites two studies from Hammond et al. (Refs. 3 and 39) among others as support for its statement that, “Numerous non-U.S. studies support the role of pictorial cigarette warnings in promoting knowledge gains in cigarette-related

127 84 Fed Reg at 42,762-65. In Hammond et al., the authors discuss Canadian warnings (2001) – 16 rotating warnings, 50% of pack face front and back with color pictures, Australia warnings (1995) – 6 warnings, 25% front/33% back of pack, black and white text only; United Kingdom warnings (1992) – 6 warnings, 6% of face of pack, black and white text only; and four 1984 U.S. Surgeon General warnings from 1984. In Thrasher et al., the authors discuss Mexico’s warnings – 50% of back of pack, black and white text only and Canadian warnings – 50% of front and back of pack, color pictures. In Fathelrahman et al., the authors discuss Malaysian warnings – 9 mock- up warnings, 50% of front and back of pack, vivid color with extremely graphic images. In Kennedy et al., the authors discuss the same warnings discussed in Hammond et al., except they looked at Australia’s 1995 warnings (text-only) vs. Australia’s 2007 warnings which include the “eye” label – full color, very graphic image.

39 health risks after implementation of those warnings.”128 But the Hammond studies do not support that statement. Only one of the four non-U.S. countries evaluated (Canada) had pictorial warnings at the time (Ref. 3) and although the prevalence of adult smoking in Canada has declined 12-19% since implementation of the pictorial warnings in 2001, in 2011 Hammond concluded that this decline cannot be solely attributed to the new health warnings since “health warnings are typically introduced against a backdrop of other tobacco control measures, including changes in price/taxation, mass media campaigns, and smoke-free legislation” (Ref. 39). Reliance on the Hammond studies (Refs. 3 and 39) represents one of many weaknesses in FDA’s evidence supporting its Proposed Rule. Fourth, FDA provides no basis for assessing the comparative effectiveness of each individual proposed warning and the warnings’ effects over time. While a few studies have examined longer-term effects and changes in graphic health warnings’ effectiveness (e.g., Brewer et al., 2019 (Ref. 84) - 4 weeks with warning labels affixed to packs; Mercincavage et al., 2018- tested recall after 5 days and exposed all subjects to all graphic health warnings), the available scientific literature indicates that more research is needed to track exposure and subsequent behavior in situ to understand potential impact at the population level (Ref. 109). Again, FDA’s failure to address this limitation, or to acknowledge contrary evidence suggesting a need for more robust data, underscores the lack of adequate evidentiary support for the Proposed Rule. Fifth, FDA supports its claim that graphic health warnings increase understanding of smoking health risks with studies that use eye-tracking software to measure visual preference and attention to various areas of a on a screen. But those studies do not measure understanding of the cigarette pack warnings (e.g., Refs. 77 and 98).129 The investigators of these studies all emphasize the need for more real-world, relevant exposure and assessment in their discussion of study limitations. Moreover, Byrne et al., 2017 notes, “We cannot speak to the effects of repeated exposure to attention-grabbing, graphic FC [full-color] warning labels” (Ref. 77). Again, these limitations in the data illustrate the weakness of the scientific evidence FDA has invoked to support its Proposed Rule. Sixth, many FDA-cited studies lack an experimental procedure for exposure to imagery in their protocols. Rather, FDA’s cited studies relied solely on self-reported measures and impressions of any knowledge after the warnings in those countries had already been commercially implemented (Refs. 3, 48, 102, 103, 104, 105, 107, and 203).130 None of these FDA-cited studies systematically assessed the impact of the graphics in the health warnings. In addition, using those studies to justify the Proposed Rule here involves an improper comparison, because the FDA’s proposed graphic warnings have not yet been implemented. Seventh, studies cited by FDA fail to evaluate whether the graphic contributes to comprehension or understanding of the warning over and above the text alone. Most FDA-cited studies failed to evaluate comprehension at all (e.g., Refs. 3, 12, 77, 81, 105, and 109).131 In a source FDA does not cite, Carolyn Clancy emphasized in the Institute of Medicine (IOM) Roundtable on Health Literacy Workshop that “measures of health literacy must go beyond individual reading

128 Id. at 42,763. 129 Based on our review of Refs. 77 and 98. 130 Based on our review of Refs. 3, 48, 102, 103, 104, 105, 107 and 203. 131 Based on our review of Refs. 3, 12, 77, 81, 105 and 109.

40 capability in order to capture how well Americans understand what they hear and what they are told.”132 Further, the majority of FDA-cited studies do not systematically test the text and graphic alone and in combination, making it impossible to understand the unique contribution of either. To our knowledge, only a few studies systematically tested graphic health warnings (e.g., Refs. 12 and 77). Byrne et al., 2017 did not find any statistically significant differences in risk 133 beliefs between text-only warnings and graphic warnings with the same text (Ref. 77). That study did not include measures of knowledge level or understanding; rather, participants agreed or disagreed with a list of known health effects based on a read-and-recall task. Mutti et al. 2013 (Ref. 12) used a similar task to evaluate health beliefs, with no measure of comprehension, and did not test for differences between the text-only and text plus graphic conditions. FDA’s cited literature does not establish that graphics contribute to understanding or comprehension beyond the text alone. Finally, some studies that FDA cites attempted to measure real-world impact of graphic health warnings on behavior and intentions, but those studies are also faulty.134 They lack data to correlate intentions to longer-term number of quit attempts, future purchase behavior, reduction in cigarettes per day, etc., all of which are direct measures of smoking behavior (Refs. 12, 77, and 81). Brewer et al. (2019) (Ref. 84) measured quit attempts over a 4-week period, but the authors acknowledged that longer-term effects cannot be extrapolated. “Future studies should evaluate relationships of self-reported effectiveness rankings with recall and behavior change (i.e., smoking behaviors, changes in risk perceptions) to determine how this construct relates to more complex processes and behavior” (Mercincavage et al., 2018). Again, this evidence provides no sound basis for FDA’s decision to impose the proposed warnings as a means of addressing the health consequences of smoking.

C. FDA ignored contrary scientific evidence Even setting aside the weakness of the evidence FDA invokes, FDA has not provided a legally sufficient basis for the rule. That is because FDA has failed to adequately address a litany of evidence that undermines virtually every premise of the Proposed Rule. First, FDA did not adequately address contrary evidence indicating that graphic warnings do not meaningfully influence consumer knowledge regarding the consequences of smoking. FDA essentially ignores findings from U.S.-based studies—the most relevant evidence—showing that graphic warnings are ineffective in improving consumer comprehension regarding the health consequences of smoking. Indeed, FDA neglected to cite several U.S.-based studies that demonstrated little or no contribution of added graphics to textual warning messages (Bekalu et al., 2019; Shadel et al., 2019; Skurka et al., 2019; Van Dessel, Smith, & De Houwer, 2018). Other U.S. studies further undercut FDA’s conclusion that graphic warnings increase knowledge and accurate health beliefs concerning the consequences of smoking. For example, Robinson and Killen (Ref. 62) reported a paradoxical, significant increase in smoking behavior from baseline to three-month follow-up in high-school students with higher awareness of pack

132 IOM Roundtable on Health Literacy Workshop, Measures of Health Literacy: Workshop Summary (2009) https://www.ncbi.nlm.nih.gov/books/NBK45384/pdf/Bookshelf_NBK45384.pdf. 133 Based on our review of Ref. 77. 134 Based on our review of Refs. 12, 77 and 81.

41 warning labels. These labels did not contain the current proposed graphic warnings, but the authors suggest that youth in general are not influenced by the possibility of distal, longer-term negative health effects and that warnings could do more harm than good in this subpopulation. Even U.S. studies that demonstrate increased attention or fixation time on color graphic warnings on packaging relative to text only do not adequately support FDA’s approach (Refs. 77 and 84). Data from both studies suggest no differences in outcomes such as changes in beliefs or risk perceptions with the addition of a color graphic (Refs. 77 and 84). Byrne et al. (2017) concluded: “We thus cannot draw definitive conclusions about the degree to which graphic [full-color] labels are necessary to achieve downstream public health goals” (Ref. 77). Another study tested cigarette ads with graphic warning labels instead of packaging, but still found that health risk beliefs did not differ by warning label condition (Surgeon General’s warnings vs. proposed warnings (text-only) vs. graphic warning labels with text included). The effects of graphic warning labels on attention and negative affect were also equivalent across all conditions, indicating that the pictures did not affect these outcomes compared to text-only exposures (Ref. 109). FDA has cited, but not adequately addressed, this evidence, which contradicts the evidence supporting the Proposed Rule.135 Second, FDA’s analysis fails to account for scientific evidence indicating that its proposed warnings could backfire by prompting consumers to avoid the warnings or reject them in fear or disgust. Results from Van Dessel et al. (2018) suggest that “in order to find more beneficial effects of [graphic warnings] it might be important to use [graphic warnings] that are less (construed as) fear-evoking and therefore lead to less reactance.” Different graphic health warnings may garner different types of responses—like fear, defensiveness, and disgust—that would not increase knowledge. Likewise, another U.S. study that FDA fails to cite found that highly graphic warnings actually prompt lower levels of recall as compared with controls and warnings with less graphic content (Kees, Burton, Andrews, & Kozup, 2010). Research into low- and high-arousal emotional responses evoked by the previously proposed nine graphic warnings concluded that “the findings suggest the need for further research into the differential effects of equally valanced [graphic health warning]-evoked emotions of high and low arousal activation among smokers…” (Bekalu et al., 2019). Findings from two large randomized trials reported by Hall et al. (2018) involving close to 3,000 adult U.S. smokers suggest that graphic warnings that elicit greater negative affect can lead to greater message avoidance after initial exposure, and could also prompt consumers to avoid the warnings, for instance among consumers who interpret the warnings as a threat to personal freedom or choice. In a similar vein, the available data on avoidance of graphic cigarette pack warnings are mixed, but many studies indicate that graphic warnings may indeed cause unintended consequences contrary to FDA’s aims. Various types of studies demonstrate that some smokers avoid cigarette pack warnings after initial exposure (for a summary, see Hall et al. 2018). Multi-country studies including Canada and the U.S. indicate that as many as 30% of smokers regularly try to avoid thinking about or looking at cigarette pack warnings (Cho et al., 2016; Hammond, McDonald, Fong, Brown, & Cameron, 2004; Thrasher et al., 2016). One U.S.-based study found that 74%

135 84 Fed. Reg. at 42,763.

42 of participants indicated they would likely avoid graphic warning information by concealing their packs in some way (McCloud et al., 2017). Yet again, FDA failed to review this body of relevant literature in the Proposed Rule. Third, FDA failed to address evidence indicating that its size requirements for packaging and advertisements do not advance consumer understanding. Those studies indicate that the FDA’s proposed 50% size requirement for packaging and its 20% requirement for advertising space is not necessary to improving consumer understanding. For instance, Skurka et al. (2019) described a well-designed study investigating size of text-only and graphic health warnings in two distinct samples of adult smokers with low socioeconomic status. The major strength of this study was the inclusion of the same text in the control and graphic warning conditions, allowing a direct evaluation of the addition of a graphic to the messages conveyed by the text. The study found that increasing warning label size from 30% of the pack to 50% did not influence negative affect; in addition, risk belief acceptance was no different between control (text-only), 30% size or 50% size graphic warnings. Finally, the graphic warnings did not affect quit intentions; if anything, the graphic warnings reduced quit intentions, as the percentage of participants expressing quit intentions trended lower when shown graphic warnings as opposed to text. FDA cites an earlier study by Skurka et al. (Ref. 110) that explored whether graphic health warning size (30% or 50%) was associated with visual attention, negative affect, risk beliefs, and intentions. FDA summarily described this and a handful of other U.S.-based experiments as having “failed to find an effect of pictorial cigarette warnings on increasing health beliefs about the negative effects of smoking,” without acknowledging their results.136 But the earlier Skurka study (Ref. 110) in fact revealed that warning size has no impact on negative affect or risk beliefs among youth and adults and no impact on smoking susceptibility among youth. Fourth, FDA fails to adequately address evidence indicating that consumers already understand the health consequences of smoking, including some of the effects FDA’s Proposed Rule seek to highlight. A recent study by Polen et al. (2015) reports that more than 80% of participants, regardless of demographic factors, recognized that “smoking during pregnancy can cause the baby to be born too small or too early,” and 70% of women of reproductive age reported that if they became pregnant, they would quit smoking without any specific reasons from their doctor.137 Similarly, our November 26, 2018 Comments note, “There is ample research indicating that the general public already believes that cigarette smoking is extremely harmful, which would suggest that exposure to graphical warnings will not meaningfully increase risk perceptions.”138 FDA’s cited studies confirm that consumers understand the health risks communicated in the current Surgeon General’s warnings. FDA argues that individuals of low socioeconomic status are less likely to know the health effects of smoking. While cited studies (e.g., Refs. 69 and 136) do show gradients in knowledge or perception, the studies also demonstrate that even among lower socioeconomic status groups, most people report at least a awareness of the negative health effects of smoking. And, FDA’s cited studies reinforce that awareness of the negative health effects of smoking have increased over time (Ref. 69).

136 Id. 137 Karen Polen, et al., “Knowledge and Attitudes of Adults towards Smoking in Pregnancy: Results from the HealthStyles© 2008 Survey” (Jan. 19, 2015), available at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4594848/. 138 See Attachment 4, November 26, 2018 Comments.

43 Fifth, FDA failed to meaningfully address the differential effect graphic warnings may have on four particularly critical subpopulations. Regarding adolescents, FDA briefly discusses studies involving less-apposite, non-U.S. data concerning the effect graphic cigarette warnings may have in improving adolescents’ knowledge about the harms of smoking.139 But FDA ignores a host of studies that demonstrate that adolescents may react to graphic warnings with potentially perverse or unintended responses. For instance, Shadel et al. (2019) examined graphic warning posters at point-of-sale locations in a simulated convenience store environment and showed that there was a significant increase in future smoking susceptibility among adolescents already at high risk for smoking, with no impact on committed never-smokers. That study shows that graphic warnings at points of sale could perversely draw adolescents’ attention toward the array of tobacco-related stimuli and product offerings. Several other publications reviewing the literature have also described the potential for warning messages to backfire in younger populations (Evans et al., 2017; Harris, Mayle, Mabbott, & Napper, 2007; Leshner, Bolls, & Thomas, 2009; Sussenbach, Niemeier, & Glock, 2013). A more recent study measured intention to avoid graphic health warnings in young adult and adult smokers and found that young adults under age 25 had significantly higher odds of intending to avoid the graphic health warning-based information compared to adults over 50 (McCloud et al., 2017). At a minimum, researchers have called for further research to “determine whether highly graphic warnings are as effective for younger segments of interest (e.g., adolescents (Kees et al., 2010)).” Adding to these difficulties, FDA does not account for adolescents who purchase loose or bagged cigarettes, who are less likely to be knowledgeable of the health effects of smoking compared to those who bought cigarettes in packages/cartons (Ref. 11). Likewise, FDA failed to meaningfully consider lower socioeconomic status, the less educated, and low health literacy in its rulemaking process, even though FDA’s quantitative studies collected information on these subpopulations. FDA’s studies assessed literacy but did not analyze the data by literacy level or test for the impact of literacy level on outcomes. Instead, FDA appears to have assumed equal exposure to the graphic warnings and equal comprehension and effectiveness of the messages. But significant other evidence calls those assumptions into question. McCloud et al. (2017) studied avoidance of graphic warning information in low socioeconomic position smokers. Almost 75% of such smokers indicated that they would likely avoid the graphic health warning altogether, although many simultaneously expressed an intent to seek out more information about the health risks of smoking. Bekalu et al. (2019) found that not all negative emotions evoked by graphic health warnings may be effective in influencing quit-related outcomes in groups with a low socioeconomic position. Another study suggests lower health literacy scores were associated with more relative viewing time of the pictorial elements of the warnings as well as decreased recall (Quisenberry et al., 2018). FDA cites two studies that assessed the impact of graphic health warnings on behavior over time (Refs. 4 and 87) among low socioeconomic status individuals, but in the longitudinal studies, the effects of graphic health warnings declined or remained relatively flat over time. Several studies, such as Ngan et al. (Ref. 209) and Bekalu (2019), show that the effectiveness of graphic warnings wanes over time for smokers with a lower socioeconomic status. Finally, FDA does not address the impact of other forms of warnings and images (like The Real Cost campaign) on lower socioeconomic status groups. For example, research suggests that the recall of messages and

139 84 Fed. Reg. at 42,763.

44 impact on health beliefs from The Real Cost campaign can vary by parental education level and race (Huang et al., 2017). Similarly, FDA failed to meaningfully consider nicotine and substance dependence as factors that may inhibit the efficacy of warnings for particular subgroups of smokers. Several studies indicate cigarette smokers with higher nicotine dependence or cigarette smokers who are substance dependent may be less impacted by graphic health warnings (McCloud et al., 2017; Pagano et al., 2017; Shadel et al., 2019). One study reported that adult smokers who perceived themselves “somewhat addicted” to cigarette smoking had 2.3 times greater odds of intending to avoid graphic health warnings compared to those who reported themselves “not at all addicted” (McCloud et al., 2017). Another study reported that graphic health warnings deterred lower- dependent but not highly-dependent smokers from purchasing their cigarettes (Shadel et al., 2019). Earlier studies demonstrate similar results (Hammond et al., 2004; Rousu & Thrasher, 2012); (Ref. 48). FDA’s Proposed Rule, however, fails to consider level of dependence at all in their study populations or in any other way. Finally, FDA failed to meaningfully consider the effects of gender and sexual orientation, despite some data suggesting that these traits correspond to differences in the effectiveness of graphic health warnings. For example, Pagano et al. (2017) reported gender differences in assessments of image salience in graphic health warnings on their cigarette packs. Males reacted most negatively to images of women and children as opposed to images of men, while females reacted most negatively to an image of a baby (Pagano et al., 2017). Effectiveness of graphic health warnings was lower in certain sexual and gender minority groups (lesbian, transgender and gender nonconforming) in a study by Tan et al. (2017), and different warnings evoked different levels of perceived effectiveness. These sexual and gender minority groups exhibit higher smoking prevalence compared to national averages, thus lower efficacy of graphic health warnings in this population could lead to additional smoking-related health disparities. As noted above, FDA’s quantitative studies collected information on gender and sexual orientation but failed to provide analyses accounting for these characteristics to determine whether the proposed warnings may differentially impact these subpopulations over time.

D. FDA failed to assess alternatives to the proposed warnings FDA’s evaluation of the scientific evidence suffers from a further deficiency: the APA requires agencies to consider alternatives in reaching reasoned decisions, but FDA’s Proposed Rule fails to adequately evaluate or justify rejecting obvious alternatives to its proposed graphic warnings. First, FDA did not adequately evaluate whether refreshing or rotating the Surgeon General warnings or requiring new, text-only warnings would be sufficient to warn consumers. FDA did not include any express reasoning as to why this approach would be insufficient, let alone provide adequate evidence rejecting this approach. FDA did point to evidence that the photorealistic images provide new information above and beyond the new textual warnings. That conclusion is questionable, see supra Section III.B and infra Section IV.D. Even if that proposition were valid, it would not negate the possibility that rotating existing warnings or requiring text-only warnings might be equally effective in educating the public. Compounding this problem, FDA failed to address the novelty effect and the wearing-out of the message impact over time and did not cite studies looking at these issues (Anshari et al., 2018). FDA thus failed

45 to consider whether this alternative would be equally effective as the proposed graphic warnings over time. Second, FDA did not evaluate whether smaller or differently placed warnings on packaging and advertising (for instance, a warning displayed on 30% of the packaging and 15% of advertising space) would advance FDA’s objectives equally well. To our knowledge, FDA did not study or evaluate different sizes or locations of the graphic warnings. Neither FDA’s own studies nor the literature it cites appears to undertake this critical analysis, and as noted, some unaddressed evidence (Skurka et al., 2019) indicates that increased warning size may not affect measures of consumer affect or risk belief acceptance. Third, FDA did not consider the potential for enhanced public education campaigns to advance its objective. See supra Section IV.B.

E. No evidence justifies FDA’s proposal to include two graphic COPD warnings The Proposed Rule proposes 12 textual warnings paired with 13 graphics. It associates two different graphics with the warning “Smoking causes COPD, a lung disease that can be fatal.” One depicts a man with a nasal canula. The other image – a pair of lungs removed from the body – appears to be the same or similar graphic that is also paired with a different textual warning, albeit slightly rotated. FDA, however, has not provided any scientific basis for requiring two graphics for the COPD textual warning when only one graphic is required for all other textual warnings. The decision to include two COPD graphics does not seem to be based on any scientific evidence. In fact, results of FDA’s studies (OMB Control No. 0910-0848 and OMB Control No. 0910-0866) show that the COPD messaging and graphics had some of the lowest effects on ratings of learning or new information relative to the other revised statements or graphic health warnings tested.140 FDA’s failure to support its decision to include two COPD graphic warnings with any scientific evidence, as well as its failure to address evidence undercutting the soundness of this decision, would render this aspect of the Proposed Rule manifestly unlawful under the APA if incorporated in a final rule. Moreover, including two different graphics for COPD produces a total of 13 warnings—an odd and prime number that, as explained above, would cause extremely difficult and potentially insurmountable technical challenges with respect to the equal and random printing and distribution of the warnings on cigarette packages, see supra Section II.A.

IV. THE PROPOSED WARNINGS VIOLATE THE FIRST AMENDMENT The proposed warnings—if finalized—would also violate the First Amendment, and thus the APA, which prohibits agency action that is “contrary to constitutional right, power, privilege or immunity.”141 The “freedom of speech” guaranteed by the First Amendment “includes both the right to speak freely and the right to refrain from speaking at all.”142 To protect that constitutional guarantee, the government can compel commercial disclosures only upon a showing under Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio, 471 U.S.

140See Study 1 Report and Study 2 Report. 141 5 U.S.C. § 706(2)(B). 142 Janus, 138 S. Ct. at 2463 (2018) (internal quotation marks omitted).

46 626 (1985), that the disclosures (i) contain “purely factual and uncontroversial information” and (ii) are not “unjustified or unduly burdensome.”143 Otherwise, the government may regulate the content of speech only by demonstrating (at the very least) that doing so “directly and materially advances a substantial state interest in a manner no more extensive than necessary to serve that interest”—that is, by satisfying intermediate scrutiny.144 The Proposed Rule falls short of satisfying those standards, and thus violates the First Amendment, for at least three reasons. First, the Proposed Rule requires cigarette manufacturers to dedicate 50% of their packages and 20% of their advertisements to promoting the government’s message through large, prominently placed warning statements. Regardless of the content of the mandated warnings, the government’s commandeering of such a large amount of premium space “drowns out the [manufacturer’s] own message” and significantly restricts manufacturer speech.145 FDA has failed to carry its burden to show that those size and placement requirements are not “unjustified or unduly burdensome” means to further its asserted interest in improving consumer understanding of smoking-related health risks, much less that the requirements are no more extensive than necessary.146 Second, the graphic warnings would turn cigarette packages and advertisements into mini- billboards that dissuade consumers from purchasing the product, and appeal to emotion to further the government’s anti-smoking position. The D.C. Circuit invalidated FDA’s 2011 graphic- warnings rule in R.J. Reynolds Tobacco Co. v. FDA, 696 F.3d 1205 (2012), because those warnings did not convey any new and “uncontroversial” “warning information” beyond the text, but instead were “primarily intended to evoke an emotional response, or, at most, shock the viewer into retaining the information in the text warning.”147 The D.C. Circuit concluded that the warnings constituted an effort to “browbeat consumers into quitting” smoking rather than make “an accurate statement about cigarettes,” and thus fell outside Zauderer’s allowance for “purely factual” disclosures.148 Further, the D.C. Circuit held, FDA could not justify the graphic warnings under the intermediate-scrutiny standard by invoking an asserted interest in “effectively communicating health information regarding the negative effects of cigarettes,” because “[a]llowing FDA to define ‘effectiveness’ however it sees fit would not only render [the] ‘substantial interest’ requirement [under intermediate scrutiny] a complete nullity, but it would also eviscerate the requirement that any restriction ‘directly advance’ that interest.”149 The new proposed warnings suffer from the same flaws. They again require manufacturers to feature content designed to provoke emotional responses that dissuade consumers from buying cigarettes, and FDA again asserts a problematic interest in “effective” communication of health information.150 Under any potentially applicable First Amendment standard, FDA has failed to satisfy its heavy burden to justify the speaker-, content-, and viewpoint-based speech restrictions that its proposed warnings would impose.

143 NIFLA, 138 S. Ct. at 2372 (quoting Zauderer, 471 U.S. at 651). 144 Ibanez, 512 U.S. at 142. 145 NIFLA, 138 S. Ct. at 2378. 146 Id. at 2377. 147 R.J. Reynolds, 696 F.3d at 1216. 148 Id. at 1216-17. 149 Id. at 1221 (internal quotation marks omitted). 150 84 Fed. Reg. at 42,756.

47 Third, the proposed graphic and textual warnings are all but certain to mislead consumers regarding the health risks associated with cigarette smoking. Because the First Amendment prohibits compelling private actors to engage in misleading speech, the fact that the proposed warnings would likely mislead consumers renders them invalid speech regulations. FDA relies upon Discount Tobacco City & Lottery, Inc. v. United States, 674 F.3d 509 (6th Cir. 2012), to justify the constitutionality of its Proposed Rule. But the Sixth Circuit in that case merely rejected a facial First Amendment challenge to the Tobacco Control Act’s provisions requiring color graphic warning labels on cigarettes. Unlike the D.C. Circuit, which carefully scrutinized and invalidated FDA’s prior proposed warnings in R.J. Reynolds, 696 F.3d 1205, the Sixth Circuit did not assess the constitutionality of any particular warnings. Further, the Sixth Circuit’s decision rests on reasoning that the Supreme Court has since questioned. For instance, the Sixth Circuit held that a compelled disclosure need only be “reasonably related to the purpose” of a law,151 but subsequent Supreme Court decisions such as National Institute of Family. & Life Advocates v. Becerra, 138 S. Ct. 2361, 2377 (2018), clarify that compelled disclosures may “extend no broader than reasonably necessary.” The Sixth Circuit’s Discount Tobacco decision thus furnishes an inadequate legal justification for the Proposed Rule.

A. FDA’s proposed size and placement requirements unconstitutionally burden speech The Proposed Rule would require cigarette manufacturers to reserve “the top 50 percent of the area of the front and rear panels of cigarette packages and at least 20 percent of the area at the top of cigarette advertisements” for the government’s prescribed anti-smoking warnings.152 Those ratios are unprecedented and violate the First Amendment based on the sheer size and placement of the warnings alone. By commandeering such a substantial portion of manufacturers’ packaging and advertising space, the warnings would impose an unjustified and undue burden on manufacturers’ ability to convey their own speech on their own products. FDA’s Proposed Rule offers no justification for why these particular size and placement requirements are necessary, beyond the illegitimate and overly broad suggestion that bigger is always better for consumer comprehension.153 And FDA’s size and placement requirements do not advance FDA’s stated interest in educating consumers in a manner no more extensive than necessary to serve that interest. Regardless of the applicable First Amendment framework, the size and placement requirements would not pass constitutional muster. Even assuming that the applicable framework is the more deferential standard for evaluating compelled speech under Zauderer , FDA’s size and placement requirements are problematic. Under that standard, the government can compel disclosures only if they (i) contain “purely factual and uncontroversial information” and (ii) are not “unjustified or unduly burdensome.”154 FDA “has the burden” of demonstrating that its proposed warnings meet those criteria.155

151 Discount Tobacco, 674 F.3d at 562. 152 84 Fed. Reg. at 42,757. 153 See id. at 42,759, 42,778-79. 154 NIFLA, 138 S. Ct. at 2372 (quoting Zauderer, 471 U.S. at 651). 155 Id. at 2377.

48 Even setting content aside, the excessively large size and prominent placement of the required warnings on advertising and packaging constitute an unjustified and undue burden under the Zauderer framework. Supreme Court “precedents require [compelled] disclosures . . . to extend no broader than reasonably necessary,” lest “they risk ‘chilling’ protected speech.”156 Thus, the Supreme Court in NIFLA invalidated a California statute requiring pregnancy centers to include a 29-word, “government-drafted statement” on “all print and digital advertising materials,” reasoning that the statement unduly burdened facilities by requiring them to “call attention to the [government’s] notice, instead of [their] own message, by some method such as larger text or contrasting type or color,” and to relay the message in multiple languages.157 Similarly, the Ninth Circuit invalidated San Francisco’s required warning for sugar-sweetened beverages solely because they would have occupied 20% of the label, reasoning that the city failed to show the warning “does not ‘drown out’” the advertisers’ messages and “effectively rule out the possibility of having an advertisement in the first place,” or that a smaller warning would not have vindicated the city’s goals.158 The reasoning of those cases is fatal to FDA’s proposed size and placement requirements, which are at least as onerous. The proposed size and placement of FDA’s warnings compel manufacturers to “call attention to the [government’s] notice, instead of [their] own message” in a way all but certain to “drown out” the manufacturer’s own message.159 The required warnings are not only excessively large—50% of packages and 20% of advertisements—but would also occupy premium real estate on packaging and in advertisements, depriving manufacturers of the space most likely to help adult tobacco consumers recognize brands. Those size and placement criteria appear to have one overriding purpose: to crowd out the (lawful) visual aspect of the manufacturers’ speech. FDA’s proposed requirements are especially burdensome considering that cigarette manufacturers maintain few “alternative channels” of communication with consumers.160 Federal law already significantly restricts the avenues through which cigarette manufacturers may advertise to consumers.161 Most states impose further restrictions on cigarette manufacturers, including limitations on where they can advertise and what they can promote.162 If the Zauderer compelled-speech framework did not apply, FDA’s size and placement requirements would face an even more rigorous standard of scrutiny. At a minimum, FDA must justify those requirements under the intermediate-scrutiny standard that the Supreme Court applied to commercial speech in Central Hudson Gas & Elec. Corp. v. Public Serv. Comm’n of N.Y., 447 U.S. 557, 563-64 (1980). That standard “is not light,”163 and FDA has failed to show that requiring manufacturers to adopt FDA’s mandatory warnings “directly and materially

156 Id. (internal quotation marks omitted). 157 Id. at 2378. 158 Am. Beverage Ass’n v. City & Cty. of San Francisco, 916 F.3d 749, 756-57 (9th Cir. 2019) (en banc). 159 NIFLA, 138 S. Ct. at 2378. 160 Fla. Bar v. Went For It, Inc., 515 U.S. 618, 634 (1995). 161 See, e.g., Lorillard Tobacco Co. v. Reilly, 533 U.S. 525, 571 (2001); 15 U.S.C. § 1335. 162 See Letter from Altria Client Services Inc. to FDA, Doc. No. FDA-2010-N-0136-0028 at 2 (July 19, 2010) (discussing Master Settlement Agreement with various States), available at https://www.regulations.gov/document?D=FDA-2010-N-0136-0028. 163 R.J. Reynolds, 696 F.3d at 1218.

49 advances a substantial state interest in a manner no more extensive than necessary to serve that interest.”164 To start, FDA’s stated interest in promulgating the rule is the promotion of “greater public understanding of the negative health consequences of cigarette smoking.”165 But the Proposed Rule does not establish that covering 50% of manufacturers’ packaging and 20% of manufacturers’ advertisements with a government-mandated warning is “no more extensive than necessary” to serve that interest.166 Quite the contrary, Congress added the 50% and 20% requirements to the TCA without conducting any fact-finding.167 Likewise, FDA’s own commissioned studies apparently did not test or evaluate different sizes (or locations) of the proposed graphic warnings. FDA suggests that larger warnings attract more notice from consumers,168 but FDA offers no explanation for why those precise requirements are necessary. Indeed, under FDA’s reasoning, FDA could require its warnings to occupy 99% of a package to maximize consumer comprehension. That is an absurd result. And FDA has cited no studies that could support it. Indeed, FDA has not cited any evidence establishing that larger warnings advance its stated interest in improving consumer understanding in a way that smaller warnings would not.169 If anything, relevant literature—literature FDA has not addressed—suggests that warning size does not meaningfully affect the impact warnings have on consumers.170 Based on these shortcomings, FDA’s Proposed Rule would not satisfy even the intermediate- scrutiny standard, let alone the more stringent requirements of “exacting” and “strict” scrutiny that arguably apply.171

B. FDA’s proposed graphic warnings are impermissible speaker-, content- and viewpoint-based regulations of speech FDA’s Proposed Rule violates the First Amendment for another, independent reason: just like the previous graphic warnings that the D.C. Circuit invalidated, the new warnings are not “purely factual” disclosures designed to provide information about the risks of cigarette smoking, but instead aim at eliciting a negative emotional response from consumers.172 FDA asserts that its proposed warnings further its interest in better educating consumers about the dangers of smoking, but the D.C. Circuit has made clear that a “vague” interest in more “effective” communication of smoking-related health risks cannot support graphic-warnings requirements under the intermediate-scrutiny standard.173 In short, the Proposed Rule reflects a quintessential

164 Ibanez, 512 U.S. at 142. 165 84 Fed. Reg. at 42,757. 166 Ibanez, 512 U.S. at 142. 167 Compare H.R. 1256, 111th Cong. § 201 (June 11, 2009) (engrossed Senate amendment), with H.R. 1256, 111th Cong. § 201 (Mar. 3, 2009) (as placed on the House calendar). See also R.J. Reynolds Tobacco v. U.S. FDA, 823 F. Supp. 2d 36, 52 (D.D.C. 2011), vacated on other grounds by R.J. Reynolds, 696 F.3d 1205. 168 84 Fed. Reg. at 42,759, 42,778-79. 169 Cf. Am. Beverage Ass’n, 916 F.3d at 757 (required warning occupying 20% of label invalid in part because “findings suggest[ed] that the Ordinance’s goals could be accomplished with a smaller warning”). 170 See supra Section III.C (discussing Skurka et al. (2019); Skurka et al. (2018)). 171 See infra Section IV.B. 172 R.J. Reynolds, 696 F.3d at 1216-17. 173 Id. at 1221.

50 speaker-, content- and viewpoint-based restriction that cannot survive constitutional scrutiny, regardless of which First Amendment framework applies. Ordinarily, content-based speech regulations are subject to strict scrutiny, i.e., the government must establish that the regulation advances a compelling governmental interest and is narrowly tailored to restrict no more speech than necessary.174 The Supreme Court in Zauderer, however, carved out a “narrow” exception that applies a less stringent standard to certain government- mandated commercial disclosures.175 But again, that standard applies only to disclosures involving “purely factual and uncontroversial information” that are also not “unjustified or unduly burdensome.”176 FDA’s proposed graphic warnings satisfy neither of those criteria. First, FDA’s proposed graphic warnings are not “the type of purely factual and uncontroversial information . . . to which the Zauderer standard may be applied.”177 The D.C. Circuit’s opinion in R.J. Reynolds invalidating FDA’s prior proposed graphic warnings is instructive. There, the D.C. Circuit rejected FDA’s argument that the warnings were permissible disclosures of “‘purely’ factual” information.178 According to the court, the 2011 graphic warnings were not designed to “convey any warning information” beyond that contained in the relevant text, but instead were “primarily intended to evoke an emotional response” in consumers—“such as worry and disgust” or “embarrassment.”179 In other words, as the district judge in R.J. Reynolds wrote, “the emotional response [the graphic warning labels] were crafted to induce is . . . an objective wholly apart from disseminating purely factual and uncontroversial information.”180 Thus, the D.C. Circuit held, “[w]hile none of [FDA’s] images are patently false, they certainly do not impart purely factual, accurate or uncontroversial information to consumers. Consequently, the images fall outside the ambit of Zauderer.”181 The same reasoning applies to the graphic warnings in the Proposed Rule. Those graphic warnings do not meaningfully convey “any warning information” not already present in the text. Rather, they are “primarily intended to evoke an emotional response or, at most, shock the viewer into retaining the information in the text warning.”182 FDA concedes as much: one of FDA’s stated goals was to choose “attention-grabbing” graphic warnings.183 And FDA appears

174 See, e.g., Ashcroft v. ACLU, 542 U.S. 656, 666 (2004). 175 R.J. Reynolds, 696 F.3d at 1212 (internal quotation marks omitted). 176 NIFLA, 138 S. Ct. at 2372 (quoting Zauderer, 471 U.S. at 651); see also CTIA – The Wireless Ass’n v. City of Berkeley, Cal., 928 F.3d 832, 845 (9th Cir. 2019). 177 R.J. Reynolds, 696 F.3d at 1216; see NIFLA, 138 S. Ct. at 2372. 178 R.J. Reynolds, 696 F.3d at 1216. 179 Id. at 1216-17. 180 R.J. Reynolds, 823 F. Supp. 2d at 45. 181 R.J. Reynolds, 696 F.3d at 1217. In American Meat Institute v. USDA, the D.C. Circuit overruled one aspect of the R.J. Reynolds decision not at issue here. See 760 F.3d 18, 22-23 (D.C. Cir. 2014) (en banc). American Meat emphasized, however, the various limitations of Zauderer discussed here, including the requirement—enforced in R.J. Reynolds—that a disclosure contain “purely factual and uncontroversial information.” Id. at 27. 182 R.J. Reynolds, 696 F.3d at 1217. What’s more, because Congress did not clearly instruct FDA to use emotional response as a factor in designing the graphic warnings, construing the TCA as authorizing FDA to do so would violate the principle of statutory interpretation that “where an otherwise acceptable construction of a statute would raise serious constitutional problems,” courts must “construe the statute to avoid such problems unless such construction is plainly contrary to the intent of Congress.” Edward J. DeBartolo Corp. v. Fla. Gulf Coast Bldg. & Constr. Trades Council, 485 U.S. 568, 575 (1988). 183 84 Fed. Reg. at 42,770.

51 to have relied on the Surgeon General’s 2014 report on the health consequences of smoking, which emphasizes, “[i]nfluential and successful [anti-smoking] campaigns contain a number of essential elements, including optimized themes, appropriate emotional tone, appealing format, clear messages, intensity, and adequate repetition.”184 As the D.C. Circuit recognized in R.J. Reynolds, graphic warnings that “evoke emotion” rather than inform are clearly designed to advance the government’s anti-smoking message (i.e., are viewpoint-based), which further precludes application of the Zauderer standard. 185 The overwhelming viewpoint behind mandatory warnings describing various smoking-related conditions—especially outsized warnings accompanied by graphics that appear calculated to shock or outrage viewers—is that consumers should never purchase or use cigarettes. But Zauderer does not permit the government to “require corporations to carry the messages of third parties, where the messages themselves are biased against or are expressly contrary to the corporation’s views.”186 Second, FDA’s proposed warnings are “unjustified” and “unduly burdensome” and thus fall outside the Zauderer standard in all events, because they restrict speech more broadly “than reasonably necessary.”187 As discussed below, the images chosen by FDA illustrate the point. Those images are clearly designed to “evoke emotion”—and thus advance an anti-smoking viewpoint—rather than impart accurate or useful information regarding the health risks of cigarettes.188 By compelling cigarette manufacturers to promote FDA’s anti-smoking message, the Proposed Rule would significantly interfere with manufacturers’ commercial speech rights. Yet FDA has failed to make the requisite showing that inclusion of the selected images—as opposed to new text-only warnings—is “reasonably necessary” to advance its asserted goal of educating consumers.189 If the government could classify the proposed graphic warnings as permissible compelled speech that position would give the government free rein to accomplish a host of other patently unconstitutional results. FDA could require manufacturers of prescription painkillers, for example, to include images of heroin addicts injecting themselves on their labels to illustrate the possibility of to opioids. The government could compel all beer, wine, or spirits to carry graphics of livers affected by liver disease or display mid-surgery photos of a liver transplant. Or the government could force cell-phone manufacturers to fund billboards showing gruesome automobile accidents caused by distracted driving. Because the content of FDA’s proposed warnings puts them outside the Zauderer standard, FDA must at least justify those warnings under the intermediate-scrutiny standard, i.e., show that its burden on speech “directly and materially advances a substantial state interest in a manner no more extensive than necessary to serve that interest.”190 Recent Supreme Court decisions suggest FDA must shoulder an even higher burden because strict scrutiny should apply to

184 U.S. Dep’t of Health and Human Services, Office of the Surgeon General, The Health Consequences of Smoking—50 Years of Progress 814 (2014). 185 R.J. Reynolds, 696 F.3d at 1217. 186 Pac. Gas & Elec. Co. v. Pub. Utilities Comm’n of Cal., 475 U.S. 1, 16 n.12 (1986) (plurality op.). 187 See NIFLA, 138 S. Ct. at 2372, 2377. 188 R.J. Reynolds, 696 F.3d at 1217. 189 NIFLA, 138 S. Ct. at 2377. 190 Ibanez, 512 U.S. at 142.

52 regulations, such as viewpoint-based or speaker-based restrictions on commercial speech, that fall outside of Zauderer.191 But, even under the intermediate-scrutiny standard—and, thus, under any more stringent standard that might apply—FDA’s proposed warnings cannot pass muster for three independent reasons. First, FDA lacks a substantial state interest in the Proposed Rule. FDA’s suggestion that consumers need pictures to understand the health risks associated with smoking is antithetical to fundamental First Amendment considerations, and thus cannot be a valid interest. FDA states that mandating graphic warnings furthers its interest in “effective” consumer education regarding the risks of smoking.192 But FDA does not suggest, much less demonstrate, that the text of its warnings—existing or new—is too complicated for consumers to understand. For instance, FDA does not claim that consumers (or even youths) cannot understand textual warnings. FDA’s justification for its graphic warnings, instead, is that the longtime Surgeon General’s warnings on cigarette packages and in cigarette advertisements have become “invisible.”193 Because consumers have stopped reading, FDA demands pictures to entice consumers to pay attention, which FDA in turn argues will lead to a “more effective” warning regime.194 The D.C. Circuit has explained, however, that FDA’s “interest in ‘effective’ communication” of health information related to smoking is “not an independent interest capable of sustaining” a graphic- warnings rule.195 Without evidence that people cannot understand the content of health warnings, FDA has no legitimate interest in going further to promote those warnings by restricting manufacturer speech. Underscoring the lack of any valid governmental interest, FDA’s suggestion that consumers need pictures to understand clear textual warnings defies key First Amendment principles. The First Amendment depends on the premise that Americans can comprehend competing ideas and decide which to espouse. Because “the best test of truth is the power of the thought to get itself accepted in the competition of the market,” any “fear that people would make bad decisions if given truthful information cannot justify content-based burdens on speech.”196 The government thus cannot attempt to overwhelm manufacturers’ own speech on packaging and advertisements with emotionally charged images. Likewise, the fear that people would make bad decisions unless the government could compel private parties to disclose information in the manner the government deems most effective cannot possibly justify the speaker-, content-, and viewpoint- based burdens the Proposed Rule would impose. Perhaps for that reason, FDA painstakingly avoids framing the Proposed Rule as a measure for advancing an anti-smoking message.197 But that is the message of these graphics, nonetheless.

191 See NIFLA, 138 S. Ct. at 2371-72 (indicating that compelled disclosure requirements falling outside of Zauderer should be analyzed as “[c]ontent-based regulations” under the “stringent” strict-scrutiny standard); see also Matal v. Tam, 137 S. Ct. 1744, 1764 (2017) (op. of Alito, J.); Sorrell v. IMS Health Inc., 564 U.S. 552, 571 (2011). 192 84 Fed. Reg. at 42,759. 193 Id. at 42,755, 42,779. 194 Id. at 42,762. 195 R.J. Reynolds, 696 F.3d at 1221. 196 Abrams v. United States, 250 U.S. 616, 630 (1919) (Holmes, J., dissenting); Sorrell, 564 U.S. at 577 (internal quotation marks omitted). 197 Cf. R.J. Reynolds, 696 F.3d at 1218 n.13 (expressing “skeptic[ism] that the government can assert a substantial interest in discouraging consumers from purchasing a lawful product, even one that has been conclusively linked to adverse health consequences”).

53 Second, the graphic warnings do not advance FDA’s stated interest in “promoting greater public understanding of the negative health consequences of cigarette smoking” directly enough to survive under any potentially applicable test.198 FDA’s selected images do not themselves enhance understanding of the health risks mentioned in the textual warnings.199 Indeed, as discussed in more detail below, many of the images convey no health information at all, but instead derive their content exclusively from the accompanying text. Therefore, even were FDA correct that some consumers cannot comprehend simple textual warnings, none of the images would materially advance FDA’s asserted interest in more effectively promoting consumer understanding. Third, FDA has failed to show that its proposed restriction of speech is no more extensive than necessary to serve its asserted interest in promoting consumer understanding. FDA already compels manufacturers to include warnings on cigarette packages and implementing the nine warning statements required by the TCA would sufficiently advance the government’s interest in providing information to the public.200 FDA suggests that the proposed graphic warnings would further its asserted interest both by informing the public of the “less-known” health risks associated with smoking and by overcoming consumers’ tendency to gloss over existing Surgeon General’s warnings.201 Even if true, FDA’s rationales do not justify compelling the graphic warnings that the Proposed Rule demands. FDA has other alternatives to promote consumer understanding that would burden less speech. For example, FDA did not consider the potential for enhanced public education campaigns to promote consumer understanding.202 The success of these campaigns demonstrates that to advance its objective FDA could “inform the [public] itself” about the risks of smoking “with a public-information campaign,” rather than requiring manufacturers to distribute its message.203 Even if public response to that campaign was “tepid,” that would not “prove that an advertising campaign is” insufficient.204 Or, if FDA is concerned that the existing warnings are stale, FDA could require manufacturers to rotate those warnings so they appear more frequently—a step FDA did not take during the many years between the invalidated original rule and this new rule. Or FDA could rely on new, text- only warnings. It could also incorporate an attention-grabbing headline in all capital letters, such as “PLEASE READ!” FDA has not shown that any of these or many other obvious alternative options would be inadequate to serve its interests. See supra Section III.D.

198 84 Fed. Reg. at 42,755. 199 See R.J. Reynolds, 696 F.3d at 1219-21 (rejecting reliance on “questionable social science” showing the impact of graphic warnings). See also infra Section IV.D. 200 See United States v. Playboy Entm’t Grp. Inc., 529 U.S. 803, 816 (2000). 201 See 84 Fed. Reg. at 42,759-761, 42,767. 202 See https://www.fda.gov/TobaccoProducts/PublicHealthEducation/default.htm (identifying multiple FDA- sponsored public education campaigns); see also https://www.fda.gov/tobacco-products/real-cost-campaign/real- cost-cost-effective-approach (describing results of FDA’s “The Real Cost” campaign as “impressive,” having reached more than 90% awareness in its target audience in the first year alone, changed perceptions and beliefs about tobacco, and reduced smoking initiation in that population by 30% from 2014 to 2016). 203 NIFLA, 138 S. Ct. at 2376. 204 Id.

54 Further, FDA’s explanation of how frequently consumers fatigue of warnings only underscores why FDA’s Proposed Rule is a poor fit for its asserted interests. FDA’s justification would suggest that all warnings will become stale, including the proposed graphic warnings. For those warnings to remain effective, FDA would have to compel additional new warnings continually, perhaps with even more sensational graphics and messages, just to attain any enduring degree of effectiveness. Either the Proposed Rule would be a temporary stopgap of limited effect, or the Proposed Rule would be the first in a series of infinite incursions on manufacturers’ speech.

C. The textual and graphic warnings are misleading The proposed graphic warnings suffer from an additional, independent First Amendment problem. Many of the warnings—both text and graphics—would mislead the public. Specifically, the text and graphic warnings mislead because they emphasize extreme and unlikely conditions and treat conditions of differing severity as if they were equal. By grouping together conditions with different smoking-attributable risks and severity (e.g., diseases with high mortality versus diseases that are not fatal), the text and graphic warnings make certain diseases appear to be equivalent or more likely than they actually are. The government cannot compel misleading speech under any First Amendment standard. As discussed, the government can compel speech only under two conditions. The government can satisfy the Zauderer standard, which requires that any compelled speech be (i) “purely factual and uncontroversial information,” and (ii) not “unjustified or unduly burdensome.”205 But misleading speech, by definition, is not “purely factual and uncontroversial.”206 If Zauderer does not apply, the government must (at a minimum) satisfy intermediate scrutiny.207 But that standard, too, is impossible for the government to satisfy if its restriction on speech requires manufacturers to convey misleading information. The government could have no legitimate interest in requiring misleading speech, as misleading speech has no First Amendment value in the commercial-disclosure context.208 The proposed textual and graphic warnings are misleading in several critical respects, as discussed in detail below, see infra Section IV.D. To start, the warnings do not distinguish among the different risks associated with each of the targeted health consequences, providing only that smoking “causes” all of the named conditions. By using that formulation, the proposed warnings misleadingly suggest to consumers that all of the targeted health risks stand on the same footing. FDA’s proposed requirement that manufacturers rotate these warnings on packages and in advertisements so as to give each warning “equal” play,209 exacerbates the risk of confusion by further making it appear that the risks of less likely conditions are on par with more likely health consequences of smoking. Aggravating the problem, the proposed warnings highlight comparatively rare conditions or outcomes, but fail to feature grave but more common conditions, such as lung cancer. Further, FDA’s proposed warnings feature numerous chronic, non-fatal conditions, but not other

205 NIFLA, 138 S. Ct. 2372 (quoting Zauderer, 471 U.S. at 651). 206 Id. 207 Ibanez, 512 U.S. at 142. 208 See In re R. M. J., 455 U.S. 191, 203 (1982). 209 84 Fed. Reg. at 42,755.

55 conditions with high mortality rates, such as lung cancer in smokers (17.6% 5-year survival rate) or pancreatic cancer (8.2% 5-year survival rate). Likewise, FDA features erectile dysfunction and diabetes, but omits conditions such as aortic aneurysms, arthritis, periodontitis, cervical cancer, asthma, stomach cancer, liver cancer, acute myeloid leukemia, and kidney and ureter cancer—all serious conditions that have a similar or stronger association with smoking than erectile dysfunction or diabetes. Again, that unexplained choice threatens to mislead consumers regarding the relative seriousness of smoking-related health risks. Even if FDA believes that most consumers are already aware of the risks of all of those omitted conditions such that further warnings would not be useful, FDA has not explained that conclusion. Nor does the Proposed Rule justify FDA’s decision to warn of the particular risks included in its proposed warnings as opposed to other lesser-known but more common risks of smoking. Further, even when describing the effects of particular conditions, various warnings appear misleading in isolation because they emphasize—either through their text, graphics, or both— extreme and relatively unlikely consequences of a given condition, as opposed to more likely effects. To be clear, PM USA and Nat Sherman in no way minimize the gravity of the conditions FDA has identified or their impact on those who experience them, but the proposed warnings emphasize extreme consequences, “suggesting that” relatively rare conditions are in fact “a common consequence of smoking.”210

D. Each Warning Is Constitutionally Deficient

• The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected image portrays a sickly-looking child in a hospital gown in need of oxygen. Adults viewing the image would be horrified at the thought of inflicting such harm on their children.

• The selected image of a sickly child does not convey the FDA Description: “The image “purely factual information” FDA asserts. No reasonable shows the head and shoulders consumer would be able to determine from the image of a young boy (aged 8-10 that the child depicted has “chronic asthma from years) wearing a hospital gown secondhand smoke exposure” or is “receiving a nebulizer and receiving a nebulizer treatment.” Rather, the hospital gown, the child’s sickly treatment for chronic asthma appearance, and the mask over the child’s face suggest resulting from secondhand only that the child is experiencing a medical emergency smoke exposure.” 84 Fed. that requires receipt of oxygen. Reg. at 42,773. • Contrary to FDA’s stated criteria for image selection, the text and image are not “concordant.” 84 Fed. Reg. at 42,765. The text of the warning indicates only that the child has suffered “harm.” That general description is

210 R.J. Reynolds, 696 F.3d at 1216.

56 not clarified by the picture, which indicates that the child is suffering from a condition requiring a hospital stay and receipt of oxygen but does not clarify what the condition is. The warning’s troubling ambiguity regarding the harm at issue adds to the fear and confusion a consumer would experience when viewing the warning. • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected image portrays a small, crying, newborn infant—an image designed to provoke an instinctive, emotional need in adult viewers to comfort the child and alleviate its distress. • Contrary to FDA’s stated criteria for image selection, the image contains a “non-essential element.” 84 Fed. Reg. FDA Description: The image at 42,770. There is no doubt that smoking during depicts “[a]n infant with low pregnancy indeed stunts fetal growth, and PM USA and birth weight resulting from Nat Sherman agree that pregnant mothers should avoid stunted fetal growth. The smoking. But depicting the infant’s apparent distress is image shows a newborn infant not essential to portraying the fetal-growth effects subject on a medical scale, and the to the warning and can only be viewed as an effort to digital display on the scale appeal to viewers’ emotions. reads four pounds.” 84 Fed. Reg. at 42,773. • The selected image is misleading and not “concordant” with the text. 84 Fed. Reg. at 42,765. Although FDA has provided scientific support for the link between maternal smoking and lower birth weight, the scale’s depiction of “4.00 lbs” conveys a very low birth weight commonly associated with premature birth. FDA has not claimed, let alone demonstrated, that a birth weight of four pounds is a likely outcome of maternal smoking. Yet the image would mislead consumers into believing that maternal smoking will commonly cause low birth weight to this degree. • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The image appears designed to provoke an emotional reaction of fear that many experience when faced with the prospect of a medical procedure involving needles. The further depiction of the insertion of the needle into a wide-open eye depicts an extremely uncomfortable

medical procedure that is designed to provoke an FDA Description: The image emotional reaction of fear, shock, and disgust. “depict[s] a closeup of an older

57 man (aged 65 years or older) • The selected image of a needle being inserted into an eye who has age-related macular does not convey “purely factual information.” No degeneration caused by reasonable consumer would be able to determine from cigarette smoking. The man is the depiction of an injection into an eye that the man receiving an injection in his depicted has “age-related macular degeneration” or that right eye to prevent additional the purpose of the injection depicted is to “prevent vessel growth.” 84 Fed. Reg. additional vessel growth.” at 42,777. • Contrary to FDA’s stated criteria for image selection, the text and image are not “concordant.” 84 Fed. Reg. at 42,765. The text of the image indicates that smoking can lead to blindness. Yet nothing about the picture indicates that the individual depicted is blind. Nor does FDA’s description of the image suggest the individual is blind. The mismatch between the text and the image depicted adds to the fear and confusion a consumer would experience when viewing the warning. • The warning is misleading because it emphasizes a condition—blindness—that occurs in only a minority of cases of age-related macular degeneration (“AMD”). Further, the warning, when viewed alongside FDA’s other proposed warnings, incorrectly suggests to consumers that their risk of exposure to smoking- attributable AMD and blindness is the same as or greater than their risk of exposure to other, more-prevalent smoking-related conditions that other warnings depict, such as COPD—especially since the AMD warning must be distributed equally alongside other warnings. See supra Table 1. Finally, the warning emphasizes a chronic, non-fatal condition. This is not to diminish the seriousness of AMD. But FDA omits from the warnings other conditions with high mortality rates, such as trachea, bronchus, and lung cancer, pancreas cancer, stomach cancer, liver cancer, and acute myeloid leukemia, raising further questions about whether the warnings directly advance FDA’s stated educational aims. See supra Fig. 8. • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The image of a red-orange colored liquid in a specimen cup— which the text describes as “bloody urine”—appears designed to provoke an emotional reaction of fear or disgust regarding the nature of the depicted liquid.

58 FDA Description: “The image • The selected image of a gloved hand holding a cup filled shows a gloved hand holding a with a red-orange colored liquid does not, on its own, urine specimen cup containing convey “purely factual information.” No reasonable bloody urine resulting from consumer would be able to determine from the image bladder cancer caused by alone that the liquid depicted is “bloody urine,” let alone cigarette smoking.” 84 Fed. “bloody urine resulting from bladder cancer.” All the Reg. at 42,774. image conveys is that an unknown person with a gloved hand is holding an unspecified red-orange liquid in a sample cup. • Contrary to FDA’s stated criteria for image selection, the text and image are not “concordant.” 84 Fed. Reg. at 42,765. The text of the image indicates that smoking can lead to bladder cancer. Yet nothing about the picture indicates that bladder cancer is the subject of the warning. The mismatch between the text and the image depicted adds to the fear and confusion a consumer would experience when viewing the warning. • The warning is misleading on its face because it suggests that bloody urine is a more serious health concern than bladder cancer, a fatal disease. Although formulating the warning in that manner allows FDA to include a troubling image with emotive power, it misleads consumers regarding the seriousness of the health consequences at issue. Moreover, the warning—when viewed as part of a rotational sequence of warnings that all get equal distribution—is misleading because it suggests to consumers that bladder cancer is just as or more prevalent than other, more-prevalent smoking- related conditions that the warnings depict, such as COPD, peripheral vascular disease, stroke, or coronary heart disease. Likewise, the warning also misleadingly suggests to consumers that their relative risk of bladder cancer is the same as their relative risk of exposure to other smoking-related conditions with smaller relative risk rates, such as AMD, cataracts, erectile dysfunction, or type 2 diabetes. • Finally, the warning is misleading because it emphasizes bladder cancer while other conditions with worse survival rates, such as kidney and ureter cancer, cervix cancer, colorectal cancer, stomach cancer, acute myeloid leukemia, trachea, bronchus, and lung cancer, liver cancer, and pancreas cancer are omitted from the warnings. See supra Fig. 8.

59 • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected image of a pair of discolored and disfigured feet with several toes amputated is disturbing and unsightly and is clearly designed to provoke either disgust at the sight of the image, fear at the prospect of undergoing an amputation, or both. • The selected image of the discolored and disfigured feet FDA Description: The “image does not convey “purely factual information.” No depict[s] the feet of a person reasonable consumer would be able to determine from who had several toes the image alone that the individual whose feet are amputated due to tissue depicted had an amputation due to “tissue damage from damage resulting from peripheral vascular disease.” Rather, the only thing peripheral vascular disease discernible from the image is that the depicted feet are caused by cigarette smoking.” blackened and missing several toes. Again, the text gives 84 Fed. Reg. at 42,776. meaning to a disturbing image, rather than the other way around. • Contrary to FDA’s stated criteria for image selection, the text and image are not “concordant.” 84 Fed. Reg. at 42,756. The text indicates that smoking can reduce blood flow to the limbs. Yet the graphic simply shows amputated toes. Nothing about the picture indicates that the amputation resulted from reduced blood flow, let alone that the reduced blood flow reflects peripheral vascular disease. The mismatch between the text and the image depicted adds to the fear and confusion a consumer would experience when viewing the warning. • The warning—when viewed as part of a rotational sequence of warnings that all get equal distribution—is misleading because it suggests to consumers that peripheral vascular disease is just as or more prevalent than other, more-prevalent smoking-related conditions that other warnings depict, such as COPD, coronary heart disease, or cataracts. Further, the warning suggests that consumers’ relative risk of exposure to peripheral vascular disease is the same as their relative risk of exposure to other smoking-related conditions with smaller relative risk rates, such as stroke, coronary heart disease, AMD, cataracts, erectile dysfunction, or type 2 diabetes. See supra Table 1. • In addition, the warning is misleading because it emphasizes peripheral vascular disease, while other conditions with worse survival rates, such as stomach

60 cancer, acute myeloid leukemia, trachea, bronchus, and lung cancer, liver cancer, and pancreas cancer are omitted from the warnings. See supra Fig. 8. • Finally, the warning is misleading because it depicts a complication that is an uncommon effect of reduced blood flow yet fails to depict the more common side effects of peripheral vascular disease, such as numbness or weakness in the legs. • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected image of a man with one artificially enhanced, discolored eye is discomforting and appears designed to shock the viewer or generate fear at the prospect of experiencing the condition in the image. • The selected image of a man with one discolored eye does not convey the “purely factual information” FDA FDA Description: The “image asserts. The image on its own, without the depict[s] a closeup of the face accompanying text, simply shows a man with one eye of a man (aged 65 years or differently colored than the other. There is no reason for older) who has a cataract a consumer to know that the depicted eye-color variation caused by cigarette smoking. represents “a large cataract.” The man’s right pupil is covered by a large cataract.” • Contrary to FDA’s stated criteria for image selection, the 84 Fed. Reg. at 42,777. text and image are not “concordant.” 84 Fed. Reg. at 42,765. The text of the image indicates that smoking can lead to blindness. Yet the picture does not clearly indicate that the individual depicted is blind. And in fact, the individual appears to not be blind. Losing sight in one eye is tragic, but it does not meet the definition of blindness. • The warning—when viewed as part of a rotational sequence of warnings that all get equal distribution—is misleading because it suggests to consumers that their risk of exposure to smoking-attributable cataracts and blindness is the same as or greater than their risk of exposure to other, more-prevalent smoking-related conditions that other warnings depict, such as COPD. See supra Table 1. • Further, the warning emphasizes a chronic, non-fatal condition. This is not to diminish the seriousness of cataracts. But FDA omits from the warnings other conditions with high mortality rates, such as trachea, bronchus, and lung cancer, pancreas cancer, stomach

61 cancer, liver cancer, and acute myeloid leukemia, raising further questions about whether the warnings directly advance FDA’s stated educational aims. See supra Fig. 8. Finally, the warning emphasizes a condition— blindness—that occurs in only a small minority of cases (0.48%) of cataracts.211 • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected image of blood-covered hands holding bloody, diseased lungs that appear to have been taken from a deceased person is intended to shock and disturb viewers with its goriness or to generate fear about the prospect of death and having one’s lungs removed postmortem. • The selected image of blood-covered hands holding FDA Description: “The image bloody, diseased lungs does not convey “purely factual shows gloved hands holding a information.” The image does not provide any pair of diseased, darkened information not already contained in the text, such as, for lungs removed from a smoker example, an illustration of the link between COPD and with COPD.” 84 Fed. Reg. at smoking. Indeed, FDA has failed to explain how a 42,775. typical layperson would understand that the lungs pictured are “diseased.” • The warning is misleading because it suggests to consumers that their relative risk of exposure to COPD is the same as their relative risk of exposure to other smoking-related conditions with smaller relative risk rates, such as bladder cancer, peripheral vascular disease, stroke, coronary heart disease, AMD, cataracts, erectile dysfunction, or type 2 diabetes. See supra Table 1. Finally, the warning emphasizes COPD while other conditions with worse survival rates, such as stomach cancer, acute myeloid leukemia, trachea, bronchus, and lung cancer, liver cancer, and pancreas cancer are omitted from the warnings. See supra Fig. 8.

211 According to the National Eye Institute, the number of cataracts cases in the U.S. in 2010 was 24,408,000. https://nei.nih.gov/learn-about-eye-health/resources-for-health-educators/eye-health-data-and-statistics. If it is assumed that blindness is represented among the incident cases of cataracts, 0.48% of cataracts cases result in blindness in the U.S.

62 • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected image of a man receiving oxygen through an oxygen tank appears designed to make consumers fearful of the prospect of needing to rely upon an oxygen tank to survive. • The selected image does not, on its own, convey “purely FDA Description: The “image factual information.” No reasonable consumer would be depict[s] a man receiving able to determine from the image alone that the man oxygen support because he has depicted suffers from COPD. All the image conveys is COPD caused by cigarette that the man needs oxygen support. smoking. The image shows • The warning is misleading because it suggests to the head and neck of a man consumers that their relative risk of exposure to COPD is (aged 50-60 years) who has a the same as their relative risk of exposure to other nasal canula under his nose smoking-related conditions with smaller relative risk supplying oxygen; the oxygen rates, such as bladder cancer, peripheral vascular disease, tank can be seen behind his left stroke, coronary heart disease, AMD, cataracts, erectile shoulder.” 84 Fed. Reg. at dysfunction, or type 2 diabetes. See supra Table 1. 42,775. Finally, the warning emphasizes COPD while other conditions with worse survival rates, such as stomach cancer, acute myeloid leukemia, trachea, bronchus, and lung cancer, liver cancer, and pancreas cancer are omitted from the warnings. See supra Fig. 8. FDA has failed to justify the requirement of two separate graphic images to accompany the text of the COPD warning. • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected image of a man with a large, recently-sutured incision is intended to disgust or shock consumers by depicting a graphic wound or to make consumers fearful of the prospect of needing to undergo major heart surgery and medical monitoring.

• Contrary to FDA’s stated criteria for image selection, the FDA Description: The image text and image are not “concordant.” 84 Fed. Reg. at “depict[s] a patient who 42,765. The text of the image indicates that smoking can recently underwent heat lead to heart disease and strokes. Yet the image, on its surgery to treat heart disease own, does not convey that the individual depicted either caused by smoking. The suffered from heart disease or a stroke. All the image image shows the chest of a shows is an individual with a large incision on his chest man (aged 60-70 years) who wears a hospital gown and has unspecified probes. wearing an open hosptial gown. The man has a large, • The warning is misleading because it suggests to recently-sutured incision consumers that their relative risk of exposure to heart

63 running down the middle of disease and stroke is the same as their relative risk of his chest and is undergoing exposure to other smoking-related conditions with post-operative monitoring.” smaller relative risk rates, such as AMD, cataracts, 84 Fed. Reg. at 42,774. erectile dysfunction, or type 2 diabetes. See supra Table 1. Further, the warning emphasizes heart disease and stroke while other conditions with worse survival rates, such as stomach cancer, acute myeloid leukemia, trachea, bronchus, and lung cancer, liver cancer, and pancreas cancer are omitted from the warnings. See supra Fig. 8. • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected image of a woman with a large tumor protruding from her neck is disturbing and unsightly and is clearly designed to provoke disgust or discomfort at the sight of the image, fear at the prospect of experiencing the same uncomfortable medical condition, or both. • The selected image does not, on its own, convey “purely FDA Description: “The image factual information.” The image, on its own, shows a shows the head and neck of a woman with a large growth in her neck region, but woman (aged 50-60 years) whether that growth is a tumor or something else is not who has neck cancer caused by obvious; only the text conveys key details. cigarette smoking. The woman has a visible tumor • The warning is misleading in the context of the Proposed protruding from the right side Rule’s warning regime because it suggests to consumers [of] her neck just below her that their relative risk of exposure to head and neck jawline.” 84 Fed. Reg. at cancer is the same as their relative risk of exposure to 42,774. other smoking-related conditions with smaller relative risk rates, such as COPD, bladder cancer, peripheral vascular disease, stroke, coronary heart disease, aortic aneurysm, AMD, cataracts, erectile dysfunction, or type 2 diabetes. See supra Table 1. Further, the warning emphasizes head and neck cancer while other conditions with worse survival rates, such as stomach cancer, acute myeloid leukemia, trachea, bronchus, and lung cancer, liver cancer, and pancreas cancer are omitted from the warnings. See supra Fig. 8.

64 • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The image appears designed to provoke the emotional reaction of fear or disgust that many experience when faced with the prospect of a medical procedure involving needles and drawing blood. • Contrary to FDA’s stated criteria for image selection, the image contains a “non-essential element.” 84 Fed. Reg. FDA Description: The “image at 42,770. The depiction of blood being drawn, which depict[s] a personal glucometer threatens to cause an emotional or fearful reaction in device being used to measure many consumers, is not necessary to informing the blood glucose level of a consumers regarding the risk of diabetes. person with type 2 diabetes caused by cigarette smoking. • The selected image of a blood being drawn from a finger The digital display reading of and the associated glucometer reading, on its own, does 175 mg/dL and a notation on not convey “purely factual information.” The average the glucometer indicate a high consumer is unlikely to be aware of the meaning of the blood sugar level.” 84 Fed. “175” reading on the glucometer (or even to recognize Reg. at 42,776. the device as a glucometer). For that reason, the image also is not “concordant” with the text, as it does not relate to diabetes without knowledge of additional information not depicted. • The warning is misleading in the context of the Proposed Rule’s warning regime because it suggests to consumers that their relative risk of exposure to type 2 diabetes is the same as or greater than their relative risk of exposure to other, more-prevalent smoking-related conditions, such as trachea, bronchus, and lung cancer, aortic aneurysm, pancreas cancer, periodontitis, cervix cancer, asthma, stomach cancer, liver cancer, acute myeloid leukemia, or kidney and ureter cancer. See supra Table 1. Further, the warning emphasizes diabetes while other conditions with worse survival rates, such as cervix cancer, colorectal cancer, stomach cancer, acute myeloid leukemia, trachea, bronchus, and lung cancer, liver cancer, and pancreas cancer are omitted from the warnings. See supra Fig. 8.

65 • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected image of blood-covered hands holding bloody, diseased lungs from a deceased individual is intended to shock and disturb viewers with its goriness or to generate fear about the prospect of death and having one’s lungs removed postmortem. • The selected image of blood-covered hands holding FDA Description: “The image bloody, diseased lungs does not convey “purely factual shows gloved hands holding a information.” The image does not provide any pair of diseased lungs information not already contained in the text, nor does it containing cancerous lesions clarify the types of lung disease nonsmokers may from chronic secondhand experience or the circumstances under which they may smoke exposure.” 84 Fed. be at risk of contracting fatal lung disease from Reg. at 42,773. secondhand smoke exposure. Nor is it clear why a typical layperson would understand that the lungs pictured are “diseased” and contain “cancerous lesions.” • The warning is misleading in context of the warning regime because it emphasizes a condition—fatal lung disease caused by secondhand smoke—that is less prevalent than other than other smoking-attributable health conditions, such as trachea, bronchus, and lung cancer, pancreas cancer, cervix cancer, asthma, stomach cancer, liver cancer, acute myeloid leukemia, or kidney and ureter cancer, raising further questions about whether the warnings directly advance FDA’s stated educational aims. See supra Table 1; 2014 Surgeon General’s Report Table 12.4. • The warning impermissibly seeks to advance FDA’s anti- smoking message by “evok[ing] an emotional response in consumers.” R.J. Reynolds, 696 F.3d at 1216. The selected bedroom scene of a despondent man and a woman “look[ing] off in another direction” is clearly designed to generate embarrassment and shame in viewers regarding the sensitive topic of sexual intimacy. • The selected image, on its own, does not convey “purely factual information.” The image does not provide any FDA Description: The “image health information. The man on the bed may be depict[s] a man who is experiencing erectile dysfunction. Or he may be experiencing erectile distressed for unrelated reasons. The text of the relevant dysfunction caused by warning gives content to the image, and the image on its smoking. The image shows a own in no way illuminates how smoking could cause man (aged 50-60 years) sitting erectile dysfunction. on the edge of a bed and

66 leaning forward, with one • The warning is misleading because it suggests to elbow resting on each knee. consumers that their risk of exposure to smoking- The man’s head is tilted down, attributable erectile dysfunction is the same as or greater with his forehead pressed into than their risk of exposure to other, more-prevalent the knuckles of his right hand. smoking-related conditions, such as trachea, bronchus, Behind him on the bed, his and lung cancer, aortic aneurysm, pancreas cancer, female partner looks off in periodontitis, cervix cancer, asthma, stomach cancer, another direction.” 84 Fed. liver cancer, acute myeloid leukemia, or kidney and Reg. at 42,776. ureter cancer. Further, the warning emphasizes a chronic, non-fatal condition. This is not to diminish the seriousness of erectile dysfunction. But FDA omits other conditions with high mortality rates, such as trachea, bronchus, and lung cancer, pancreas cancer, stomach cancer, liver cancer, and acute myeloid leukemia, raising further questions about whether the warnings directly advance FDA’s stated educational aims. See supra Fig. 8. Finally, the warning focuses on erectile dysfunction while omitting mention of more common side effects of low blood flow, such as numbness or weakness in the legs.

V. CONCLUSION We appreciate the opportunity to provide our views and look forward to continuing to engage with FDA.

67 VI. SCIENTIFIC REFERENCES Anshari, D., Yong, H. H., Borland, R., Hammond, D., Swayampakala, K., & Thrasher, J. (2018). Which type of tobacco product warning imagery is more effective and sustainable over time? A longitudinal assessment of smokers in Canada, Australia and Mexico. BMJ Open, 8(7), e021983. doi:10.1136/bmjopen-2018-021983 Aparicio, H. J., Himali, J. J., Beiser, A. S., Davis-Plourde, K. L., Vasan, R. S., Kase, C. S., . . . Seshadri, S. (2017). Overweight, Obesity, and Survival After Stroke in the Framingham Heart Study. J Am Heart Assoc, 6(6). doi:10.1161/jaha.116.004721 Bacon, C. G., Mittleman, M. A., Kawachi, I., Giovannucci, E., Glasser, D. B., & Rimm, E. B. (2006). A prospective study of risk factors for erectile dysfunction. J Urol, 176(1), 217- 221. doi:10.1016/s0022-5347(06)00589-1 Bekalu, M. A., Ramanadhan, S., Bigman, C. A., Nagler, R. H., & Viswanath, K. (2019). Graphic and Arousing? Emotional and Cognitive Reactions to Tobacco Graphic Health Warnings and Associated Quit-Related Outcomes Among Low SEP Population Groups. Health Commun, 34(7), 726-734. doi:10.1080/10410236.2018.1434733 Bhat, A., Shilpashree, K. B., Krishnamurthy, A., Manjunath, C., Shwetha, R., & Madhusudhan, S. (2018). Adolescent's Perception about the Introduction of New Cigarette Packaging and Plain Packaging of Cigarette Packs: A Qualitative Study. Indian J Community Med, 43(Suppl 1), S47-s51. doi:10.4103/ijcm.IJCM_97_18 BRFSS. (2018). Behavioral Risk Factor Surveillance System (BRFSS) Prevalence & Trends Data, Centers for Disease Control and Prevention. Retrieved from https://www.cdc.gov/brfss/brfssprevalence/ Cho, Y. J., Thrasher, J. F., Swayampakala, K., Yong, H.-H., McKeever, R., Hammond, D., . . . Borland, R. (2016). Does Reactance against Cigarette Warning Labels Matter? Warning Label Responses and Downstream amongst Adult Smokers in Australia, Canada, Mexico and the United States. PLoS One, 11(7), e0159245-e0159245. doi:10.1371/journal.pone.0159245 Christen, W. G., Manson, J. E., Seddon, J. M., Glynn, R. J., Buring, J. E., Rosner, B., & Hennekens, C. H. (1992). A prospective study of cigarette smoking and risk of cataract in men. Jama, 268(8), 989-993. Cochran, J. R., Kydd, R. R., Lee, J. M. J., Walker, N., & Consedine, N. S. (2018). Disgust but not Health Anxiety Graphic Warning Labels Reduce Motivated Attention in Smokers: A Study of P300 and Late Positive Potential Responses. Nicotine Tob Res, 20(7), 819-826. doi:10.1093/ntr/ntx158 Colamesta, V., D'Aguanno, S., Breccia, M., Bruffa, S., Cartoni, C., & La Torre, G. (2016). Do the smoking intensity and duration, the years since quitting, the methodological quality and the year of publication of the studies affect the results of the meta-analysis on cigarette smoking and Acute Myeloid Leukemia (AML) in adults? Crit Rev Oncol Hematol, 99, 376-388. doi:10.1016/j.critrevonc.2016.01.003 Congdon, N., Vingerling, J. R., Klein, B. E., West, S., Friedman, D. S., Kempen, J., . . . Taylor, H. R. (2004). Prevalence of cataract and pseudophakia/aphakia among adults in the United States. Arch Ophthalmol, 122(4), 487-494. doi:10.1001/archopht.122.4.487 Cornuz, J., Sidoti Pinto, C., Tevaearai, H., & Egger, M. (2004). Risk factors for asymptomatic abdominal aortic aneurysm: systematic review and meta-analysis of population-based screening studies. Eur J Public Health, 14(4), 343-349. doi:10.1093/eurpub/14.4.343

68 Criqui, M. H., Langer, R. D., Fronek, A., Feigelson, H. S., Klauber, M. R., McCann, T. J., & Browner, D. (1992). Mortality over a period of 10 years in patients with peripheral arterial disease. N Engl J Med, 326(6), 381-386. doi:10.1056/nejm199202063260605 Di Giuseppe, D., Discacciati, A., Orsini, N., & Wolk, A. (2014). Cigarette smoking and risk of rheumatoid arthritis: a dose-response meta-analysis. Arthritis Res Ther, 16(2), R61. doi:10.1186/ar4498 Eke, P. I., Dye, B. A., Wei, L., Slade, G. D., Thornton-Evans, G. O., Borgnakke, W. S., . . . Genco, R. J. (2015). Update on Prevalence of Periodontitis in Adults in the United States: NHANES 2009 to 2012. Journal of periodontology, 86(5), 611-622. doi:10.1902/jop.2015.140520 Evans, A. T., Peters, E., Shoben, A. B., Meilleur, L. R., Klein, E. G., Tompkins, M. K., . . . Tusler, M. (2017). Cigarette Graphic Warning Labels Are Not Created Equal: They Can Increase or Decrease Smokers' Quit Intentions Relative to Text-Only Warnings. Nicotine Tob Res, 19(10), 1155-1162. doi:10.1093/ntr/ntw389 Foucher, P., Baudouin, N., Merati, M., Pitard, A., Bonniaud, P., Reybet-Degat, O., & Jeannin, L. (1998). Relative Survival Analysis of 252 Patients With COPD Receiving Long-term Oxygen Therapy. Chest, 113(6), 1580-1587. doi:10.1378/chest.113.6.1580 Freedman, N. D., Abnet, C. C., Leitzmann, M. F., Hollenbeck, A. R., & Schatzkin, A. (2007). Prospective investigation of the cigarette smoking-head and neck cancer association by sex. Cancer, 110(7), 1593-1601. doi:10.1002/cncr.22957 Gandini, S., Botteri, E., Iodice, S., Boniol, M., Lowenfels, A. B., Maisonneuve, P., & Boyle, P. (2008). and cancer: a meta-analysis. Int J Cancer, 122(1), 155-164. doi:10.1002/ijc.23033 Hackshaw, A., Morris, J. K., Boniface, S., Tang, J.-L., & Milenković, D. (2018). Low cigarette consumption and risk of coronary heart disease and stroke: meta-analysis of 141 cohort studies in 55 study reports. Bmj, 360, j5855. doi:10.1136/bmj.j5855 Hall, M. G., Mendel, J. R., Noar, S. M., & Brewer, N. T. (2018). Why smokers avoid cigarette pack risk messages: Two randomized clinical trials in the United States. Soc Sci Med, 213, 165-172. doi:10.1016/j.socscimed.2018.07.049 Hammond, D., McDonald, P. W., Fong, G. T., Brown, K. S., & Cameron, R. (2004). The impact of cigarette warning labels and smoke-free bylaws on smoking cessation: evidence from former smokers. Can J Public Health, 95(3), 201-204. Harris, P. R., Mayle, K., Mabbott, L., & Napper, L. (2007). Self-affirmation reduces smokers' defensiveness to graphic on-pack cigarette warning labels. Health Psychol, 26(4), 437- 446. doi:10.1037/0278-6133.26.4.437

69 Hirsch, A. T., Haskal, Z. J., Hertzer, N. R., Bakal, C. W., Creager, M. A., Halperin, J. L., . . . Riegel, B. (2006). ACC/AHA 2005 Practice Guidelines for the management of patients with peripheral arterial disease (lower extremity, renal, mesenteric, and abdominal aortic): a collaborative report from the American Association for Vascular Surgery/Society for Vascular Surgery, Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, Society of Interventional Radiology, and the ACC/AHA Task Force on Practice Guidelines (Writing Committee to Develop Guidelines for the Management of Patients With Peripheral Arterial Disease): endorsed by the American Association of Cardiovascular and Pulmonary Rehabilitation; National Heart, Lung, and Blood Institute; Society for Vascular Nursing; TransAtlantic Inter-Society Consensus; and Vascular Disease Foundation. Circulation, 113(11), e463- 654. doi:10.1161/circulationaha.106.174526 Huang, L. L., Lazard, A. J., Pepper, J. K., Noar, S. M., Ranney, L. M., & Goldstein, A. O. (2017). Impact of The Real Cost Campaign on Adolescents' Recall, Attitudes, and Risk Perceptions about Tobacco Use: A National Study. Int J Environ Res Public Health, 14(1). doi:10.3390/ijerph14010042 Jayes, L., Haslam, P. L., Gratziou, C. G., Powell, P., Britton, J., Vardavas, C., . . . Leonardi-Bee, J. (2016). SmokeHaz: Systematic Reviews and Meta-analyses of the Effects of Smoking on Respiratory Health. Chest, 150(1), 164-179. doi:10.1016/j.chest.2016.03.060 Kees, J., Burton, S., Andrews, J., & Kozup, J. (2010). Understanding How Graphic Pictorial Warnings Work on Cigarette Packaging. Journal of Public Policy & Marketing, 29, 265- 276. doi:10.2307/20798415 Kim, J. K., Baker, L. A., Davarian, S., & Crimmins, E. (2013). Oral health problems and mortality. Journal of dental sciences, 8(2), 10.1016/j.jds.2012.1012.1011. doi:10.1016/j.jds.2012.12.011 Leite, F. R. M., Nascimento, G. G., Scheutz, F., & López, R. (2018). Effect of Smoking on Periodontitis: A Systematic Review and Meta-regression. American Journal of Preventive Medicine, 54(6), 831-841. doi:https://doi.org/10.1016/j.amepre.2018.02.014 Leshner, G., Bolls, P., & Thomas, E. (2009). Scare' em or disgust 'em: the effects of graphic health promotion messages. Health Commun, 24(5), 447-458. doi:10.1080/10410230903023493 Lu, L., Mackay, D. F., & Pell, J. P. (2014). Meta-analysis of the association between cigarette smoking and peripheral arterial disease. Heart, 100(5), 414. doi:10.1136/heartjnl-2013- 304082 Mani, K., Bjorck, M., Lundkvist, J., & Wanhainen, A. (2009). Improved long-term survival after abdominal aortic aneurysm repair. Circulation, 120(3), 201-211. doi:10.1161/circulationaha.108.832774 McCloud, R. F., Okechukwu, C., Sorensen, G., & Viswanath, K. (2017). Cigarette graphic health warning labels and information avoidance among individuals from low socioeconomic position in the U.S. Cancer Causes Control, 28(4), 351-360. doi:10.1007/s10552-017- 0875-1 Mercincavage, M., Burdge, J., Lochbuehler, K., Souprountchouk, V., McCullough, A. A., & Strasser, A. A. (2018). Visual Attention Patterns Differ by Pictorial Health Warning Label Features. Tob Regul Sci, 4(6), 8-17. doi:10.18001/trs.4.6.2 Mukamal, K. J., Nesto, R. W., Cohen, M. C., Muller, J. E., Maclure, M., Sherwood, J. B., & Mittleman, M. A. (2001). Impact of diabetes on long-term survival after acute myocardial

70 infarction: comparability of risk with prior myocardial infarction. Diabetes Care, 24(8), 1422-1427. doi:10.2337/diacare.24.8.1422 National Eye Institute. (2019). Age-Related Macular Degeneration (AMD) Data and Statistics Retrieved from https://nei.nih.gov/learn-about-eye-health/resources-for-health- educators/eye-health-data-and-statistics/age-related-macular-degeneration-amd-data-and- statistics Ordonez-Mena, J. M., Schottker, B., Mons, U., Jenab, M., Freisling, H., Bueno-de-Mesquita, B., . . . Brenner, H. (2016). Quantification of the smoking-associated cancer risk with rate advancement periods: meta-analysis of individual participant data from cohorts of the CHANCES consortium. BMC Med, 14, 62. doi:10.1186/s12916-016-0607-5 Pagano, A., Gubner, N., Tajima, B., Yip, D., Henderson, C., & Guydish, J. (2017). Addiction Treatment Clients' Reactions to Graphic Warning Labels on Cigarette Packs. J Drug Issues, 47(3), 433-447. doi:10.1177/0022042617699196 Pan, A., Wang, Y., Talaei, M., Hu, F. B., & Wu, T. (2015). Relation of active, passive, and quitting smoking with incident type 2 diabetes: a systematic review and meta-analysis. Lancet Diabetes Endocrinol, 3(12), 958-967. doi:10.1016/s2213-8587(15)00316-2 Peng, K. A., Grogan, T., & Wang, M. B. (2014). Head and neck sarcomas: analysis of the SEER database. Otolaryngol Head Neck Surg, 151(4), 627-633. doi:10.1177/0194599814545747 Polen, K. N., Sandhu, P. K., Honein, M. A., Green, K. K., Berkowitz, J. M., Pace, J., & Rasmussen, S. A. (2015). Knowledge and attitudes of adults towards smoking in pregnancy: results from the HealthStyles(c) 2008 survey. Matern Child Health J, 19(1), 144-154. doi:10.1007/s10995-014-1505-0 Quisenberry, A. J., Scott, A. E., Shoben, A. B., Ferketich, A. K., Cooper, S. E., Berman, M., . . . Klein, E. G. (2018). Health Literacy and Attention to Cigarette Health Warning Labels among Rural Smokers. Tob Regul Sci, 4(6), 38-46. doi:10.18001/trs.4.6.4 Ringbaek, T., Seersholm, N., & Viskum, K. (2005). Standardised mortality rates in females and males with COPD and asthma. Eur Respir J, 25(5), 891-895. doi:10.1183/09031936.05.00099204 Robbins, A. S., Manson, J. E., Lee, I. M., Satterfield, S., & Hennekens, C. H. (1994). Cigarette smoking and stroke in a cohort of U.S. male physicians. Ann Intern Med, 120(6), 458- 462. doi:10.7326/0003-4819-120-6-199403150-00002 Roger, V. L., Go, A. S., Lloyd-Jones, D. M., Adams, R. J., Berry, J. D., Brown, T. M., . . . Wylie-Rosett, J. (2011). Heart disease and stroke statistics --2011 update: a report from the American Heart Association. Circulation, 123(4), e18-e209. doi:10.1161/CIR.0b013e3182009701 Rousu, M. C., & Thrasher, J. F. (2012). Demand Reduction from Plain and Pictorial Cigarette Warning Labels: Evidence from Experimental Auctions. Applied Economic Perspectives and Policy, 35(1), 171-184. doi:10.1093/aepp/pps049 SEER. (2018a). SEER [Surveillance, Epidemiology, and End Results Program] Cancer Statistics Review, 1975-2014, National Cancer Institute. Bethesda, MD. Retrieved from https://seer.cancer.gov/csr/1975_2014 SEER. (2018b). Table 1.5: Age-Adjusted SEER [Surveillance, Epidemiology, and End Results Program] Incidence and U.S. Death Rates and 5-Year Relative Survival (Percent) By Primary Cancer Site, Sex and Time Period Retrieved from https://seer.cancer.gov/archive/csr/1975_2014/results_merged/topic_survival.pdf

71 SEER. (2019). Cancer Stat Facts: Leukemia – Acute Myeloid Leukemia (AML) [Surveillance, Epidemiology, and End Results Program]. Retrieved from https://seer.cancer.gov/statfacts/html/amyl.html Selvin, E., Burnett, A. L., & Platz, E. A. (2007). Prevalence and risk factors for erectile dysfunction in the US. Am J Med, 120(2), 151-157. doi:10.1016/j.amjmed.2006.06.010 Shadel, W. G., Martino, S. C., Setodji, C. M., Dunbar, M., Scharf, D., & Creswell, K. G. (2019). Do graphic health warning labels on cigarette packages deter purchases at point-of-sale? An experiment with adult smokers. Health Educ Res, 34(3), 321-331. doi:10.1093/her/cyz011 Skurka, C., Kalaji, M., Dorf, M. C., Kemp, D., Greiner Safi, A., Byrne, S., . . . Niederdeppe, J. (2019). Independent or synergistic? Effects of varying size and using pictorial images in tobacco health warning labels. Drug and Alcohol Dependence, 198, 87-94. doi:https://doi.org/10.1016/j.drugalcdep.2019.01.034 Surgeon General (SG). (2004). Surgeon General's Report—The Health Consequences of Smoking. Retrieved from https://www.cdc.gov/tobacco/data_statistics/sgr/2004/index.htm Sussenbach, P., Niemeier, S., & Glock, S. (2013). Effects of and attention to graphic warning labels on cigarette packages. Psychol Health, 28(10), 1192-1206. doi:10.1080/08870446.2013.799161 Tan, A. S. L., Bigman, C. A., Nagler, R. H., Minsky, S., & Viswanath, K. (2017). Comparing perceived effectiveness of FDA-proposed cigarette packaging graphic health warnings between sexual and gender minorities and heterosexual adults. Cancer Causes Control, 28(10), 1143-1155. doi:10.1007/s10552-017-0954-3 Thornton, J., Edwards, R., Mitchell, P., Harrison, R. A., Buchan, I., & Kelly, S. P. (2005). Smoking and age-related macular degeneration: a review of association. Eye, 19(9), 935- 944. doi:10.1038/sj.eye.6701978 Thrasher, J. F., Swayampakala, K., Borland, R., Nagelhout, G., Yong, H. H., Hammond, D., . . . Hardin, J. (2016). Influences of Self-Efficacy, Response Efficacy, and Reactance on Responses to Cigarette Health Warnings: A Longitudinal Study of Adult Smokers in Australia and Canada. Health Commun, 31(12), 1517-1526. doi:10.1080/10410236.2015.1089456 Van Dessel, P., Smith, C. T., & De Houwer, J. (2018). Graphic cigarette pack warnings do not produce more negative implicit evaluations of smoking compared to text-only warnings. PLoS One, 13(3), e0194627. doi:10.1371/journal.pone.0194627 van Osch, F. H., Jochems, S. H., van Schooten, F.-J., Bryan, R. T., & Zeegers, M. P. (2016). Quantified relations between exposure to tobacco smoking and bladder cancer risk: a meta-analysis of 89 observational studies. Int J Epidemiol, 45(3), 857-870. doi:10.1093/ije/dyw044

72

Attachment 1: CTP Technical Project Lead Review (TPL) Review of IQOS PMTA: PM0000424, PM0000425, PM0000426, and PM0000479 (April 29, 2019) (“IQOS® TPL Review”)

Technical Project Lead (TPL):

Digitally signed by Priscilla Callahan-lyon -S Date: 2019.04.29 15:13:25 -04'00'

Priscilla Callahan-Lyon, M.D. Deputy Director Division of Individual Health Science

Signatory Decision:

X Concur with TPL recommendation and basis of recommendation ܆ Concur with TPL recommendation with additional comments (see separate memo) ܆ Do not concur with TPL recommendation (see separate memo)

Digitally signed by Matthew R. Holman -S Date: 2019.04.29 16:46:36 -04'00'

Matthew R. Holman, Ph.D. Director Office of Science

5 TABLE OF CONTENTS

I. EXECUTIVE SUMMARY ...... 10

II. REVIEW OF PMTA ...... 14

A. Regulatory History ...... 14

B. Product Description: Engineering, Chemistry, Stability, and Manufacturing ...... 14 1. General ...... 14 2. Heatsticks ...... 16 a. Tobacco Ingredients ...... 16 b. Non-tobacco Ingredients ...... 17 c. Manufacturing, Process, and Controls ...... 19 d. Quality Assurance/Sample Testing ...... 19 e. Product Stability ...... 22 3. Heating System ...... 25 a. Holder ...... 25 b. Charger ...... 25 c. Manufacturing, Process, and Controls ...... 25 d. Quality Assurance/Sample Testing ...... 26 4. Inspections of Manufacturing Facilities ...... 27 5. Summary of Engineering, Chemistry, Product Stability, and Manufacturing Findings ...... 29

C. Toxicological Risk Assessment ...... 30 1. Harmful and Potentially Harmful Constituents (HPHCs) ...... 30 a. General Overview ...... 30 b. Constituents Unique to IQOS ...... 32 c. Comparison to Cigarette Smoke ...... 33 d. Environmental Exposure from Heatstick Aerosol ...... 35 2. In Vitro Studies ...... 35 a. Neutral Red Uptake (NRU) Assays ...... 35 b. Bacterial Reverse Mutation (Ames) test ...... 36 c. Mouse Lymphoma Assay (MLA) ...... 36 d. Summary of In Vitro Studies ...... 37 3. In Vivo Studies ...... 37 a. 90-day Nose-only Inhalation Studies ...... 37 b. 18-month Carcinogenicity Study with A/J mice ...... 38 c. Nicotine Pharmacokinetic (PK) Study with Rats ...... 38 d. Summary ...... 38 4. Systems Toxicology Studies ...... 39 a. Acute and Repeated Exposure Studies with Human Organotypic Tissues ...... 39 b. ApoE-/- Mouse Switching Study ...... 39 5. Nonclinical Evaluation of Carbon Monoxide from Heatstick Aerosol ...... 40 6. Summary of Toxicological Findings ...... 41

6

D. Behavioral and Clinical Pharmacological Assessment ...... 42 1. Pharmacokinetics, Exposure/Response and Clinical Pharmacology ...... 42 a. Pharmacokinetic/Pharmacodynamic (PK/PD) Studies ...... 42 b. Intrinsic and Extrinsic Factors Affecting Nicotine Pharmacokinetics ...... 44 c. Nicotine Equivalents (NEQ) in Urine ...... 44 d. Summary ...... 45 2. Behavioral Pharmacology...... 45 a. Use Behavior and Topography ...... 45 b. Product Use/Consumption...... 45 c. Product Acceptability ...... 47 d. Abuse Liability ...... 47 e. Summary ...... 49 3. Summary of Behavioral and Clinical Pharmacology Findings ...... 49

E. Individual Health Impact ...... 50 1. Biomarkers of Exposure (BOE) ...... 50 a. COHb – BOE for Carbon Monoxide ...... 53 b. Monohydroxybutenyl-mercapturic acid (MHBMA) – BOE for 1,3-Butadiene ...... 53 c. 3-Hydroxypropylmercapturic acid (3-HPMA) – BOE for ...... 54 d. S-phenylmercapturic acid (S-PMA) – BOE for Benzene ...... 54 e. Total NNAL - BOE for NNK ...... 54 f. CYP1A2 Activity ...... 55 g. Urine Mutagenicity ...... 55 h. Summary ...... 55 2. Biomarkers of Potential Harm (BOPH) ...... 56 a. Assessment of Inflammation ...... 57 b. Assessment of Oxidative Stress ...... 57 c. Assessment of Cardiovascular (CV) Risks ...... 57 d. Assessment of Lung Function...... 57 e. Assessment of Genotoxicity and Mutagenicity ...... 58 f. Summary ...... 58 3. Clinical Effects of IQOS ...... 59 a. Analysis of Adverse Events in Clinical Studies ...... 59 b. Review of Published Clinical Literature ...... 60 c. Summary ...... 60 4. Likelihood of Product Misuse or Malfunction ...... 61 5. Bioresearch Monitoring (BIMO) Inspection ...... 62 6. 2017 Safety Update Report ...... 62 7. Summary of Individual Health Findings from Clinical Studies, Literature, Adverse Experience Reports, and Safety Updates ...... 64

F. Population Health ...... 65 1. Likelihood of IQOS Use by Current Cigarette Smokers ...... 68 2. Poly-use of IQOS and cigarettes or other tobacco products ...... 72 3. Use of IQOS by Former or Never Smokers and Youth ...... 73 4. Likelihood of IQOS leading to Conventional Cigarette Smoking Cessation ...... 76

7

5. Population Modeling ...... 77 6. Amendment MR0000117; Study ZRHR-ERS-09-US ...... 79 7. Summary of Population Health Findings ...... 83

III. PRODUCT LABELING, CONSUMER COMPREHENSION, AND MARKETING PLAN ...... 85

A. Proposed PMTA Labeling ...... 85

B. Consumer Comprehension ...... 85

C. Marketing Plan ...... 86

C. Nicotine is Addictive Labeling ...... 87

D. Carbon Monoxide (CO) Warning ...... 89

IV. CONCLUSIONS AND RECOMMENDATIONS ...... 90

A. Recommendation for Marketing ...... 98

B. Postmarketing Recommendations ...... 99

C. Postmarketing Recordkeeping, Retention, Reporting and Marketing Requirements ...... 99 1. Record Retention ...... 100 2. Serious and Unexpected Adverse Experiences Reporting ...... 102 3. Manufacturing Deviations ...... 103 4. Periodic Reporting ...... 103 5. Marketing Requirements ...... 107

APPENDIX: THE PUBLIC HEALTH RATIONALE FOR RECOMMENDED RESTRICTIONS ON NEW TOBACCO PRODUCT LABELING, ADVERTISING, MARKETING, AND PROMOTION ...... 111

8

Table 1: Select HPHCs in Aerosol and Filler as Tested by Southeast Tobacco Laboratory (STL) ...... 20 Table 2: Comparison of HPHC Yields from Heatstick Aerosols and Combusted Cigarettes ...... 34 Table 3: Selected HPHCs, Chemical Class, Measured BOE, and Toxicity Class ...... 51 Table 4: Summary of Device Issues in IQOS Clinical Trials ...... 61 Table 5: Summary of Pre- and Post-Market Studies of IQOS Use Patterns Among Current Tobacco Users 66 Table 6: Intention to Try and Use IQOS among Current Smokers (PBA-05-NOC) ...... 69 Table 7: Prevalence of Heatstick Trial and Switching at Study End, by Country ...... 70 Table 8: IQOS and CC Use Over 6-Week Observational Actual Use Study (PBA-07) ...... 72 Table 9: Study Population ZRHR-ERS-09-US ...... 80 Table 10: Primary Analysis of CREs between THS Use and CC Use at Six Months for ZRHR-ERS-09-US ...... 81 Table 11: CO Acceptance Criteria for Heatsticks ...... 90

Figure 1: The IQOS Charger, Holder, and Heatstick ...... 15 Figure 2: Heatstick Components ...... 16 Figure 3: Carboxyhemoglobin Levels after Repeated Exposure to 3R4F, ...... 40 Figure 4: Plasma Carboxyhemoglobin Levels at Week 6 for 90-day Rat Inhalation Study ...... 41 Figure 5: Product Consumption (Number of Products Used) During the REX Studies...... 46 Figure 6: Arithmetic Mean and 95% CI Total Scores for QSU-Brief Total ...... 48 Figure 7: Percent Change from Baseline of Geometric Mean Levels and 95% CIs at Day 90 in ZRHM-REXA- 08-US ...... 52 Figure 8: Geometric Mean and 95% CIs of COHb Concentrations (%) ...... 53 Figure 9: Likelihood of Heatstick Purchase among Current Smokers who used at least 100 Heatsticks in PBA-07 Actual Use Study...... 71 Figure 10: THS:CC (light grey) and Dual Use:CC (dark grey) Ratios (%) and Difference (ppm) and 95% CI at Month 6 ...... 81 Figure 11: Perceived Addiction Risk for Heatsticks Pack Arm 2 ...... 88

9

I. Executive Summary On May 15, 2017, FDA received PMTAs for the IQOS Tobacco Heating System (THS)4 with Marlboro Heatsticks, Smooth Menthol Heatsticks, and Fresh Menthol Heatsticks from Philip Morris Products S.A. (PMP S.A. or the applicant). PMP S.A.’s parent company, Philip Morris International Management S.A. (PMI) has entered into a distribution agreement with Altria Client Services LLC (ALCS) by which ALCS and an ALCS affiliate, Philip Morris USA Inc. (PM USA), will be licensed to distribute and sell the IQOS system and the Marlboro Heatsticks in the U.S. upon receipt of a marketing authorization.5

The THS consists of three main components: 1. The IQOS Heatstick: a tobacco plug consisting of crimped cast reconstituted tobacco sheet made from ground tobacco powder. Three different Heatsticks will be available - Regular, Smooth Menthol and Fresh Menthol. Heatsticks will be marketed under the Marlboro brand in packs of 20. 2. The IQOS Holder: an electrically powered and rechargeable unit designed to hold and heat the Heatsticks during consumer use to generate the nicotine-containing aerosol. 3. The IQOS Charger: used to recharge the Holder after each use. The Charger stores sufficient energy for the use of approximately 20 Heatsticks before requiring recharging itself. It can be recharged from household power.

A new tobacco product, including a tobacco product modified in any way (“including a change in design, any component, any part, or any constituent, including a smoke constituent, or in the content, delivery, or form of nicotine, or any other additive or ingredient” after February 15, 2007 (section 910(a)(1)(B)), generally requires premarket review and an order from FDA authorizing the marketing of the product (section 910(a)(2)(A)).

A PMTA must be submitted to FDA under section 910(b) of the FD&C Act and a marketing authorization order must be received from FDA under section 910(c)(1)(A)(i) prior to marketing any new tobacco product, unless FDA has found that the new tobacco product is substantially equivalent to a tobacco product commercially marketed in the US as of February 15, 2007 (see section 910(a)(2)(A)(i)) or is exempt from a substantial equivalence determination pursuant to regulation (see section 910(a)(2)(A)(ii)).

FDA will deny a PMTA and issue a no marketing authorization order that the product may not be introduced or delivered for introduction into interstate commerce under section 910(c)(1)(A)(ii) where FDA finds that: • there is a lack of a showing that permitting the product to be marketed would be appropriate for the protection of the public health; • the methods, facilities, or controls used in manufacturing, processing, or packing do not conform to manufacturing regulations issued under section 906(e) (21 U.S.C. 387f(e)); • based on a fair evaluation of all material facts, the proposed labeling is false or misleading; or • it is not shown that the product complies with any tobacco product standard in effect under section 907 (21 U.S.C. 387g), and there is not adequate information to justify deviation from the standard.

4 Throughout the remainder of this review, the Tobacco Heating System will be referred to as either “THS” or “IQOS” and the tobacco sticks will be referred to as the “Heatsticks.” Unless otherwise designated, the terms THS, THS 2.2, and IQOS refer to the same thing; mTHS 2.2 refers to mentholated Heatsticks. 5 Altria Client Services LLC is a wholly-owned subsidiary of Altria Group, Inc. and provides certain services to the Altria family of companies. PM USA is not part of Philip Morris International group of companies.

10

The statute provides that the finding as to whether the marketing of a product for which a PMTA is submitted would be appropriate for the protection of the public health shall be determined with respect to the risks and benefits to the population as a whole, including users and nonusers of the tobacco product, and taking into account ─ (A) the increased or decreased likelihood that existing users of tobacco products will stop using such products; and (B) the increased or decreased likelihood that those who do not use tobacco products will start using such products.

Scientific review of these applications has demonstrated the following: • There are adequate process controls and quality assurance procedures to help ensure the IQOS Holder, IQOS Charger, Marlboro Heatsticks, Fresh Menthol Heatsticks, and Smooth Menthol Heatstick are manufactured consistently to meet the applicant’s specifications. • Marlboro, Smooth Menthol, and Fresh Menthol Heatstick aerosols contain some chemicals which are different from those found in combusted cigarettes (CC).6 Although some of the chemicals are genotoxic or cytotoxic, these chemicals are present in very low levels and potential effects are outweighed by the substantial decrease in the number and levels of HPHCs found in CC (see below). • The toxicological profiles of Marlboro, Smooth Menthol, and Fresh Menthol Heatsticks are essentially identical except for the quantity of menthol. The available toxicological data indicates the potential for a relative benefit compared to CC for smokers who switch completely to IQOS. • Smooth Menthol Heatsticks contain 6.98 mg menthol/Heatstick. Fresh Menthol Heatsticks contain 13.23 mg menthol/Heatstick. The applicant compared this to 23 mentholated cigarette brands in the U.S. which had 2.9-19.5 mg menthol/cigarette. • PK studies show Marlboro, Smooth Menthol, and Fresh Menthol Heatsticks have nicotine delivery, addiction potential, and abuse liability similar to CC. This is potentially beneficial for smokers trying to switch to IQOS as they are more likely to have satisfactory results and not resume CC smoking. The nicotine levels do pose an addiction risk for non-tobacco users who initiate use of these products; however, the risk is no higher than for other, currently available, tobacco products and initiation is expected to be low generally. (See also the discussion regarding the inclusion of a nicotine addiction warning below.) • The 5-day studies demonstrate improved biomarkers of exposure (BOE) which indicates reduced HPHC exposures. These improvement trends persisted in the 90-day studies despite reduced compliance and use of other tobacco products. Additionally, the applicant recently submitted data from a six-month which demonstrated reduction in eight BOE as well as NNAL and COHb for self-reported users of IQOS compared to CC users. Although the studies conducted by the applicant do not demonstrate reduction in long-term disease risks, the currently available evidence indicates CC smokers who switch completely to IQOS will have reduced toxic exposures and this is likely to lead to less risk of tobacco-related diseases. The data for CC smokers who use IQOS while continuing to smoke (dual use) is less clear but the available evidence shows no increase in HPHC exposures for those who dual use. • There have been no specific, short-term health-related or product quality issues unique to IQOS in the clinical studies, the current world-wide markets, or the published literature. • Misuse of IQOS is uncommon and the product design makes it unlikely users will have a satisfactory experience (e.g., no significant nicotine is delivered with reusing a Heatstick). • Dual use of IQOS and CC was common in all countries in the pre- and post-market studies though the CC users in the U.S. actual use study who switched to exclusive IQOS use during the study remained

6 For the purposes of this review CC=combusted cigarette(s) or conventional cigarette(s)

11

generally stable during the 6-week observational period. Individuals who initiate IQOS and maintain exclusive IQOS use over time can potentially replace their use of CC with Heatsticks long-term. The toxicological and clinical studies do not show an increase in HPHC exposures when consumers are using both IQOS and CC and, although the decreases are not statistically significant, some HPHC exposures appear to be decreased. • Although the data for IQOS uptake by never smokers, former smokers, and youth is limited, there are some data from countries where IQOS is marketed - Italy and Japan - which show low uptake by youth and current nonsmokers. In these countries, the likelihood of uptake is slightly higher in former smokers, but still low. Appropriately, the population most likely to use IQOS are current CC smokers. The proposed marketing and advertising restrictions will help ensure lower youth exposure and access to the products. Additionally, the applicant will be required to monitor consumer use patterns and demographic information and provide FDA with regular reports.

As discussed in more detail in Sections III C, III D, and IV F of this review, I recommend the PMTAs be authorized subject to the following changes to the proposed product labeling and advertising for IQOS: 1. Inclusion of the warning: “WARNING: This product contains nicotine. Nicotine is an addictive chemical.” on the package labels of all Heatsticks packs and of all kits containing Heatsticks packs as well as in all advertisements for such products and kits. Data shows that consumers do not accurately perceive the addiction risks of IQOS. Permitting IQOS to be marketed without this warning would not be appropriate for protection of public health. 2. Removal of the warning: “SURGEON GENERAL’S WARNING: Cigarette Smoke Contains Carbon Monoxide.” from the required warnings to be displayed on the product package labels and advertisements under FCLAA. Based on a fair evaluation of all material facts, the warning is misleading with respect to these products which, although categorized as cigarettes, do not produce carbon monoxide above environmental levels and do not increase CO-related health risks.

In conclusion, none of the grounds specified in Section 910(c)(2) of the FD&C Act apply. Specifically, I find the following: 1. Permitting the marketing of the products is appropriate for the protection of the public health, as described in Section 910(c)(4) of the FD&C Act (subject to the labeling and advertising changes described above); 2. The methods used in, and the facilities or controls used for, the manufacture, processing, and packing of these products do not fail to conform to the requirements in 906(e);7 3. Based on a fair evaluation of all material facts, the labeling (when subject to the changes described above) is not false or misleading in any particular; and 4. The products do not fail to conform to a tobacco product standard in effect under Section 907 of the FD&C Act.

7 FDA has not yet promulgated any regulations under Section 906(e) of the FD&C Act.

12

I recommend FDA grant marketing authorization for the products described in the STNs, subject to the changes to the products’ package labels and advertisements, as described above: 1. PM0000424: Marlboro Heatsticks 2. PM0000425: Marlboro Smooth Menthol Heatsticks 3. PM0000426: Marlboro Fresh Menthol Heatsticks 4. PM0000479: IQOS System Holder and Charger8

8 Originally, FDA assigned the STNs as: PM0000424 - IQOS System with Marlboro Heatsticks, PM0000425 - IQOS System with Marlboro Smooth Menthol Heatsticks, and PM0000426 – IQOS System with Marlboro Fresh Menthol Heatsticks. For administrative convenience, a decision was made to change the STNs and assign the Heatsticks to PM0000424, PM0000425, and PM0000426 and the IQOS System Holder and Charger to a separate STN; PM0000479. (See memo dated February 19, 2019.)

13

II. Review of PMTA A. Regulatory History On May 15, 2017, FDA received PMTAs for the IQOS system including the IQOS Holder and Charger with three Heatsticks: Marlboro, Smooth Menthol, and Fresh Menthol from Philip Morris Products S.A. (PMP S.A. or the applicant). The applications were accepted and acknowledged June 14, 2017 and filed for scientific review on August 4, 2017. As per agreement with FDA, PMP S.A. included only PMTA-specific information in these submissions and cross-referenced other pertinent materials contained in the MRTPAs (MR000059-61) submitted for the products with modified risk information.

There have been several amendments submitted for the applications including applicant responses to FDA information requests, confirmation of laboratory samples, and responses to clarifying questions. In addition, amendments have been submitted to the MRTPAs that have been reviewed as part of the PMTA process, e.g., a safety update and a recently completed clinical study. All relevant information submitted to the agency, including information from the MRTPAs, the TPSAC meeting on the MRTPAs and the public comments to the MRTPAs, to the extent relevant to the PMTAs, has been considered in review of these applications.

The new tobacco products that are the subjects of the PMTAs include the IQOS Holder and Charger, and three different Heatstick packs: Marlboro Heatsticks (non-mentholated), Marlboro Smooth Menthol Heatsticks (1.35 mg menthol in smoke/ stick)9 and Marlboro Fresh Menthol Heatsticks (2.3 mg menthol in smoke/ stick).10 Throughout the remainder of this review, unless the products are specifically designated, the Marlboro Heatsticks and general discussion will refer to Heatsticks and mentholated Heatsticks includes both Smooth Menthol and Fresh Menthol Heatsticks.

Review Format The applicant provided information on each new tobacco product included in this review. IQOS is the commercial name of the Tobacco Heating System (THS), which includes a tobacco heating device (THD) and tobacco sticks. Throughout the remainder of this review, the Tobacco Heating System will be referred to as either “THS” or “IQOS” and the tobacco sticks will be referred to as the “Heatsticks.” Unless otherwise designated, the terms THS, THS 2.2, and IQOS refer to the same thing. Mentholated Heatsticks are designated mTHS 2.2. Section IV summarizes the technical project lead’s conclusions and recommendations for these applications.

B. Product Description: Engineering, Chemistry, Stability, and Manufacturing 1. General IQOS is the commercial name of the Tobacco Heating System (THS), which includes a THD with Holder and Charger and Heatsticks: Marlboro Heatsticks, Marlboro Smooth Menthol Heatsticks, and Marlboro Fresh Menthol Heatsticks. In response to FDA’s request for clarification, the applicant provided additional information about the product development history and the evolution of product naming. The applicant states ZRH, P1, and THS 2.2 all refer to the same tobacco heating system. THD 2.2 was the developmental device; the planned commercial device is THD 2.4. The proposed products, also known as IQOS or THS 2.2, uses THD 2.4. All products tested in the reduced exposure studies (REX) and most toxicology studies correspond to THS 2.2. The applicant made changes to the (b) (4) during product development and

9 Target level in aerosol, using the Canadian Intense Smoking Regime. 10 Target level in aerosol, using the Canadian Intense Smoking Regime.

14

(b) (4) . The applicant changed to (b) (4) and has provided bridging information to support that the products that are the subject of these applications are comparable to the study products.

The applicant confirmed that THD 2.4 (b) (4) ), which reflects certain modifications as part of continuous product development (See Section II.B.3.c), is the device that is the subject of these applications. This device, including these modifications, has been evaluated by FDA.

The IQOS Tobacco Heating System (referred to in this document as IQOS or THS) consists of three main components (see Figure 1 below): • The IQOS Heatstick, which contains a tobacco plug consisting of crimped cast reconstituted tobacco sheet made from ground tobacco powder. There are three different Heatsticks –Regular, Smooth Menthol and Fresh Menthol. Approximately half the length of a conventional cigarette, the Heatsticks will be marketed under the Marlboro brand in packs of 20. Heatsticks are designed to be electrically heated to release nicotine-containing aerosol and are not intended to be combusted. Heatsticks use reconstituted tobacco blended with glycerin to allow the generation of aerosol. HeatSticks are not designed or intended to be reused; the glycerin, which generates the aerosol, is depleted after one use. Heatsticks are filtered non-combusted cigarettes. • The Holder is an electrically-powered and rechargeable unit designed to hold and heat the Heatsticks during consumer use to generate the nicotine-containing aerosol. (b) (4)

This heating blade is inserted into a Heatstick to heat the tobacco. The user activates the Holder by pressing the activation button for a set period until the light begins to blink, signaling that the product may be used. It is designed to be used for a single Heatstick (a period of use of 6-7 minutes or 12-14 puffs) after which the Holder requires recharging and the used Heatstick is discarded. The Holder is electronically controlled to maintain a specific temperature range that allows generation of aerosol and prevents reaching temperatures where combustion can occur. • The IQOS Charger is used to recharge the Holder after each use. The Charger stores sufficient energy for the use of approximately 20 Heatsticks before requiring recharging itself. It can be recharged from household power. The Charger is designed to initiate and control the automatic cleaning cycle of the Holder heating blade at regular intervals; (b) (4) (b) (4) Cleaning (b) (4) (b) (4) and can only be performed when the Holder is securely held in the Charger, preventing the user from using the cleaning mode to heat a tobacco stick. The Charger electronically monitors and manages the Holder battery as well as the Charger battery.

Figure 1: The IQOS Charger, Holder, and Heatstick

15

To operate the THS, the user inserts a Heatstick into the IQOS Holder and turns on the device, which initiates the heating of the tobacco via the heating blade inserted into the tobacco plug. The Heatstick is not designed or intended to ignite or burn. The applicant states the electronically controlled heating at a set temperature range, in combination with the uniquely processed tobacco, prevents combustion from occurring. The temperature of the heating blade is controlled and the energy supply to the blade is cut off if its operating temperature exceeds 350°C. The temperature measured in the tobacco plug is designed to not exceed 300 °C.

2. Heatsticks Heatsticks consist of a tobacco plug and a non-tobacco component. Heatsticks do not contain tobacco cut- filler (tobacco leaf cut in small pieces found in CC); instead, the tobacco is ground and reconstituted into sheets (termed cast-leaf) following the addition of water, glycerin, guar gum and cellulose fibers. The Heatstick contains smaller amounts of tobacco than a CC. The weight of the tobacco plug in the Heatstick is approximately 320 mg compared with the 550-700 mg of cut-filler found in CC. The reconstituted tobacco cast-leaf is fashioned into a small plug through “crimping” that allows aerosol to flow through the tobacco plug during heating. The tobacco plug portion is composed of crimped cast tobacco sheet made from ground tobacco powder, , and flavorings.

The non-tobacco component includes a hollow tube (HAT), polylactic acid (PLA) filter, mouth piece filter (MPF), outer paper, and tipping paper. Unlike a conventional cigarette, the Heatstick contains two independent filters: (1) a polymer-film filter to cool the aerosol and (2) a low-density cellulose acetate filter that functions as a mouthpiece. In addition, a hollow acetate tube separates the tobacco plug and the polymer-film filter to prevent contact with the heating blade during use. Various papers are used to hold the Heatstick together. The plugs are individually wrapped with a plug wrap paper. The tobacco plug, the HAT and the PLA filter are held together with a cigarette paper and attached to the MPF using a tipping paper. Although typical cigarette papers and wraps are used in the construction of the Heatstick, they only serve as structural components and do not have any functionality as they would in a CC.

Figure 2: Heatstick Components Source: MR0000059-61, Section 3.1, Figure 3

a. Tobacco Ingredients The tobacco in the three Heatsticks includes (b) (4) (b) (4) blend types. In comparison, the Kentucky reference cigarette 3R4F includes flue-cured (35%), burley (22%), oriental (12%), (1%), and reconstituted (30%) tobacco blend types. The mainstream smoke of cigarettes made solely from reconstituted tobacco can produce high levels of carbon monoxide, nitrogen oxides, and tobacco-specific (TSNAs) during

16

combustion. The applicant included a description of the tobacco by (b) (4) , (b) (4) of the tobacco in Section 3.2.2.3.2 of the MRTPAs.

To maintain the blend characteristics over time, each individual tobacco lot is analyzed (b) (4) (b) (4) ) to ensure consistency and comparability. The total amount of tobacco b in each of the three Heatsticks is(b) (4) mg. This is(b) (4) less than the mass of tobacco in the Kentucky reference cigarette 3R4F (760 mg/cigarette).

b. Non-tobacco Ingredients Section 3.1.3 of the MRTPAs lists some of the non-tobacco ingredients included in the Heatsticks. Based on this information, the Heatsticks do not include any preservatives, which are frequently added to prevent undesirable microbial growth. This is discussed in Section II.B.2.e of this review. The applicant stated that the detailed list of ingredients and their quantities are commercially sensitive and were submitted via a TPMF (MF0000278) on November 30, 2017. The TPMF has been reviewed and found to include sufficient information regarding the tobacco blend in PM0000424 – PM0000426.

In the three Heatsticks, (52.3 mg/Heatstick) is 26% of the total weight of the tobacco in the Heatstick compared to levels of 1-5% typically added to tobacco in CC.11 In the three Heatstick products, (~2 mg/Heatstick) constitutes 1% of the total tobacco weight. Glycerol degradation produces mainly glycidol and acrolein, while propylene glycol degradation produces acetol and 2-propen-1- ol. Both glycerol and propylene glycol produce , which could increase acrolein generation by IQOS systems with Heatsticks compared to CC; however, the applicant provides data to show this does not occur. (See Section II.C.1.c).12

(b) (4)

A study of 48 mentholated cigarette brands available in the U.S. market between 2002 and 2003 includes a menthol range of 1.61 to 4.38 mg/cigarette.13 In addition, an Altria Client Sciences report submitted to the Tobacco Products Scientific Advisory Committee in 2010, includes a menthol range of 2.2 to 9.8 mg/cigarette.14 The total amount of menthol in PM0000426 (13.23 mg/Heatstick) is 35% higher than the upper limit of menthol reported in the U.S. market for combusted cigarettes (9.8 mg/cigarette).

Triacetin is included in the hollow acetate tube (10.7 mg) and in the mouth piece filter (2.22 mg) of the three Heatsticks. In CC, triacetin can increase the menthol amounts captured in the filter due to a change in filter efficiency and smoke transfer, and thus, affect menthol yield in mainstream smoke.15 The three Heatsticks

11 Carmines, E. and Gaworski, C. Toxicological evaluation of glycerin as a cigarette ingredient. Food and Chem Toxicol. 2005, 43, 1521-1539. 12 Sleiman, M.; Logue, J.; Montesinos, V.; Russell, M. et al. Emissions from Electronic Cigarettes: Key Parameters Affecting the Release of Harmful Chemicals. Environ. Sci. Technol. 2016, 50, 9644−9651. 13 Celebucki, C.; Ferris Wayne, G.; Connolly, G.; Pankow, J; Chang, E. Nicotine & Tob. Res. 2005, 7 (4), 523–531. 14 Altria Client Services. Background Information to Tobacco Products Scientific Advisory Committee, Menthol Discussion, 2010. 15 Wilson SA. Theoretical aspect of menthol migration and transfer. Recent Advances in Tobacco Science. 1993; 19, 129-153.

17

include 7.85 mg/stick of guar gum in the tobacco blend. In CC guar gum produces formaldehyde, benzo[a]pyrene, benzene, , and styrene.16

The three Heatsticks include cellulose in the tobacco (5.23 mg/Heatstick), wrap papers (4-45 mg/Heatstick), outer paper (14-23 mg/Heatstick), mouth piece filter (27 mg/Heatstick), tipping paper (12 mg/Heatstick), and cellulose acetate in the hollow acetate tube (56 mg/Heatstick). In addition, PM0000425 and PM0000426 include 20.9 mg of cellulose acetate in the polylactic acid filter. Thermal degradation of carbohydrates such as cellulose, pectins, starch, and sugars produce polyaromatic hydrocarbons (PAHs), phenols, , and ketones.17

The three Heatsticks include 214.8 mg/Heatstick of polylactic resin in the polylactic acid filter. The polylactic acid is biodegradable and the main degradation product is lactic acid.18 No harmful and potentially harmful constituents (HPHCs) are known to increase due to the presence of polylactic resin or lactic acid in the mainstream smoke of cigarettes.

The three Heatsticks include the copolymer ethylene-vinyl acetate in the outer paper adhesive (2.98 mg/Heatstick) and in the tipping paper adhesive (7.26 mg/Heatstick). Copolymer ethylene-vinyl acetate is also present at ~0.35 mg/Heatstick in the tobacco plug, polylactic acid filter, and mouth piece filter of the three Heatsticks. The copolymer ethylene-vinyl acetate decomposes at temperatures above 230°C19 to produce straight-chain hydrocarbon products.20 No HPHCs are known to increase due to the presence of straight-chain hydrocarbon products in the mainstream smoke of cigarettes.

The three Heatsticks include titanium dioxide in the tipping paper (1.34 mg/Heatstick). There is no significant difference in , nicotine and carbon monoxide (TNCO) between cigarettes using a filter with or without titanium dioxide.21 The three Heatsticks also include kaolin (0.39-3.00 mg/Heatstick) in the tipping paper, polylactic acid filter plug wrap paper, and mouth piece filter plug wrap paper. No HPHCs are known to increase due to the presence of kaolin in the mainstream smoke of cigarettes.

The three Heatsticks include calcium carbonate (0.7-19 mg/Heatstick) in the tobacco plug wrap paper, outer paper, hollow acetate tube plug wrap paper, polylactic acid filter, and tipping paper. A search of documents and patents indicates that calcium carbonate is added to CC to reduce side-stream smoke visibility.22 Calcium carbonate, in combination with alkali citrates, or ammonium phosphates, regulates the porosity of the cigarette paper.23 Higher permeability of the cigarette paper

16 Nair, U. Fact sheet on the tobacco additive guar gum created by the German Cancer Research Center (DKFZ), 2012, Heidelberg, Germany. 17 Rodgman, A. and Perfetti, T. The Chemicals Components of Tobacco and Tobacco Smoke, CRC Press, 2013, 1325. 18 Jamshidian, M.; Tehrany, E.; Imran, M.; Jacquot, M. et al. Poly-Lactic Acid: Production, Applications, Nanocomposites, and Release Studies. Comprehensive Reviews in Food Science and Food Safety. 2010, 552 – 571. 19 Beckman Sundh, U.; Binderup, M.-L.; Bolognesi, C.; Brimer, L.; Castle, L. et al. Scientific Opinion on the safety assessment of the substance ethylene-vinyl acetate copolymer wax, CAS No 24937-78-8 for use in food contact materials. EFSA Journal 2014, 12(2):3555. 20 McGrattan, B. Decomposition of Ethylene—Vinyl Acetate Copolymers Examined by Combined Thermogravimetry, Gas Chromatography, and Infrared Spectroscopy. Chapter 8. Hyphenated Techniques in Polymer Characterization, 1994, 103-115. 21 Wendeborn R.; Leutner T. Bull. Spec. CORESTA Symposium, Kallithea, 1990, 196, T09. 22 Connolly, G.; Wayne, G.; Lymperis, D.; Doherty, M. How cigarette additives are used to mask environmental tobacco smoke. Tobacco Control, 2000, 9:283–291. 23 Klus, H.; Scherer, G.; Müller, L. Influence of Additives on Cigarette Related Health Risks. Beiträge zur Tabakforschung International/Contributions to Tobacco Research, 2012, 25, 3, 411-493.

18

The chemistry reviewers conducted a search of peer-reviewed literature and identified eight additional studies reporting results of chemical analyses of “heat-not-burn” tobacco products. The individual studies are described in the chemistry PMTA review. Auer et al.27 compared the concentrations of eight volatile organic compounds (VOCs), 16 PAHs, three inorganic compounds, and nicotine in mainstream aerosol generated at 330 °C in the IQOS system with Heatsticks and in mainstream cigarette smoke at 684 °C. Although the results indicated significantly elevated levels of acenaphthene and formaldehyde in the IQOS product, the chemists concluded the data published by Auer et al. are not considered adequate for comparing the levels of HPHCs between the IQOS products and CC due to analytical issues – specifically lack of testing reference samples, low number of replicates, and a lack of sensitivity on some analytical methods. Other studies and conclusions include: • Farsalinos et al.28 compared nicotine levels among IQOS, e-cigarettes (EC), and commercially available cigarettes and conclude that the “HnB29 product delivers nicotine to the aerosol at levels higher than ECs but lower than a tobacco cigarette when tested using Health Canada Intense puffing regime.” • Savareear et al.30 reported on a list of 205 compounds identified in the aerosol of Heatsticks, including flavor and fragrance agents, humectants, natural substances, and a plasticizer. The article lists 82 compounds that were not previously reported in cigarette smoke, including 43 compounds previously reported in tobacco leaves. Savareear et al. conclude the chemical composition of the aerosol of Heatsticks is significantly less complex compared to the smoke of a combustible product, although the aerosol is not fully characterized. • Bekki et al.31 compared nicotine, tar, carbon monoxide (CO), and TSNA levels in mainstream smoke and tobacco filler between the IQOS products and the reference cigarettes 1R5F and 3R4F. CO was found to be 99% lower in the Heatstick aerosol compared to mainstream cigarette smoke; NNN was reduced by 90-94% and NNK by 87-95%. • Davis et al.32 evaluated the performance of the IQOS system using two different cleaning protocols. This study found evidence of release of formaldehyde cyanohydrin (glycolonitrile) when the cleaning protocol described by the applicant (clean after every 20 Heatsticks) is followed. In amendment PM0000466, the applicant submitted a chromatographic study of PLA in response to the release of formaldehyde cyanohydrin reported by Davis et al. The applicant stated that, based on chromatographic data and literature,33 the compound that Davis et al. identified as formaldehyde cyanohydrin is likely meso-lactide, a condensation product of lactic acid.

27 Auer, R., Concha-Lozano, N., Jacot-Sadowski, et al. Heat-not-burn tobacco cigarettes: smoke by any other name. (2017). JAMA Internal Medicine, 177, 1050-1052. 28 Farsalinos, K., Yannovits, N., Sarri, T., et al. (2017). Nicotine delivery to the aerosol of a heat-not-burn tobacco product: comparison with a tobacco cigarette and e-cigarettes. Nicotine Tob Res, 1-6. 29 HnB = Heat not burn 30 Savareear, B., Lizak, R., Brokl, M., et al. (2017). Headspace solid-phase microextraction coupled to comprehensive two- dimensional gas chromatography–time-of-flight mass spectrometry for the analysis of aerosol from tobacco heating product. J. Chromatogr, 1520, 135-142. 31Bekki, K., Inaba, Y., Uchiyama, S., & Kunugita, N. (2017). Comparison of chemicals in mainstream smoke in heat-not-burn tobacco and combustion cigarettes. J University of Occupational and Environmental Health (UOEH), Japan, 39(3), 201-207. 32 Davis, B.; Williams, M.; Talbot, P. (2018). IQOS: evidence of and release of from plastic. Tob Control, 1-8. 33 Arrieta, M.; Parres-Garcia, F.; Lopez-Martinez, J.; Navarro-Vidal, R.; Ferrandiz, S. (2012). Pyrolysis of biBplastics Waste: Obtained Products from Poly(lactic acid) (PLA). Dyna, 87 (4), 395-399.

21

• Stephens34 compared the quantities published in the literature for 13 HPHCs in mainstream cigarette smoke, in mainstream aerosol of e-cigarettes, and in a prototype of “heat-not-burn” device (THS 2.2).35 The quantities of the 13 HPHCs are 1-3 orders of magnitude lower in THS 2.2 compared to CC and the quantities of four HPHCs (acetaldehyde, formaldehyde, NNN, and NNK) are 1-2 orders of magnitude higher in THS 2.2 compared to e-cigarettes. • Mallock et al.36 compared the levels of nicotine, tar, TPM, water, four aldehydes (acetaldehyde, acrolein, formaldehyde, and crotonaldehyde) and five VOCs (1,3-butadiene, benzene, isoprene, styrene, and toluene) in the aerosol of the IQOS system with the three different Heatsticks with data from combustible cigarettes published in Counts et al.37 The level of nicotine was comparable to combustible cigarettes, and lower for aldehydes (80-96%) and VOCs (97-99.8%) in the aerosol of the IQOS system with Heatsticks compared to mainstream cigarette smoke. • Li et al.38 compared the levels of TPM, water, tar, nicotine, propylene glycol, glycerin, carbon monoxide, and 25 HPHCs in the IQOS system with Heatsticks and the Kentucky reference cigarette 3R4F, under International Organization for Standardization (ISO) and Canadian Intense (CI) smoking regimens. The level of tar in the IQOS system with Heatsticks was comparable to the level found in the mainstream smoke of Kentucky reference cigarette 3R4F; however, nicotine was 29% lower. Other measures were similar to those reported by the applicant. Li et al. also compared the chemicals obtained in the IQOS system with Heatsticks, two commercial CC, and reconstituted tobacco blend during simulated pyrolysis at 350°C. When the tobacco from the four products (Heatsticks, commercial CC, and reconstituted tobacco blend) were heated to the same temperature (350°C), all emitted comparable levels of chemicals., This suggests that the temperature of the IQOS system, rather than the tobacco filler ingredients, has a major impact on the levels of harmful constituents.

In summary, the level of nicotine, tar, glycerol, HPHCs, and other components in the aerosol reported in six of the eight peer-reviewed articles is similar to data reported by the applicant. Auer et al. reported higher levels of some compounds compared to the applicant but there may be methodological issues with this study. For the reasons set forth above, the chemistry reviewers do not believe these differences raise any concerns.

e. Product Stability To maintain tobacco blend characteristics over time, (b) (4) (b) (4) , analysis of (b) (4) . The applicant has established tolerance limits for (b) (4) (b) (4)

34 Stephens, W. (2017). Comparing the cancer potencies of emissions from vapourised nicotine products including e-cigarettes with those of tobacco smoke. Tob Control 2017; 0:1–8. 35 Schaller, J.P.; Pijnenburg, J.; Ajithkumar, A.; Tricker, A. (2016). Evaluation of the tobacco heating system 2.2. Part 3: Influence of the tobacco blend on the formation of harmful and potentially harmful constituents of the tobacco heating system 2.2 aerosol. Regul Toxicol Pharmacol 81(Suppl 2): S48–58. 36 Mallock, N.; Böss, L.; Burk, R.; Danziger, M. et al. (2018). Levels of selected analytes in the emissions of “heat not burn” tobacco products that are relevant to assess human health risks. Arch Toxicol, in press. 37 Counts, M.; Morton, M.; Laffoon, S.; Cox, R. et al (2005). Smoke composition and predicting relationships for international commercial cigarettes smoked with three machine-smoking conditions. Regul Toxicol Pharmacol 41(3), 185–227. 38 Li, X.; Luo, Y.; Jiang, X.; Zhang, H. et al (2018) Chemical analysis and simulated pyrolysis of tobacco heating system 2.2 compared to conventional cigarettes. Nicotine Tob Res., 1-8.

22

product.41,42,43,44,45,46 Additionally, factors such as pH, moisture content, nitrate/nitrite concentrations, microbial content and product storage temperature are reported to influence microbial stability and TSNA formation during tobacco product storage.47,48,49 The applicant states that the product specifications for shelf life are a (b) (b) (4) (4)

Amendments MR0000085 and MR0000096 provided complete stability testing data for all the Heatsticks measured over a period of (b) (4) of product storage. Samples were collected at (b) (4) and evaluated for (b) (4) . (b) (4)

The applicant concluded that a shelf life of (b) (4) is acceptable at(b) (4) and a shelf life of(b) (4) is acceptable at(b) (4) for all Heatstick varieties. (b) (4) .

Changes in the(b) (4) may be of concern if they indicate microbial changes. A key factor in determining potential to support microbial growth is the amount of water that is available, which is described in terms of aw. Aw limit varies with various solutes (water) and humectants. The applicant provided no explanation for why (b) (4) (b) (4) Additionally, the applicant did not provide any aw data for the Heatsticks to show that the (b) (4) is not a microbiological concern. However, the applicant states (in Amendment MR0000085) that the tobacco portion of the finished product has an approximate moisture content of (b) (4) and the (b) (4) (b) (4) concentration exceeds , under which the aw is not expected to exceed . It is generally recognized that 50 no microbial proliferation occurs with aw <0.60. Additionally, the applicant submitted a graphical representation of the moisture content data of the tobacco plug recorded as part of a 15-month product monitoring study of Heatsticks that were shipped by air from the manufacturing center in Bologna, Italy to several warehouses in Japan and stored under standard warehouse conditions. Based on this data, the tested moisture content of the Heatsticks was approximately (b) (4) - below the 20% level considered necessary for microbial growth.

During inspection of the PMP S.A. site in Switzerland, FDA obtained pictures taken as part of the visual inspection of the Heatsticks. The applicant states that (b) (4) , observations on

41 Andersen, R.A., Fleming, P.D., Hamilton-Kemp, T.R., and Hildebrand, D.F. 1993. pH changes in smokeless undergoing nitrosation during prolonged storage: Effects of moisture, temperature, and duration. J. Agric. Food Chem. 41:968-972. 42 Chopyk, J., Chattopadhyay, S., Kulkarni, P., Claye, E., et al., 2017a. Mentholation affects the cigarette microbiota by selecting for bacteria resistant to harsh environmental conditions and selecting against potential bacterial pathogens. Microbiome. 5:22. 43 Ibid, Chopyk, J., Chattopadhyay, S., Kulkarni, P., Smyth, E.M., et al., 2017b 44 Mutasa, E.S., Seal, K.J., and Magan, N. 1990. The water content/water activity relationship of cured tobacco and water relations of associated spoilage fungi. Int Biodeterior. 26: 381-396. 45 Smith, T.E. 1964. A literature review of aging and fermentation of tobacco. http://legacy.library.ucsf.edu/tid/ubb93f00/pdf. 46 St. Charles, F.K. 1989. Reduction of the water activity of wet snuff/259. Brown & Williamson Tobacco Corporation Research and Development. http://legacy.library.ucsf.edu/tid/aaj41f00/pdf. 47 Brunnemann, K.D., Prokopczyk, B., Djordjevic, M.V., and Hoffmann, D. 1996. Formation and analysis of tobacco-specific N- nitrosamines. Crit. Rev. Toxicol. 26:121-137. 48 Fisher, M.T., Bennett, C.B., Hayes, A., Kargalioglu, Y., et al., 2012. Sources of and technical approaches for the abatement of tobacco specific formation in moist smokeless tobacco products. Food Chem Toxicol. 50:942–948. 49 Rutqvist, L.E., Curvall, M., Hassler, T., Ringberger, T., and Wahlberg, I. 2011. Swedish sSnus and the GothiaTek Standard. Harm Reduct J. 8:11. 50 Beuchat, L.R. 1983. Influence of water activity on growth, metabolic activities, and survival of yeast and molds. J. Food Protect. 46:135-141.

24

visual quality of the product ((b) (4) ) were not related to safety or product performance. However, the applicant states that it will add aw testing to future high humidity stability studies, and, if high aw is detected, microbiological testing will be performed.

3. Heating System a. Holder The Holder heats the tobacco using a g(b) (4) ceramic blade, which is pushed into the tobacco plug by the act of inserting the HeatStick into the Holder. The Holder has a small battery, which stores enough energy for a single use (i.e. complete use of one HeatStick). (b) (4)

The Holder needs to be recharged prior to each HeatStick use. The Holder is activated by a button and its status is indicated through an interface that includes a colored light-emitting diode (LED).

In Amendment MR0000085, the applicant describes changes to(b) (4) (b) (4) . These changes were made to (b) (4) .

b. Charger The Charger is a pocket-sized recharging case for the Holder. It contains a larger battery and charging electronics, which recharges the Holder battery when the Holder is placed inside. The Charger battery holds sufficient charge to recharge the Holder 20 times. The Charger is recharged using an AC adaptor.

When the Holder is inserted in the Charger, it is possible to initiate the blade cleaning process. This procedure heats the blade to a higher temperature than during Holder use to facilitate the removal of deposits left by multiple inhalation experiences, and thus ensures consistent heating performance in normal use. The Charger status is displayed and controlled through an interface that includes colored LEDs and two buttons.

Amendment MR0000085 describes changes in (b) (4) ; this change was made to (b) (4) . The amendment also (b) (4) (b) (4) .

c. Manufacturing, Process, and Controls The IQOS THS is designed and manufactured in accordance with published external standards when available and applicable for the product category and all systems/sub-systems. These standards have been third-party tested per the regulatory compliance standards. The engineers requested the IQOS Charger battery be tested under IEC62133:2012 and all test units passed. In addition, the IQOS THS follows applicable European Directive which is intended to improve environmental policies associated with batteries sold in the European Union.

The manufacturing and assembly processes for the product components are described in the applications and reviewed in detail in the Engineering review. The applicant submitted part-by-part and sub-assembly details for the following components: • (b) (4)

25

(b) (4)

The Holder and Charger contain microcontrollers and firmware (b) (4) (b) (4) Details are in the Engineering review. The Engineering review notes that the Holder and Charger firmware architecture is based on (b) (4)

The heating blade cleaning function (b) (4)

(b) (4)

The product is designed to use interchangeable batteries. The applicant provided the supplier manufacturing specifications, which are aligned with the product battery specifications for the Holder and the Charger. To ensure a “full experience” with each use, the applicant has established minimum battery standards.

The applicant submitted product battery samples of the IQOS Holder and Charger batteries (50 samples each) to Winchester Engineering and Analytical Center (WEAC) in September 2017. Engineering requested evaluation of conformance to certain requirements of IEC62133:2012 (second edition) to be measured for the new products. Testing was performed on all the products with 5-10 replicates, depending on the parameter. No individual data points were out of specification when compared to the applicant’s range limits.

4. Inspections of Manufacturing Facilities FDA inspections were performed of the applicant’s research and manufacturing sites in Lausanne, Switzerland (product testing), Neuchatel, Switzerland (product design and research), Bologna, Italy (HeatStick production), (b) (4) ).

(b) (4)

27

(b) (4) There were no discussion items or deficiencies identified.

PMP S.A., located in Neuchatel, Switzerland, is a subsidiary of PMI51 with headquarters in Lausanne, Switzerland. Activities conducted at this site include (b) (4)

. (b) (4)

(b) (4)

PM Manufacturing and Technology Bologna is in Bologna, Italy. (b) (4)

There were no discussion items or deficiencies identified.

During the inspection of the Bologna (Italy) facility, CoAs for batch releases collected included the specifications and the results obtained for one batch of each of the three Heatsticks, PM0000424-426. All the results obtained were within specifications. The inspection report (FEI: 3011169041) for the facility in Italy notes discrepancies in weight specifications between the application and the batch records collected. The batches collected were for the Japanese market. In the batch records the specifications for weight of the crimpled PLA filter are (b) (4) for PM0000425 and (b) (4) for PM0000426. In the

51 The cover letter of the PMTAs notes: The application refers to Philip Morris International (PMI) which includes: Philip Morris International Inc., Philip Morris Products S.A. (the PMT and MRTP applicant), Philip Morris International Management S.A., Philip Morris International Research Laboratories Pte. Ltd., and Philip Morris Manufacturing & Technology Bologna S.p.A.

28

applications, the weight specifications are (b) (4) for PM0000425 and (b) (4) for PM0000426. In addition, in the (b) (4) process, the specification for weight is (b) (4) for the three Heatsticks. In the batch record for Marlboro Smooth Menthol Heatsticks and Marlboro Fresh Menthol Heatsticks, the specification for weight is (b) (4) . The applicant indicated that the differences are mainly due to different cigarette paper used in the Japanese and the U.S. markets.

5. Summary of Engineering, Chemistry, Product Stability, and Manufacturing Findings The engineering review concludes that the PMTAs contain adequate information with respect to the following: • A complete characterization of the design parameters • An adequate description of manufacturing steps and quality control measures • Adequate process controls and quality assurance procedures to help ensure that the products meet manufacturing specifications for the IQOS Holder, Charger, and Marlboro, Smooth Menthol, and Fresh Menthol Heatsticks and that the products are manufacture in a consistent manner that minimizes the variability in product quality • Performance testing to verify the product design The engineering review concludes that these PMTAs contain sufficient information to characterize the product design and adequate processes and controls to help ensure that the products meet the manufacturer’s specifications.

As TPL, I agree with the engineering conclusions. In addition to the above information, the applicant has made changes to the Holder and Charger which are likely to lead to a more consistent manufacturing process and improve product reliability. The applicant has provided a description of the function and design of the Heatstick filters. The applicant has no efficiency requirements for Heatstick filters as the filters do not control nicotine delivery. The applicant has provided an adequate description of the firmware functionality for control of the heating blade temperature and cleaning function, as well as for function and battery management of the Holder and Charger. Additionally, the applicant has described battery specifications for the vendors that will help to ensure product consistency and reduce concerns of malfunction. The battery testing performed at the Winchester Engineering and Analytical Center demonstrated consistent battery performance, which reduces concerns of malfunction.

The chemistry review concludes these PMTAs contain adequate information as follows: • A complete list of uniquely identified components, ingredients, and additives by quantity in each new tobacco product as well as the applicable specifications and a description of the intended function for each • An adequate description of manufacturing steps and quality control measures in place • Sufficient information to assure FDA that the products meet manufacturing specifications for (b) (4) , nicotine, (b) (4) , phenol, carbon monoxide, acrylamide, and menthol and that the products are manufactured in a consistent manner that minimizes the variability in product quality • Data on chemical endpoints establishing the stability of the product through the stated shelf life • Product analyses for verifying the product formulations • Testing data to demonstrate that the new products contain significantly lower levels of certain HPHCs including formaldehyde, acrolein, carbon monoxide, NNN, NNK, compared to major types of combusted cigarettes on the U.S. market • There were small weight differences between the application and the batch noted during the inspection of the Bologna Italy facility. The difference was due to different cigarette paper used in Japan vs. the U.S. and do not raise any concerns regarding product quality.

29

The chemistry review concludes that these PMTAs contain sufficient information to characterize the product composition in terms of ingredients and additives and describe the manufacturing processes and controls that can affect the product composition, chemical stability, and HPHC levels to help ensure that the products meet the manufacturer’s specifications.

As TPL, I agree with the chemistry conclusions.

The microbiology review concludes the applicant provided adequate microbiology-related information to demonstrate full product characterization and stability over product shelf-life and to address factors that can potentially affect the microbial stability of the product as well as adequate quality control information. Specifically: • The applicant has provided information to support (b) (4) shelf life for all three Heatsticks. • Based on the information provided, adequate measures are being taken to address the quality (b) (4) and stability of the Heatsticks exposed to high heat and humidity conditions.

As TPL, I agree with the microbiology conclusions. Although there was (b) (4) (b) (4) , it is unlikely this is related to product safety or performance as the moisture content of the Heatsticks is well below the level necessary for microbial growth.

The OCE manufacturing review identified no significant compliance issues during the five manufacturing inspections conducted.

C. Toxicological Risk Assessment 1. Harmful and Potentially Harmful Constituents (HPHCs) a. General Overview Tobacco fermentation is a microbial-mediated reduction of nitrate reacting with alkaloids present in tobacco to produce tobacco-specific nitrosamines (TSNAs).52,53 TSNAs are primarily formed during tobacco curing and fermentation of processed tobacco, as well as during aging/storage of the processed and packaged tobacco product. Factors such as nitrate and nitrite concentrations, moisture content, microbial content, pH, and storage temperature are reported to influence microbial stability and TSNA formation during tobacco product storage. Although the tobacco component of the Heatsticks does not include any fermented tobacco, (b) (4) tobacco is included.

HPHCs are formed by the incomplete combustion and thermal degradation of the tobacco, additives, and paper as the cigarette burns. The temperature at the center of a burning cigarette is 600–800⁰ C but can reach temperatures as high as 900⁰ C.54 The IQOS system is designed to heat tobacco to approximately 300⁰ C and employs a thermal monitoring system that prohibits temperatures from exceeding 350⁰ C. The HPHC analysis submitted by the applicant demonstrates that some thermal degradation products that are generated as tobacco burns are also found in Heatstick aerosols, albeit at lower levels. It is possible that

52 Di Giacomo, M., Paolino, M., Silvestro, D., Vigliotta, G., et al.,2007. Microbial community structure and dynamics of dark fire-cured tobacco fermentation. Appl Environ Microbiol. 73:825–837. 53 Ibid, Fisher, M.T., Bennett, C.B., Hayes, A., Kargalioglu, Y., et al., 2012. 54 Baker RR. Temperature distribution inside a burning cigarette. Nature. 1974;247(5440):405-406.

30

other unmeasured constituents may be formed at temperatures below the combustion threshold for tobacco.55 The applicant conducted a series of in vitro and in vivo nonclinical studies assessing certain toxicities of Heatstick aerosol compared to 3R4F, which should detect adverse effects caused by aerosol constituents not identified by physical characterization of the aerosol.

The applicant submitted the following testing data obtained for PM0000424 – PM0000426 manufactured under commercial manufacturing conditions: • TNCO levels using the ISO regimen • FDA 18+6: This study measured yields of 18 chemicals on the current FDA HPHC list in the Heatstick aerosols (measured under ISO and modified CI regimens) and six constituents found in Heatstick filler. • PMI-58 study: This study compared yields of 55 chemicals (measured under the modified CI regimen) on the current FDA HPHC list that are found in Heatstick aerosols and 3R4F smoke. The applicant also reported yields of (b) (4) , nicotine, tar, (b) (4) . For 18 of the aerosol compounds, the applicant compared the levels found in Heatstick aerosols to mean levels in the smoke of 30 commercially available cigarettes. A comparison was also done for six Heatstick filler constituents. • Amendment MR0000114 (study 93-FDA-HPHCs) included additional information on yields of all 93 chemicals on the current FDA HPHC list for both Heatstick aerosols and 3R4F smoke. In Heatstick aerosols, levels of 39-40 of the chemicals were too low too to be quantified; for the other 53-54 chemicals that could be quantified, the previously reported levels in the PMI-58 study were verified. • Non-Targeted Differential Screening: This study, submitted in Amendment MR0000097, provides the levels of 80 individual constituents present in the aerosol of one or more of the Heatsticks at higher concentrations than in the mainstream smoke of 3R4F. • P1 characterization: This study includes chemical constituents present at concentrations higher than 100 ng/Heatstick in the aerosol of MR0000059 under a modified CI smoking regimen.

In the FDA 18+6 study and the Non-Targeted Differential Screening, the applicant compared the quantity of each constituent to data obtained from the Kentucky reference cigarette 3R4F. The comparison was performed both per unit (quantity in Heatstick aerosol compared to quantity in cigarette smoke) and per amount of nicotine.

The FDA 18+6, PMI-58, and 93-FDA-HPHCs studies were performed by Labstat International ULC. Labstat submitted method details in TPMF (b) (4) in December 2017. TNCO and Non-Targeted Differential Screening studies were performed by PMP S.A. The analytical method used to determine water and tar content by PMP S.A. was modified to account for a larger amount of water in the aerosol (80%) of the IQOS compared to the amount of water in the mainstream smoke of CC (20%). Ghosh et al. demonstrated that the difference in tar mean values between the in-situ methodology and the standard ISO 4387 methodology was -50% for the and -4% for CC by ISO smoking regimen. In the heated tobacco product, tar is 9.39 mg/Heatstick using the standard extraction and 4.71 mg/Heatstick using the in-situ extraction.56 This approach is appropriate for analysis of this product.

55 Auer R, Concha-Lozano N, Jacot-Sadowski I, Cornuz J, Berthet A. Heat-not-burn tobacco cigarettes: Smoke by any other name. JAMA Internal Medicine. 2017. 56 Ghosh D., Jeannet C. An Improved Filter Pad Extraction Methodology to Obtain More Accurate Water and “Tar” Values. In Situ Cambridge Filter Pad Extraction Methodology. Beiträge zur Tabakforschung International/Contributions to Tobacco Research. 2014, Volume 26 (2), 38-49.

31

b. Constituents Unique to IQOS The non-targeted differential screening of Heatstick aerosols and 3R4F cigarette smoke found 80 chemicals that were either present in higher concentration in Heatstick aerosols than 3R4F smoke or not found in 3R4F smoke: 4 are possibly carcinogenic, 30 are identified by the applicant as Generally Recognized as Safe (GRAS), and 46 additional ingredients (mostly flavoring ingredients).

The applicant indicates the four possible (glycidol, 3-chloro-1,2-propanediol [3-MCPD], 2- furanmethanol, and furfural) do not pose a toxicological concern because the levels are below recognized dietary or occupational exposure limits. The applicant provided the following toxicological assessments: • Comparison against occupational exposure limits (OELs) • Use of OSHA’s Permissible Exposure Limit (PEL) as a standard for some exposures • Compared the exposure from IQOS aerosol for the four chemicals to maximum dietary intake

The assessment of these carcinogens is not considered adequate. Comparison of estimated exposures from use of tobacco products to OELs is not appropriate for a risk assessment of chemicals found in tobacco product smoke and aerosols. OELs are not health values and are not intended for use to evaluate potential health hazards from inhaled tobacco products. OSHA PELs are intended for a specific scenario in the workplace including exposure during an 8-hour work shift within a 40-hour work week. PELs are also intended to be used together with proper engineering controls (e.g., monitoring the work environment, application of feasible technological controls) and good work practices (e.g., wearing respirators) to minimize hazardous substance generation and exposure. Extrapolation of risk from dietary exposure to determine risk from inhalation is inappropriate, as the most sensitive effects and target organs drastically differ depending on whether a is ingested or inhaled. Extrapolation from dietary limits for inhalation exposure ignores differences in toxicokinetics or distinct effects at the portal of entry. The explanation provided by the applicant does not support a conclusion that these pose no risk to IQOS users; however, the levels of exposure to these possible carcinogens appear low and when considered with other data does not preclude a conclusion the products are appropriate for protection of public health.

Initially the applicant did not provide any analysis of the GRAS compounds. In response to a request for additional information, the applicant provided predictive toxicology modeling and available toxicological data for 30 chemicals present in higher levels in Heatstick aerosol compared to 3R4F smoke. Four of the 30 chemicals have known respiratory effects (irritation, sensitization, respiratory depression) and one has potential to influence nicotine metabolism. For other chemicals, toxicological data via the inhalation route is not available and their individual contributions in inhalation toxicology are unknown. Genotoxicity and carcinogenicity information for many of these chemicals is not available. The applicant analyzed all 30 chemicals with the OECD quantitative structure-activity relationship (QSAR). Eleven chemicals were identified with genotoxic potential. Based on the available toxicological data and predictive toxicology modeling analysis submitted by the applicant, 20 of the 30 chemicals exhibit concerns for potential health effects. Many of the chemicals do not have sufficient inhalation toxicity or genotoxicity/carcinogenicity data to inform the toxicological evaluation of heated tobacco products. The data provided by the applicant is not sufficient to support their conclusion that these compounds pose no risk to IQOS users; however, although there is potential for genotoxicity with some of these compounds, the exposure levels appear low and the available data does not preclude a conclusion the products are appropriate for protection of public health.

The applicant analyzed the remaining 46 chemicals (primarily flavor ingredients) with the OECD QSAR Toolbox to detect structural alerts for DNA binding or carcinogenicity. Of these 46 chemicals, 8 were identified as potentially genotoxic and/or carcinogenic. Along with the 11 noted above, the applicant

32

indicates that 19 of the 80 chemicals that were either unique to Heatstick aerosols or found at higher concentrations than in 3R4F produced a structural alert for genotoxicity and 20 more GRAS compounds have potential health effects.

c. Comparison to Cigarette Smoke PMI-58 study The PMI-58 study included measurement of 55 HPHCs in the mainstream aerosol generated from PM0000424, PM0000425, and PM0000426 and in smoke generated from the 3R4F reference cigarette using a CI smoking regimen. On a per stick basis, measured HPHC levels (except nicotine) were reduced in Heatstick aerosols by ~54-99.9% compared to 3R4F. Nicotine levels in Heatstick aerosols were reduced by ~26%-39% compared to 3R4F.

While the Heatstick aerosols generated by smoking machines contained less nicotine than smoke from 3R4F research cigarettes on a per stick basis, the clinical data indicate that humans can absorb nicotine from Heatstick aerosols at levels comparable to their current cigarette brands. Consequently, HPHC yields normalized to nicotine yield are more likely to reflect actual human exposure levels than HPHC yields expressed on a per stick basis. When normalized to nicotine yield, the yields of HPHCs that were measured were reduced by 24.8%-99.8% compared to the smoke from the CC evaluated by the applicant.

The applicant measured 18 HPHCs plus tar and water in smoke generated from 31 different Philip Morris USA brand CC marketed in the U.S. with a rotary smoking machine using the CI smoking regimen. Results are shown in Table 2 below. Except for nicotine, HPHC yields in Heatstick aerosols on a per stick basis are reduced by 40.0%-99.9% when compared to the smoke from the CC; nicotine levels are 36-42% lower. When normalized to nicotine yield, HPHC yields are reduced by 38.2%-99.8% when compared to the smoke from the CC evaluated by the applicant. The applicant quantified the levels of 6 HPHCs (nicotine, ammonia, cadmium, arsenic, NNN, and NNK) in the tobacco filler.

Data were not provided comparing these levels to 3R4F; however, the applicant indicates there is (b) (4) mg of nicotine in the tobacco filler of an unused Heatstick, but only 1.19-1.29 mg of nicotine is volatilized into the aerosol. The applicant did not provide the nicotine levels for used Heatsticks.

The applicant also included measures of tar, water and total particulate matter (TPM). Although TPM is 20- 32% higher in the aerosol of the Heatsticks than in CC, the composition is different. The TPM produced by the IQOS system contains 76% (b) and 10% (b) (4) while the TPM produced by the reference cigarette 3R4F contains 32% water and 5%(4) glycerol.57 In the three products, the level of tar is 20-36% lower in the aerosol compared to the reference cigarette 3R4F.

Formaldehyde and acrolein are produced by glycerol and propylene glycol.58 Despite the higher level of glycerol and propylene glycol in the Heatsticks than in cigarettes, the levels of acrolein and formaldehyde in the aerosol of the Heatsticks are lower than in cigarette smoke. Acrolein is 89-95% lower and formaldehyde is 66-91% lower in the aerosol of the Heatsticks compared to cigarette smoke.

57 Schaller, J.P., Keller, D., Poget, L., et al. (2016). Evaluation of the Tobacco Heating System 2.2. Part 2: Chemical composition, genotoxicity, cytotoxicity, and physical properties of the aerosol. Regulatory Toxicology and Pharmacology, 81, S27-S47. 58 Sleiman, M., Logue, J., Montesinos, V., et al. (2016). Emissions from electronic cigarettes: key parameters affecting the release of harmful chemicals. Environ. Sci. Technol., 50, 9644-9651.

33

NNN and NNK are formed by nitrosation of alkaloids present in the tobacco plant during tobacco processing, curing, and storage. NNN and NNK are 92-98% lower in the aerosol of the Heatsticks than in cigarette smoke. CDC studies show that the mainstream smoke of burley and reconstituted tobaccos contain much higher TSNA levels than bright and oriental tobacco.59 While NNN and NNK levels may be lower in the aerosol of the IQOS system due to the lower heating temperature for the tobacco, the main reduction is likely caused by selecting tobacco blends with a lower propensity for TSNA formation and by limiting the use of nitrogen fertilizer.60

Nitrogen oxides are 97-99% lower in the aerosol of the Heatsticks than in mainstream cigarette smoke. Reconstituted tobacco can produce high levels of CO and nitrogen oxides during combustion.18 The lower levels are likely related to lack of combustion.

Table 2: Comparison of HPHC Yields from Heatstick Aerosols and Combusted Cigarettes

P1 Characterization study In the P1 characterization study, the applicant reported 498 compounds present in the aerosol of PM0000424 at concentrations of 100 ng/Heatstick or higher. All compounds were also identified in the mainstream smoke of the Kentucky reference cigarette 3R4F. (Chemicals that are unique to Heatstick aerosols are present in concentrations < 100ng/Heatstick.) All the compounds reported in the Non-Targeted Differential Screening study with concentrations higher or equal to 100 ng/Heatsticks were reported in the P1 characterization study, except for eight compounds that did not overlap between the studies (lanost-8- en-3-ol, 24-methylene; 12,14-labdadiene-7,8-diol, (8a,12E); isolinderanolide; ethyl 2,4-dioxohexanoate;

59 Ding, Y.S., Zhang, L., Jain, R.B., et al. (2008). Levels of tobacco-specific nitrosamines and polycyclic aromatic hydrocarbons in mainstream smoke from different tobacco varieties. Cancer Epidemiology, Biomarkers & Prevention, 17 (12), 3366-3371. 60 Centers for Disease Control and Prevention (CDC). (2010). Chemistry and Toxicology of Cigarette Smoke and Biomarkers of Exposure and Harm in How Tobacco Smoke Causes Disease: The Biology and Behavioral Basis for Smoking-Attributable Disease: A Report of the Surgeon General. Atlanta, GA: CDC.

34

benzoic acid, 2,5-dihydroxy-methyl; ergosterol; isoquinoline 3-methyl; and pyridoxin). For these eight compounds, there were limits of detection, changes in the evaluation of the fragmentation pattern, and some of the compounds in the Non-Targeted Differential Screening study were identified as adducts of the compounds reported in the P1 characterization study.

d. Environmental Exposure from Heatstick Aerosol Cigarette combustion generates environmental tobacco smoke (ETS), which consists of both side-stream smoke emitted from the cigarette and smoke exhaled by a smoker. The applicant states that the Heatstick does not produce side-stream smoke and submitted three peer-reviewed manuscripts as well as slides from a meeting presentation in support of this assertion.61,62,63,64 Overall, the studies indicate that heated tobacco products, including Heatsticks, emit detectable levels of some HPHCs, but those levels are much lower than emissions from CC.

2. In Vitro Studies A common generation and collection method was used for the Heatstick aerosols and 3R4F smoke for the in vitro studies (Ames test, mouse lymphoma assay, neutral red uptake assay) submitted by the applicant. Some of the methodology raises questions regarding interpretation of the data; however, there are no validated regimens for generating Heatstick aerosols. The applicant does not provide any rationale or justification for the differences in TPM or gas vapor phase (GVP) collection, and it is unclear what effect the collection methods may have on results of the studies; however, similar results for nicotine and acrolein levels measured by the two different methods indicate that the applicant’s methodology is acceptable.

a. Neutral Red Uptake (NRU) Assays Neutral Red Assay (NRU) can determine cytotoxicity. Studies RLS-ZRH-2015-249 and RLS-ZRH-2015-250 measured nicotine concentrations in the TPM from Regular and Fresh Menthol Heatstick aerosols as well as 3R4F smoke. Similarly, the applicant measured acrolein concentrations in the GVP from Regular and Fresh Menthol Heatstick aerosols as well as 3R4F. Cytotoxicity in the NRU results is expressed as the reciprocal of the effective concentration that reduces the number of viable cells by 50% (1/EC50). The 1/EC50 values were calculated for both TPM and GVP fractions and expressed on both a per item and per nicotine basis. On a per stick basis, 1/EC50 values were 94%-95% lower for Regular Heatstick aerosols compared to 3R4F. Similarly, 1/EC50 values were 95% lower for Fresh Menthol Heatstick aerosols when compared to 3R4F. When normalized to nicotine yield, the NRU 1/E50 values were 91-92% lower for TPM and GVP from Heatsticks than for 3R4F RCS, indicating a reduced cytotoxic potential.

Clinical evidence provided by the applicant indicates that Heatstick users are frequently exposed to nicotine levels that are comparable to cigarette smokers. As such, assay results that are normalized to nicotine content are most appropriate for comparing cytotoxicity of Heatstick aerosols to 3R4F smoke.

61 Frost-Pineda K, Zedler BK, Liang Q, Roethig HJ. Environmental tobacco smoke (ETS) evaluation of a third-generation electrically heated cigarette smoking system (EHCSS). Regulatory Toxicology and Pharmacology. 2008;52(2):118-121. 62 Tricker AR, Schorp MK, Urban H-J, et al. Comparison of Environmental Tobacco Smoke (ETS) Concentrations Generated by an Electrically Heated Cigarette Smoking System and a Conventional Cigarette. Inhalation Toxicology. 2009;21(1):62-77. 63 O’Connell Peter Wilkinson G, Burseg K, J Stotesbury S, D Pritchard J. Heated Tobacco Products Create Side-Stream Emissions: Implications for Regulation. Journal of Environmental Analytical Chemistry. 2015;02. 64 Goujon-Ginglinger C. MS. Indoor Air Quality Assessment of the Tobacco Heating System THS 2.2, Electronic Cigarettes and Cigarettes using a Dedicated Exposure Room. Paper presented at: Atmos'Fair 2016; Lyon, France.

35

b. Bacterial Reverse Mutation (Ames) test The Ames test (studies RLS-ZRH-2015-253 and RLS-ZRH-2015-254) detects chemicals that induce mutations in bacteria that restore the functional capability to synthesize an essential amino acid (e.g., histidine). Bacteria that undergo these changes are called revertants – the more revertants, the more mutagenic the substance. In these studies, the applicant exposed five Salmonella typhimurium strains to varying concentrations of TPM from Regular and Fresh Menthol Heatstick aerosols, as well as 3R4F reference smoke, for 48-72 hours. The study reports did not contain information from an Ames test with GVP from Heatstick aerosols or 3R4F. The HPHC information submitted by the applicant indicates that Heatstick aerosols contain that are typically found in GVP (e.g., formaldehyde, propylene oxide). As such, an Ames test with GVP from Heatstick aerosols and 3R4F smoke would provide additional information about the mutagenic potential of the products.

The TPM fraction from Regular and Menthol Heatsticks did not produce a positive mutagenic response at any dose used in the Ames tests submitted by the applicant. In contrast, TPM from the 3R4F dose- dependently increased revertants in three bacterial strains, but only with metabolic activation. The positive controls used by the applicant produced a several-fold increase (typically by 300%-500%) in revertants when compared to untreated or vehicle-treated cultures.

c. Mouse Lymphoma Assay (MLA) The MLA is a qualitative test that can determine clastogenicity and mutagenicity in a mammalian cell line by measuring the resistance to a lethal pyrimidine analogue (i.e., triflurothymidine [TFT]). The frequency with which these mutations occur (i.e., mutant frequency [MF]) is commonly expressed as the number of mutants per million (106) viable cells. The applicant submitted study reports (studies RLS-ZRH-2015-251 and RLS-ZRH-2015-252) on MLAs conducted with aerosols from Regular and Fresh Menthol Heatsticks, as well as 3R4F. The applicant reported that relative total growth (RTG) was measured for the cytotoxicity assessments. The MFs were derived from the plating efficiencies of cells grown in TFT selective and non- selective media.

Both TPM and GVP from Regular and Fresh Menthol Heatstick aerosols produced cytotoxicity in the MLA, with the highest concentrations reducing RTG to less than 20%. The concentrations of TPM and GVP from 3R4F and Heatstick aerosols that were used in the MLA produced similar maximum levels of cytotoxicity (15- 20% RTG), but TPM and GVP from 3R4F produced these effects at much lower concentrations, indicating greater cytotoxic potency. For example, the concentration of TPM from 3R4F that reduced growth by 50% was about 13 times less than the concentration of TPM from Regular Heatsticks required to produce the same effect. The difference in cytotoxicity from GVP was even more pronounced: GVP from 3R4F produced 50% RTG at concentrations 29 times lower than GVP from Regular Heatsticks.

The study reports by the applicant indicate that TPM and GVP from Regular and Fresh Menthol Heatstick aerosols, as well as 3R4F, are cytotoxic; however, 3R4F produces cytotoxicity at much lower concentrations than Heatstick aerosols. Similarly, TPM and GVP from Regular and Fresh Menthol Heatstick aerosols, as well as 3R4F, are mutagenic. The lowest observed genotoxic effect levels (LOGELs) produced by 3R4F TPM were 15-30 times lower than the IQOS TPM. The LOGELs of 3R4F GVP were 8-24 times lower than the IQOS GVP. The applicant indicates this difference in LOGEL is an index of mutagenic potency. However, guidance from major public health resources (e.g., OECD, ICH, Health Canada, EPA) does not support this method of relative comparisons of mutagenic/genotoxic potency between tobacco products (or other chemicals).

36

d. Summary of In Vitro Studies Limitations of these assays affect the conclusions that can be drawn from test results. For example, while the Ames assay can robustly detect DNA damage from mutagens that directly interact with DNA, the bacterial strains used in these assays do not possess the complex DNA repair mechanisms of mammalian cells. The NRU test detects cytotoxicity in a mammalian cell line. When normalized to nicotine yield, TPM and GVP from Regular and Menthol Heatsticks were approximately 90% less cytotoxic than TPM and GVP from 3R4F reference cigarette smoke. The MLA detects mutagenicity in a mammalian cell line. Evidence from the MLA also indicates that GVP and TPM from 3R4F smoke produces cytotoxicity at a lower concentration than TPM and GVP from Heatstick aerosols. Since different mutagenicity assays detect different types of genetic damage; it is not expected that a chemical will generate uniformly positive or negative results in the various assays.

Overall, the evidence submitted indicates that although both Heatstick aerosols and 3R4F smoke produce cytotoxic changes in vitro, 3R4F produces cytotoxicity at much lower concentrations than Heatstick aerosols. Similarly, both Heatstick aerosols and 3R4F are mutagenic, though 3R4F appears to produce genotoxic effects at a much lower level than Heatstick aerosols. As noted above, the level of substance required to produce these effects may not be an accurate indicator of mutagenic potency. Consequently, it is difficult to determine from these in vitro evaluations whether long-term use of Heatsticks will have the same carcinogenic potential as CC smoke.

3. In Vivo Studies a. 90-day Nose-only Inhalation Studies The applicant submitted study reports from two separate 90-day nose-only inhalation studies with adult male and female Sprague-Dawley rats with a 42-day post-exposure recovery period. The first 90-day inhalation study (Study #15006) determined toxicity produced by repeated exposure to either aerosols from Regular Heatsticks, 3R4F smoke, or filtered air (sham control). In the second study (Study #15025), rats were exposed to aerosols from either Fresh Menthol Heatsticks, 3R4F, smoke from one of two mentholated versions of the 3R4F, or filtered air (sham control).

The applicant reported urinary levels of BOE to the harmful and potentially harmful constituents NNK (total NNAL), acrolein (HPMA, 3-hydroxypropylmercapturic acid), benzene (SPMA, S-phenylmercapturic acid), and acrylonitrile (CEMA, 2-cyanoethylmercapturic acid) for all groups. Levels of these BOE were typically lower in rats exposed to Regular and Fresh Menthol Heatstick aerosols than in rats exposed to 3R4F smoke and similar to the sham control. Other measures collected during the study included: food consumption and weight, plasma BOE, respiratory physiology, lung inflammation, hematology and clinical chemistry measures, and necropsy with gross pathology and histopathology.

Overall, the incidence of basal cell hyperplasia (nose and larynx) and squamous cell hyperplasia (nose and larynx) were similar in rats exposed to either Heatstick aerosols or 3R4F, while goblet cell hyperplasia/hypertrophy (lung) and macrophage aggregation (lung) were only observed in rats exposed to 3R4F. Hyperplasia, metaplasia, and immune cell infiltration are adaptive responses to acute stressors, which often reverse once the causative agent is removed. However, if the exposure continues, as with smoking, hyperplasia and metaplasia can be interpreted as pre-neoplastic changes while intra-alveolar macrophage aggregation can be an early indicator of fibrosis and goblet cell hyperplasia can be an early sign of chronic bronchitis. The applicant considers such findings to be adaptive as they partially reverse during the recovery period, yet the data suggest that not all effects are reversible.

37

b. 18-month Carcinogenicity Study with A/J mice The applicant conducted an 18-month carcinogenicity study where A/J mice were exposed to either Marlboro Heatstick aerosol, 3R4F smoke, or fresh air (sham control). Data supplied by the applicant suggest that male mice are more sensitive to the toxic effects of Marlboro Heatsticks than female mice; the study was halted in Month 15 for male mice. The applicant reports lung tumor incidence data during Month 5 and 10 in all groups of female mice; no lung tumor incidence data from male mice was provided.

The preliminary report, submitted after 10 months of data, showed that Female A/J mice repeatedly exposed to 3R4F exhibited significant hematological effects (i.e., increased RBCs, decreased WBCs, increased mediators of immune response), altered clinical chemistry, elevated mediators of immune response, elevated markers of inflammation in bronchial-alveolar lavage fluid (BALF), impaired lung function consistent with emphysema, histopathological changes in lungs, and elevated incidence neoplastic and pre- neoplastic lesions in the lungs when compared to sham controls. Female A/J mice repeatedly exposed to the same concentration of Regular Heatstick aerosols (i.e., 13.4 µg/L) commonly exhibited some changes to these same parameters, but the changes were typically less severe or transient.

The incidence of bronchioloalveolar adenoma during Month 10 is similar in female mice exposed to either 3R4F or Regular Heatstick aerosols (50% and 54.5%, respectively). However, the applicant reports the incidence of bronchioloalveolar adenoma during Month 10 to be numerically higher in sham controls (25%) than in female mice repeatedly exposed to the low- or high-concentration of Regular Heatstick aerosols (9.1% and 16.7%, respectively). No explanation for the lack of a dose-response relationship for Regular Heatstick aerosols and bronchioloalveolar adenoma incidence was provided.

The final report for this study was received September 4, 2018. The applicant concludes the study demonstrated no increase in lung cancer risk due to THS 2.2 aerosol exposure compared to sham group. Per the applicant, toxicity is limited to adaptive responses in the upper respiratory tract organs and stress- related responses to exposure, both of which were of lower severity compared to the mice exposed to 3R4F smoke.

c. Nicotine Pharmacokinetic (PK) Study with Rats The objective of this study was to model the inhalation of the nicotine-containing aerosol, conversion of nicotine to cotinine in the liver, distribution of nicotine and cotinine between the blood and liver compartments, and cotinine excretion. While the applicant reported trends in estimated values for t1/2 and Cmax, the PK model developed was not sensitive enough to detect significant group differences between test articles. Due to this variation, the applicant noted that the study power was not sufficient to detect significant group differences. This study does not provide relevant information for determining the health effects of Heatsticks; however, human PK studies were submitted and are more informative. These are reviewed in Section II.D.1.a.

d. Summary In vivo studies, such as the 90-day inhalation study, can provide important information about non-cancer toxicology endpoints, but are not generally sensitive enough to determine systemic toxicities from chronic tobacco product use. Despite that, the data submitted by the applicant indicate that sub-chronic exposure to Heatstick aerosols produce fewer or less severe histopathological changes than sub-chronic exposure to similar concentrations of 3R4F. Development of PK modeling of nicotine and cotinine sensitive enough to distinguish differences in nicotine and cotinine exposure will require the addition of sampling time points during inhalation exposure and model adjustments.

38

Similar to the in vitro studies, it is difficult to determine the carcinogenic potential of long-term exposure to Heatstick aerosols from these evaluations. The data suggest there is potential for carcinogenic effects from Heatstick aerosols, but at much higher exposure levels than required for CC smoke. The 18-month carcinogenicity study results reported by the applicant showed no increase in risk due to the Heatstick aerosol exposure compared to CC smoke and the changes noted were similar to the sham control group. How this correlates with clinical changes in humans is unknown.

4. Systems Toxicology Studies a. Acute and Repeated Exposure Studies with Human Organotypic Tissues The applicant submitted data from five separate in vitro organotypic studies assessing the effects of acute and repeated exposure to Regular HeatStick aerosols and 3R4F smoke on human gingival, buccal, nasal, bronchial, and coronary arterial epithelium cultures. Both 3R4F and Regular Heatstick aerosols produce toxicity (e.g., oxidative stress, DNA damage, increased proinflammatory mediators) in human gingival, bronchial, buccal, nasal, and small airway tissues, as well as epithelial tissues from human coronary arteries. The toxic effects produced by 3R4F smoke were generally more severe than those produced by Regular Heatstick aerosols or similar toxic effects were produced at much lower 3R4F smoke concentrations. However, there is variability in the 3R4F and Regular Heatstick aerosol concentrations and the post- exposure timepoints used in the applicant’s statistical analysis.

The experimental approach taken in these studies included using methods that are exploratory, have not been independently validated, and have unknown utility for regulatory use. The applicant attempts to extrapolate from acute exposure studies with naïve tissues that have little or no genetic variability to predict toxicity in a diverse population with a history of cigarette smoking. This limits the use of these data. Thus, this data does not significantly contribute to the overall toxicological profiles of the products under review in these applications.

b. ApoE-/- Mouse Switching Study The applicant conducted an 8-month switching and cessation study with female ApoE-/- mice; the report is in Section 7.5 of the submission. In this study, mice were exposed to 3R4F smoke, Regular Heatstick aerosols, or sham conditions 3 hours per day, 5 days per week, for 8 months. Additional groups of mice were exposed to 3R4F smoke under the same regimen but were switched to either Regular Heatstick aerosols (the “switching group”) or to filtered air (the “cessation group”) after 2 months. The study report includes information about biomarkers of HPHC exposure, hematologic effects, BALF, histopathology, lung function and volume, aortic arch morphometry, and tissue functional genomics evaluations (i.e., transcriptomics, lipidomics, and proteomics) performed at multiple time points.

This study was intended to model continued cigarette smoking vs. switching to Heatsticks vs. smoking cessation. There were limitations to the study design that affect interpretation of the data. Specifically, no male A/J mice were used in this study and the 3R4F exposure period for the switching group may have been too brief to allow determination of how Heatstick aerosols affect progression of the toxic effects caused by cigarette smoke. The histopathologic changes seen in the switching group were similar to cessation, but it is not clear whether a longer smoking period would lead to the same result. However, the overall pattern of changes related to switching from 3R4F to Heatstick was positive. Although the results from Heatstick exposure were not the same as sham (or smoking abstinence), some effects seen after 3R4F exposure were either less prominent or occurred less frequently in the mice that “switched” to Heatsticks, indicating that

39

switching to IQOS could be beneficial to smokers. Dual exposure to cigarette smoke and Heatstick aerosol was not evaluated.

5. Nonclinical Evaluation of Carbon Monoxide from Heatstick Aerosol Reconstituted tobacco can produce high levels of CO and nitrogen oxides during combustion. However, despite containing only reconstituted tobacco, the three Heatsticks produce much lower CO compared to regular CC. The applicant considered the(b) (4) (b) (4) The product specifications for CO, as set by the applicant, are (b) (4) /stick for PM0000424 and (b) (4) /stick for PM0000425 and PM0000426 – as measured in the respective Heatstick aerosols generated by a CI smoking regimen.

The established product specifications were achieved. The measured CO produced by Heatstick aerosol was 0.2-0.3 mg CO/Heatstick (ISO regimen) and 0.3-0.5 mg CO/Heatstick (CI regimen). This is 99% lower than the CO level in combusted cigarette smoke (~28-33 mg/cig CI regimen). During the inspection of the Bologna (Italy) facility, CoAs for batch release collected included the specifications and the results obtained for one batch of each of the three products; measured CO levels were 0.41-0.43mg/Heatstick.

In the ApoE-/- Mouse Switching Study, the applicant indicated that carboxyhemoglobin (COHb) levels in the blood were measured during Months 2, 4, and 7. Mean COHb levels were consistently higher in mice from the 3R4F group than in mice from the switching, cessation, or sham control groups, though the applicant did not provide a statistical analysis of these data. The applicant also indicates that mice in the cessation and sham control groups had similar COHb levels but did not indicate whether COHb levels in mice from the switching group were significantly different from controls. See Figure 3 below.

Figure 3: Carboxyhemoglobin Levels after Repeated Exposure to 3R4F, Regular Heatstick Aerosols, or Sham Conditions Source: “15015_CVD_Resp_ApoE_SW_SR_Part 1”, Figure 9, page 59

In the 90-day nose-only Inhalation studies conducted in rats, COHb levels in all Heatstick Regular and Menthol exposure groups were equivalent to sham controls while rats exposed to 3R4F smoke were

40

substantially elevated. During the 42-day recovery period, COHb levels in the 3R4F exposed group returned to normal levels.

Data source: “15006 THS SR Part 1”, page 170 and “15025 THS SR Part 2, page 48; Measurement made at day 47 (week 6) for the Heatstick menthol study; measurement date not noted for Heatstick regular study Figure 4: Plasma Carboxyhemoglobin Levels at Week 6 for 90-day Rat Inhalation Study

6. Summary of Toxicological Findings The toxicology review provides this summary of key findings: • There are HPHC reductions in Heatstick aerosols relative to smoke from 3R4F reference cigarette and 30 commercially available CCs. o In the PMI-58 study, the 54 measured HPHC levels were reduced by 54.4-99.9% on a per stick basis when compared to 3R4F smoke. Machine-generated nicotine yields were reduced 35.9- 39.4%, but clinical data indicates human CC smokers and Heatstick users absorb similar amounts of nicotine. o For 18 of these compounds, the applicant determined that yields in Heatstick aerosols were reduced by 40-99.8% when compared to the mean of 31 CC commercially available in the U.S. (on a per stick basis). o Side-stream aerosol from Heatsticks emit detectable levels of some HPHCs, but levels are significantly lower than emissions from CC. • The non-targeted differential screening assay indicates that Heatstick aerosols contain four probable of possible carcinogenic chemicals that are unique to IQOS or present in higher levels than 3R4F smoke. The aerosols also contain 15 other chemicals that are possibly genotoxic and 20 more GRAS compounds that have potential health effects. When balanced against the significant decreases in the number of HPHCs and HPHC yields, however, these chemicals, which are present at very low levels, do not raise significant concerns from a public health perspective. • TPM from Heatsticks did not produce a positive response for mutagenicity in the Ames assay at any dose tested, either with or without metabolic activation. • The NRU assay results indicate that cytotoxicity for the Heatstick TPM and GVP are reduced by ~95% (per stick) and ~90% when normalized to nicotine content and compared to 3R4F. • The in vitro MLA shows a biologically relevant mutagenic response in mammalian cells from Heatstick aerosols and 3R4F smoke (both with and without metabolic activation) after 4 hours of exposure, but the minimum Heatstick TPM concentration required to produce this positive result was 15-30 times greater than the concentration required for the 3R4F cigarette. The applicant

41

indicates that this difference is an index of mutagenic potency; however, this concept is not supported by guidance documents from major public health resources (e.g., OECD, ICH, Health Canada, EPA). The applicant’s conclusion that both Heatsticks and 3R4F have cytotoxic and mutagenic potential appears accurate; however, CTP agrees with the public health groups that the level of a substance required to produce a genotoxic effect may not be an accurate indicator of mutagenic potency. • Heatstick aerosols generally produced fewer pathophysiological changes and adverse effects than 3R4F smoke in organotypic studies. However, the experimental approach taken in these studies included using methods that are exploratory, have not been independently validated, and have unknown utility for regulatory use. Thus, this data does not significantly contribute to the overall toxicological profiles of the products under review in these applications. • The 90-day inhalation study in rats showed that changes related to Heatstick aerosol exposures were not observed or much less severe than changes noted due to 3R4F exposure. • An 8-month mouse switching/cessation study suggested switching to Heatsticks after a short period of cigarette smoke exposure led to histopathological changes similar to smoking cessation. However, as noted in the discussion above, study design limitations preclude reliance on these data. • The 18-month carcinogenicity study shows that the incidence of neoplastic lesions appeared to be higher in some groups exposed to either Heatstick aerosols or reference cigarette smoke compared to the sham control group. However, other evidence indicates repeated exposure to Heatstick aerosols produced fewer histopathological changes than repeated exposure to 3R4F smoke. The applicant concludes this long-term study demonstrated no increase in lung cancer risk due to THS 2.2 aerosol exposure compared to sham group. Per the applicant, toxicity is limited to adaptive responses in the upper respiratory tract organs and stress-related responses to exposure, both of which were of lower severity compared to the mice exposed to 3R4F smoke.

As TPL, I agree with the toxicology review conclusion. After consideration of all the toxicological data presented, the demonstrated reductions in measured HPHC exposures and reduced histopathological changes with reduced potential for atherosclerotic effects indicate the potential for a relative benefit compared to CC for smokers who switch completely to IQOS. The toxicological profiles of the Marlboro, Smooth Menthol, and Fresh Menthol Heatsticks are essentially identical; the only difference in the Heatsticks is the quantity of menthol added to each product. Marlboro, Smooth Menthol, and Fresh Menthol Heatstick aerosols contain some chemicals which are different from those found in CC. Although some of the chemicals are genotoxic or cytotoxic, these chemicals are present in very low levels and potential effects are outweighed by the substantial decrease in the number and levels of HPHCs found in CC.

D. Behavioral and Clinical Pharmacological Assessment 1. Pharmacokinetics, Exposure/Response and Clinical Pharmacology a. Pharmacokinetic/Pharmacodynamic (PK/PD) Studies Four single-use, randomized, 2-period, 4-sequence cross-over studies of PK/PD (ZRHR-PK-01-EU, ZRHR-PK- 02-JP, ZRHM-PK-05-JP, ZRHM-PK-06-US) were conducted to assess and compare the rate and extent of nicotine uptake in participants using THS 2.2 compared to smoking own-brand CC and nicotine replacement therapy (NRT) products. The NRT product varied by the location of the study. The primary PK parameters in these studies were maximum nicotine plasma concentration (Cmax) and area under the nicotine plasma concentration vs. time curve from time zero to the last observation (AUC0-last). Products were administered following ≥ 24 hours nicotine abstinence; participants used their assigned product once: one Heatstick, one CC ad libitum, two 1 mg sprays of nicotine nasal spray (NNS) dosed at one spray per nostril or one piece of 2

42

mg used for 30 minutes. As a PD response measure, exposure to CO was assessed as exhaled CO (eCO) and as carboxyhemoglobin (COHb) blood saturation. Measures related to craving (QSU-Brief) and reinforcement (MCEQ) were part of the PD assessment.

The clinical pharmacology review analyzed the PK/PD studies using non-compartmental data analysis and population PK data analysis. In the two Japanese studies (ZRHR-PK-02-JP and ZRHM-PK-05-JP), the geometric mean (GM) values of Cmax and AUC0-last calculated by noncompartmental analysis were similar between THS 2.2 and CC for both Regular and Menthol products. In the Irish and U.S. studies (ZRHR-PK-01-EU and ZRHM- PK-06-US), the GM values of nicotine exposure parameters in the THS 2.2 arm were lower than in the CC arm for both Cmax and AUC0-last. The Irish and U.S. studies resulted in lower nicotine exposure in the THS 2.2 arm than the Japanese studies. In all studies, the CC arms had similar GM Cmax values while the GM AUC0-last values were highest in the U.S. study. This finding may be explained by the CC characteristics (no limit of 1 mg ISO nicotine level in own brand CC as used in three other studies), prolonged duration of ad libitum CC use (longer time to reach Cmax, Tmax values), genetic differences, and differences in puffing behavior between the populations.

Four randomized, controlled, open-label, 3-arm parallel group studies (ZRHR REXC-03-EU, ZRHR-REXC-04-JP, ZRHM-REXA-07-JP, ZRHM-REXA-08-US) were conducted with the primary aim to investigate systemic exposure to BOE in smokers who switched to THS 2.2, continued to smoke CC, or abstained from smoking (SA). Nicotine plasma concentrations were measured as a secondary objective once daily and frequently on Day 5 of the confinement period and were compared across the three arms. Two REX studies (ZRHM-REXA- 07-JP and ZRHM-REXA-08-US) had an 85-day ambulatory phase extension after the 5-day confinement period for total study duration of 90 days. Plasma nicotine and cotinine concentrations were measured in the mornings of Days 30, 60, and 90. In addition, urinary nicotine equivalents (NEQ) (nicotine, cotinine, 3′- Hydroxycotinine and their glucuronides) were measured in 24-hour urine daily on Days 1-5, and on Days 30, 60, and 90 in the extended studies. The GM values of nicotine and cotinine concentrations in plasma and GM values of urinary NEQ were similar between the THS 2.2 and CC arms.

The population PK modeling of nicotine (ZRH-POP-PK-01) was conducted with four goals: • Describe the nicotine PK with physiologic parameters (clearance, volume of distribution) • Assess sources of variability in nicotine PK parameters • Assess the predictive performance of the model • Distinguish between the exposure due to product use and background exposure

The plasma concentration data were combined from all studies following use of nicotine-containing products (i.e., THS 2.2 and its comparators: CC, nicotine gum, and NNS). The analysis dataset included all participants that used a product at least once and had at least one measurable nicotine plasma concentration. It included baseline demographic variables, daily cigarette consumption at baseline, and some product-related information (e.g., type of product, presence of menthol, and nicotine dose). Model evaluation and selection of the base model were adequately performed using the standard statistical criteria of goodness-of-fit.

The typical initial and terminal half-lives of nicotine were 1.35 hours and 17 hours, respectively. These values are markedly longer than previously published; however, they appear to be reasonably assessed. The previous reports of a shorter terminal half-life may have failed to capture the terminal phase of nicotine PK due to the shorter sampling period as well as lower assay sensitivity (most analytical methods report low limit of quantification [LLOQ] from 0.5 to 1 ng/mL; the applicant reported LLOQ as 0.2 ng/mL). The model adequately captured the median nicotine PK profile for every product in both periods of studies. The

43

observed 90th percentile was generally within the prediction interval. The lower nicotine concentration range was difficult to predict, possibly due to fewer observations in that range.

b. Intrinsic and Extrinsic Factors Affecting Nicotine Pharmacokinetics The covariate analysis for the selected base model was conducted in accordance with the FDA Guidance on Population Pharmacokinetics. The covariate analysis of the final model explained the effects of the following intrinsic factors: baseline body weight, CYP2A6 activity, sex and race. Clearance in female participants was 26% higher than in males. In addition, clearance was positively correlated with the baseline activity of cytochrome P450 2A6 isoform (CYP2A6): a doubling in CYP2A6 activity appeared to increase nicotine clearance by 25%. In addition, the effect of African American race increased C0 by 50%.

The covariate analysis of the final model evaluated the effects of the following extrinsic factors: the nature of product, the nicotine ISO yield, and the presence of menthol. The bioavailability of CC, NNS, and nicotine gum relative to THS 2.2 was 102%, 24%, and 61%, respectively. Bioavailability decreased with increasing nicotine ISO yield (a doubling in nicotine ISO yield would result in a 33% relative decrease) and with increasing body weight (a 10% increase in body weight would result in a 6.6% relative decrease). The apparent central volume of distribution was 9.5% larger with Menthol than with Regular products. The absorption duration (Tdur) from the nicotine gum lasted 45 minutes vs. 5.3 minutes for other products. Menthol products marginally, but significantly, increased Tdur by 5% vs. Regular products. Menthol had no effect on bioavailability relative to THS and apparent clearance, which are the determinants of plasma exposure (AUC) at a given nicotine ISO yield.

The population PK model structure allowed the separate derivation of exposure parameters based on both total and background-adjusted concentrations of nicotine from four different nicotine-containing products (CC, THS 2.2, NNS and NRT gum) with different routes of administration (inhalation, oral, and nasal) in different populations (American, European and Japanese).

c. Nicotine Equivalents (NEQ) in Urine NEQ measured in 24-hour urine is often used to estimate nicotine exposure in clinical studies since it reflects at least 80% of the daily nicotine uptake in smokers.65 NEQ consists of nicotine and five major metabolites: nicotine-glucuronide, cotinine and its glucuronide, trans-3′-hydroxycotinine and its glucuronide. Urinary NEQ adjusted for creatinine were measured daily in the confinement period of all REX studies and on Days 30, 60, and 90 in the ambulatory period of studies ZRHM-REXA-07-JP and ZRHM-REXA-08-US.

The 24-hour NEQ urinary concentrations adjusted for creatinine between the THS 2.2 and CC arms were similar on each day of studies ZRHR-REXC-03-EU and ZRHR-REXC-04-JP. Nicotine exposures measured by NEQ between mTHS 2.2 and menthol CC (mCC) arms during the confinement period were also similar in study ZRHM-REXA-07-JP. In study ZRHM-REXA-08-US, NEQ was slightly lower after the use of mTHS 2.2 compared to the mCC arm. On Day 5, the differences in NEQ in each of the studies were not statistically significant. During the ambulatory period of studies ZRHM-REXA-07-JP and ZRHM-REXA-08-US, differences in NEQ were not statistically different at any day with confidence intervals (CIs) including zero.

65 Wang, J., Liang, Q., Mendes, P., & Sarkar, M. (2011). Is 24h nicotine equivalents a surrogate for smoke exposure based on its relationship with other biomarkers of exposure? Biomarkers, 16(2), 144-154.

44

d. Summary Overall, the population PK model accounts for the variability in nicotine PK among all clinical studies with consideration of the influence of the statistically significant intrinsic (body weight, CYP2A6 activity, sex, and race) and extrinsic (nicotine ISO yield, presence of menthol) factors. Based on this model, nicotine PK in smokers who switched to THS 2.2 is similar to those who continued to smoke CC.

2. Behavioral Pharmacology a. Use Behavior and Topography The applicant conducted four PK/PD studies, four REX studies, and one Actual Use study. These studies collected data on product use behavior (e.g., daily consumption, topography), subjective effects of product use on nicotine dependence, product satisfaction and reinforcement, and product misuse at different durations of use (single use, 5 days, 6 weeks, and 3 months). Tobacco product use behavior plays a critical role in exposure to nicotine and other constituents and can signal compensatory behaviors. Subjective effects can indicate the likelihood of continued use of a product and abuse liability. Although these measures can be considered from a public health perspective, for both users and nonusers, the provided studies only include current daily CC smokers, who may also use other tobacco products.

THS 2.2 topography is limited by its intrinsic properties, which limit the number of puffs to 14 and smoking duration to a maximum of six minutes. (b) (4) The applicant states tha (b) (4)

In the CC arm of the REX studies, topography measures were generally stable over time. Differences in topography over time were expected when consumers switched from CC to THS due to adaptation to the new product. Compared to the CC arm, participants in the THS 2.2 arm took more puffs (three of four studies), had a shorter smoking duration (two of four studies), had a higher puff frequency (four of four studies), and did not differ in total puff volume. The applicant attributes these differences to adaptation to the intrinsic properties of the new product as well as to differences in nicotine delivery, product satisfaction, ritual, sensory factors, and taste. Findings were similar for Menthol and Regular flavor products.

b. Product Use/Consumption Tobacco product consumption rates play a critical role in exposure to nicotine and other constituents. Level of consumption (secondary outcome) was measured as the number of CC or Heatsticks used per day. In the conducted studies, product use was recorded by study staff, documented by product dispensation and collection, or recorded by participants in a diary. Use of other nicotine products was recorded (yes/no), but frequency was not measured. Self-report data is susceptible to inaccuracies. Compliance to the assigned study product was controlled during confinement periods. During ambulatory periods participants were instructed to use their assigned product but noncompliance did not result in study removal and accuracy of reporting is likely not optimal.

In the 5- and 90-day REX studies, both THS 2.2 and CC arms showed minimal changes in product use over time (see Figure 5 below). Inclusion of dual use data in the analyses (i.e., combined CC and Heatstick use) did not have a major effect on changes in consumption for the THS 2.2 arm during the confinement or

45

ambulatory periods. These results were consistent across the REX studies, despite differences in environment (confined vs. ambulatory), populations, dual/exclusive use, and flavor (Menthol vs. Regular).

Figure 5: Product Consumption (Number of Products Used) During the REX Studies Source: Section 6.2.2 Note: Day 0 = Baseline

One single-group, prospective observational study (THA-PBA-07-US) was conducted with a primary aim to investigate how U.S. adult daily smokers of CC used THS 2.2 in near to real-world conditions (i.e., naturalistic setting). During this Actual Use study, participants were exposed to IQOS system material and those with positive “intention to use” were invited to participate in the study. After a one-week baseline period where they smoked own brand CC and potentially used other tobacco products as well as NRTs, participants received the IQOS system kit and instructions for use. Menthol preferences were honored. Participants could consume THS 2.2, CC, and other tobacco products as well as NRTs ad libitum for a period of six weeks. Participants completed interviews prior to the baseline period, after the baseline period, and every two weeks during the six-week observational period. Relevant secondary endpoints included product consumption and THS 2.2 misuse; data on hypothetical purchasing of THS 2.2 were also reported, but not listed as an endpoint.

In the Actual Use study, the average number of products (CC and Heatsticks combined) used per day was slightly lower during the observational period compared to baseline (9.3 products per day vs. 10.2 CPD) for the overall sample (N=987). This decrease was similar for participants who used THS 2.2 > 70% of the time (N=141): participants used 8.1 products per day during the observational period compared to 9.0 CPD at baseline. Findings were similar for users of Menthol and Regular products. Data were descriptive, and no statistical analysis was provided.

Several aspects of study design may have contributed to differences in product consumption across study arms of the REX studies, as well as overall use and use patterns in the Actual Use study. Participants received the THS 2.2 for free while those in the CC arm of the REX studies continued to pay for their cigarettes. Information provided to the participants about THS 2.2 might have influenced perceptions of product safety profiles, thereby contributing to differences in use behavior and exposure. In addition to differences in

46

labeling on CC compared to THS 2.2 Heatstick packages, participants in different REX studies were exposed to different product information in the informed consent documents. For example, the informed consent for the ZRHR-REXC-04-JP study suggests the investigational product is less harmful than CC, stating, “a number of clinical studies have been conducted… with the previous version of the device (THS 1.0 and THS 2.1)… showed reductions in exposure to selected smoke constituents in subjects who used the THS 1.0 or THS 2.1, as compared to subjects continuing smoking conventional cigarettes,” whereas this language was not included in the informed consent for study ZRHM-REXA-08-US.

c. Product Acceptability Dual use of THS 2.2 and CC was evident in the REX studies and the Actual Use study. During the last 30 days of the ambulatory period in the REX studies, dual use was low in the Japanese study; at least 84.6% reported using THS 2.2 exclusively (100%) and at least 85.9% reported using THS 2.2 > 95% of the time. In contrast, dual use was higher in the U.S. REX study; at least 55.0% reported using THS 2.2 exclusively, and at least 63.8% reported using THS 2.2 > 95% of the time. Notably, only 7.5% of cigarette smokers reported using THS 2.2 > 95% of the time at the end of the Actual Use study. The higher rates of complete switching in the REX studies may have occurred because participants were instructed to use the THS 2.2 exclusively and were confined and monitored to ensure compliance during the first five days of the study. Concurrent use of other tobacco products or NRTs was not considered in this analysis.

A further analysis considered use of tobacco products other than mCC in the mTHS 2.2 arm of the REX studies. To meet criteria for “compliant” exclusive use, participants were not allowed to use “any nicotine or tobacco-containing product other than the assigned product.” In the extended REX study conducted in Japan, 65 of 78 participants (83.3%) met this stricter exclusive use criteria during the last 30 days of the ambulatory period; however, in the U.S. extended REX study only 41 of 80 participants (51.3%) met this stricter exclusive use criteria during the last 30 days of the ambulatory period.

d. Abuse Liability The degree to which current smokers and nonsmokers are likely to use the product and become addicted or dependent was evaluated by self-report questionnaires. Participants reported perceived effects of THS 2.2 on nicotine dependence and dependence symptoms (e.g., craving, withdrawal), reward/reinforcement following use, product valuation (i.e., hypothetical purchasing) compared to own-brand CC, and product misuse.

The Questionnaire of Smoking Urges – Brief (QSU-Brief) measures craving from two perspectives: (1) the intention and desire to smoke and anticipation of positive effects from smoking (positive reinforcement) and (2) the anticipation of relief from negative affect and , and urgent and overwhelming desire to smoke (negative reinforcement). In the four PK/PD studies, relief from craving (QSU-Brief) showed a similar time curve following both THS 2.2 and CC arms: highest smoking urge prior to use, sharp decline following use, and continued decline to approach baseline over 12 hours. In the four REX studies, relief from craving (QSU-Brief) did not differ significantly between the THS 2.2 and CC arms and remained stable throughout the study. Findings were similar for Menthol and Regular products.

47

Figure 6: Arithmetic Mean and 95% CI Total Scores for QSU-Brief Total Source: Section 7.3.1 ZRHM-PK-06-US Final Study Report, Section 7.3.1 ZRHM-REXA-08-US Final Study Report Note. QSU-Brief scores reported on a 7-point scale. Higher values indicate greater intensity of urge.

The Minnesota Nicotine Withdrawal Scale – Revised (MNWS) measures relief from withdrawal based, in part, on withdrawal symptoms identified in the Diagnostic and Statistical Manual of Mental Disorders (DSM) and the International Classification of Diseases – 10th Edition (ICD-10) and a question on craving. In the REX studies, MNWS was administered prior to product use and reflected the previous day’s experience; these data were exploratory. No differences were found between study arms, at Days 5 or 90, on the MNWS questionnaire.

The Fagerström Test for Nicotine Dependence (FTND) is a test of physical dependence. In the two 90-day REX studies, FTND was administered at baseline and at the end of the ambulatory period (Day 90). No statistical analyses were conducted. No differences in dependence severity were found between THS 2.2 and CC arms at Day 90; both arms showed no change in the Japanese study or equally reduced symptom severity in the U.S. study.

The Prochaska “Stage of Change” Questionnaire assessed participants’ past-year quit attempts and quit intent in the 30 days or six months. This questionnaire was administered in the U.S. REX study at screening, baseline (Day -2), Day 30, Day 60, and Day 90. No statistical analysis was conducted. At Day 90, some participants in both study arms reported “seriously thinking of quitting smoking” within the next 30 days (10.4% THS 2.2 arm, 13.8% CC arm).

Reinforcing and aversive effects were measured using a self-report questionnaire (Modified Cigarette Evaluation Questionnaire [MCEQ]) whose subscales include: smoking satisfaction, psychological reward, aversion, enjoyment of respiratory tract sensations, and craving reduction. In the PK/PD studies, MCEQ was administered after product use and was a secondary outcome measure. In the REX studies, MCEQ was administered at the end of the day (8-11 pm). Data were exploratory and descriptive.

In the PK/PD studies, when compared to CC, THS 2.2 had significantly lower MCEQ ratings on four of five MCEQ subscales: craving reduction (two studies), enjoyment of respiratory tract sensations (four studies), psychological reward (one study), and smoking satisfaction (four studies). Scores on the aversion subscale

48

showed no difference between THS 2.2 and CC. Findings were similar for Menthol and Regular products. In the REX studies, THS 2.2 had significantly lower ratings than CC on Day 5 (end of confinement) but no differences were found on Day 90 for the two extended REX studies. This may reflect a learning or adaptation period to the new product.

In the Actual Use study, participants were asked at the end of the observational period (Week 6) about their likelihood to purchase the IQOS system “if the iQOS device were available for $79.99 and a pack of Marlboro HeatSticks were available at a price comparable to a pack of Marlboro cigarettes.” In the overall sample (N=987), nearly 20% of participants reported that they “probably or definitely” would buy IQOS. Findings were similar based on menthol/non-menthol preference, across age groups, and across baseline smoking rates. In a subsample of participants who used THS 2.2 > 70% of the time (Week 6, N=138), nearly 50% reported they “probably or definitely” would buy IQOS. Although descriptive data were provided, this was not listed as an outcome measure. It is unclear if participants assumed that they had already owned the IQOS system and were being asked about buying Heatsticks only, or if they assumed the question was referring to purchasing both the IQOS system and Heatsticks.

There are limitations to interpretation of questionnaire data, including: • Recall bias and other inaccuracies associated with self-report • No assessment of the relationship between subjective measures and behavior • No data was provided on the validation of translated versions of the questionnaires used in studies not conducted in the U.S. • The QSU-Brief was not modified to replace references to cigarettes with Heatsticks/IQOS/THS 2.2, so it is unclear if participants were aware of which product was being asked about • Scoring of the MNWS differed between studies • Intent to quit results have limited generalizability to participants who have quit intent of greater than six months (inclusion criteria) and are difficult to interpret because it was not clear whether the intent to quit refers to quitting all tobacco vs. switching completely from smoking CC to THS 2.2

e. Summary As noted above, there are limitations for self-reported data; however, this is an informative method for obtaining information of this type. These are validated questionnaires for outpatient tobacco research and are commonly used in studies of tobacco use behaviors. Systemic nicotine exposure was similar after single and multiple uses of THS 2.2 and CC (both Regular and Menthol). Nicotine exposures appear sufficient to provide user satisfaction, which can facilitate partial or complete switching to THS 2.2. THS 2.2 use rates were similar to CC use rates. Self-report questionnaires found that THS 2.2 produces reinforcing effects reaching or close to levels of CC reinforcement. Likeability scores for THS 2.2 increased over the 90-day period for those who used it more consistently which may indicate the need for an “adjustment” or transition phase from CC, (i.e., dual use).

The data indicate that THS 2.2 has addictive potential and abuse liability similar to CC. This is important as it signifies THS 2.2 can provide an adequate nicotine source for dependent populations, including current CC users; however, there is also a risk tobacco-naïve new THS users will develop nicotine addiction.

3. Summary of Behavioral and Clinical Pharmacology Findings The behavioral and clinical pharmacology (BCP) review concludes that the similar systemic exposure to nicotine as well as similar use rates, reinforcement, and withdrawal/craving reduction profiles between THS 2.2 and CC suggest a similar abuse liability of these tobacco products. Thus THS 2.2 use may sustain

49

addiction to a similar level as CC in current smokers and have a similar risk of nicotine addiction as CC in nonsmokers.

As TPL, I agree with the BCP review conclusions. IQOS provides nicotine at a high enough level to satisfy the withdrawal and craving symptoms of current smokers. The nicotine levels do pose an addiction risk for non- tobacco users who initiate use of these products; however, the risk is no higher than for other, currently available, tobacco products.

E. Individual Health Impact 1. Biomarkers of Exposure (BOE) The BOE selected correspond with 14 HPHCs identified by FDA as being found in cigarette smoke or filler. 1- hydroxypyrene is considered a general proxy for PAHs. HEMA, and the aromatic amine o-Toluidine were also measured. Exposures to acetaldehyde, formaldehyde, isoprene and ammonia were not assessed as biomarkers as there are not suitable biomarkers for these exposures.

The four clinical REX studies assessed changes in systemic exposure of HPHCs and their metabolites in smokers who switched to THS 2.2 or abstained from smoking CC during the 5-day confinement period. Two of these studies (ZRHM-REXA-07-JP and ZRHM-REXA-08-US) had an ambulatory period extension of 85 days. All REX investigations were randomized, controlled, open-label, 3-arm parallel group studies. Studies ZRHR-REXC-03-EU and ZRHR-REXC-04-JP investigated Marlboro (non-mentholated) Heatsticks, whereas studies ZRHM-REXA-07-JP and ZRHM-REXA-08-US investigated Menthol Heatsticks. Participants who were “willing and able” to use IQOS after a demonstration were randomized to one of three study arms in a 2:1:1 ratio by sex and past month mean smoking rate (10-19 CPD vs >19 CPD at screening).

During the five-day confinement period, assigned products were used ad libitum from 6:30 am to 11:00 pm. Dual use of THS 2.2, CC, and other tobacco products as well as NRTs was not permitted during the confinement period. Participants in the SA arm were not provided with “medication supportive for smoking abstinence” during confinement. Two of the four studies followed participants for a prolonged period (90 days) in an ambulatory setting (i.e., home environment, near to real-world conditions) after the 5-day confinement period to evaluate if the results achieved under controlled conditions were maintained. Participants in the SA arm were instructed to remain abstinent with or without NRT, for which they were reimbursed, during the ambulatory period. Data were collected daily during confinement and on days 30, 60, and 90 during the ambulatory period. Nicotine abstinence was not required prior to the assessments during Day 30, 60, or 90 visits.

Exposure to 16 HPHCs (including nicotine for a total of 17) were evaluated in the clinical studies, by either measuring the parent compound (e.g., 4-aminobiphenyl), by measuring one or several of their metabolites, or by using a surrogate BOE as representative of a chemical class of compounds. The list of all selected biomarkers, their classes, and related major toxicities is presented in Table 3. An assessment of markers reflecting an overall exposure to HPHCs, was also performed, including: • Activity of CYP1A2, an enzyme which can be induced by polycyclic aromatic amines • Urine mutagenicity potential, a measure to assess exposure to various carcinogenic/mutagenic substances These markers are not associated with a specific HPHC.

50

Figure 7: Percent Change from Baseline of Geometric Mean Levels and 95% CIs at Day 90 in ZRHM-REXA-08-US Source: Application Section 6.1.3.2 Note. Because of the limited number of participants in the SA arm and outliers, percent change from Baseline values for total NNAL and HEMA are reported as median (and Q1; Q3) for both for the THS 2.2 and SA arms.

The reductions of systemic exposure to 15 HPHCs or their biomarkers seen after switching from CC to THS 2.2 in all REX studies were statistically significant. The exposures to S-BMA, a biomarker of o-toluene, were not different among the THS 2.2, CC, and SA arms. This is most likely due to environmental sources of exposure to this biomarker. The profiles of decline of the other BOE observed in the THS 2.2 arm were to similar levels observed in the SA arm.

The only BOE with urinary levels that were slightly higher in the THS 2.2 than the SA arms, throughout the whole exposure period, were 3-HPMA and total NNN; however, the urine levels for these BOE in THS 2.2 were lower than in the CC arms. A smaller reduction in urinary 3-HPMA was reported in the Japanese studies (ZRHR-REXC-04-JP by 47%; ZRHM-REXA-07-JP by 49%), probably due to the presence of acrolein in non-combusted tobacco and its possible formation from glycerin at temperatures below 400Co. From the studies with ambulatory periods, the Japanese study (ZRHM-REXA-07-JP) had high compliance and thus provided convincing evidence of BOE reduction. Compliance in the SA arm of the U.S. study (ZRHM-REXA-08- US) was poor and the variability was high; results from the SA arm of this study may be less reliable. Details for primary endpoint BOE measures are discussed below.

FDA statistical reviewers analyzed the results of the four PK/PD studies and the BOE measures. Their statistical analyses demonstrated a statistically significant 50% or more reduction in levels on Day 5 of the studies for the primary BOE: MHBMA, 3-HPMA, S-PMA, COHb; and for total NNAL (urine) on Day 90 when comparing THS to CC. Reductions associated with the secondary BOE on Day 90 of the studies ranged from 25% to 90%.

52

a. COHb – BOE for Carbon Monoxide At baseline, COHb levels ranged from 4.65% to 6.66% across the REX studies. A normal COHb level for non- smokers is < 2% and for smokers is 5-13%.66 By Day 5 across all four studies, COHb decreased in study participants randomized to switch to THS 2.2 (1.06-2.48%) and SA study arms (0.99-2.5%). In contrast, COHb at Day 5 stayed similar to baseline levels in smokers who continued to smoke cigarettes (CC study arms) (4.5-6.07%). On Day 90 of the extended REX studies, COHb levels ranged from 2.66% to 2.97% and from 2.84% to 3.04% in the THS 2.2 and SA arms, respectively.67 The COHb levels in the CC arms remained generally unchanged, ranging from 5.62% to 5.73% at Day 90.

Figure 8: Geometric Mean and 95% CIs of COHb Concentrations (%) Source: Section 7.3.1

b. Monohydroxybutenyl-mercapturic acid (MHBMA) – BOE for 1,3-Butadiene 1,3-butadiene is a as well as a respiratory and reproductive or developmental toxicant. Baseline values of MHBMA, the BOE for 1,3-butadiene, varied between studies from 490 pg/mg creatinine in study ZRHR-REXC-04-JP to a mean level of 2317 pg/mg creatinine in study ZRHR-REXC-03-EU. This high variability may be explained by differences in CC between geographical locations, as well as differences in daily CC consumption, which tends to be substantially higher in Poland.

In the THS 2.2 arm, urinary MHBMA concentrations decreased to minimal values over a two-day period in all studies and were similar to values measured in the SA arm. In the two ambulatory studies, this decrease was maintained with percentage decreases from baseline to Day 90 ranging from 64% to 73% in the THS 2.2 arms and 27% to 66% in the SA arms. Participants who continued with their own CC brand showed variable levels across studies, from a relative decrease compared to baseline of 12% in the Japanese study to an increase of 36% in the U.S. study.

66 Prockop, Leon D; Chichkova, Rossitza I; Carbon monoxide intoxication: an updated review, Journal of the neurological sciences. 2007, Vol.262(1-2), p.122-130. 67 CO levels would be expected to be higher in a community environment than in a confined laboratory setting. The levels of CO were not statistically different between THS and SA arms and were well within the range for normal environmental CO.

53

c. 3-Hydroxypropylmercapturic acid (3-HPMA) – BOE for Acrolein Acrolein is a respiratory and cardiovascular toxicant. At baseline, urinary levels of 3-HPMA, adjusted for creatinine, were similar overall among studies and across arms. In participants who switched to THS 2.2 as well as those in the SA arm, a rapid reduction of 24-hour urinary 3-HPMA compared to CC use at baseline was found in all four REX studies. Urinary 3-HPMA levels during the confinement periods in all four studies decreased by 47% to 59% from baseline at Day 5 for the THS 2.2 arms and 65% to 76% for the SA arms. In ambulatory studies these reductions were maintained, to a lesser magnitude, with Day 90 ranging between 37% and 54% in the THS 2.2 arms and 48% to 57% in the SA arms. A significant reduction in acrolein exposure was achieved for Menthol and Regular THS 2.2 as well as for both Caucasian and Asian ethnicities (p’s < 0.001).

The applicant explains that the higher levels of 3-HPMA in the THS 2.2 arm compared to the SA arm may be due to residual acrolein in THS 2.2 that can be produced within a relatively low temperature range. Acrolein is naturally present in tobacco and is further produced by combustion. In non-combusted tobacco, it may also be formed through heating of glycerin, a constituent that is present in THS 2.2. The reduction of 3- HPMA urinary concentrations in smokers who switched to THS 2.2 was statistically significant at Day 5, with a sustained reduction on Day 90.

d. S-phenylmercapturic acid (S-PMA) – BOE for Benzene Benzene is a carcinogen as well as a cardiovascular and reproductive or developmental toxicant. At baseline, urinary S-PMA concentrations adjusted to creatinine ranged from 784 to 2765 pg/mg creatinine, with the highest levels observed in the Polish population and the lowest levels in the Japanese populations. (The 2.6- fold differences in S-PMA urinary concentrations were explained by the lower emission profiles in Japanese CC and the prevalence of heavy smokers in the Polish study; this rationale appears acceptable.)

S-PMA decreased during the confinement period between 77% and 92% for THS 2.2 and between 84% and 92% for the SA arms. In the ambulatory periods, urinary S-PMA concentrations in the THS 2.2 arm were reduced from baseline by 81% and 65% in ZRHM-REXA-07-JP and ZRHM-REXA-08-US studies, respectively. After Day 30, the urinary S-PMA concentrations in the U.S. study’s THS 2.2 arm slightly increased compared to the SA arm; however, the reduction of S-PMA systemic exposures in the THS 2.2 arms compared to CC arms were statistically significant in each study.

e. Total NNAL - BOE for NNK Total 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (total NNAL) is a major metabolite of nitrosamine ketone (NNK); both are classified as carcinogens. At baseline, urinary levels of total NNAL concentrations adjusted to creatinine were highly variable among studies and across arms, with values ranging from 77.0 to 150.01 pg /mg creatinine. The highest levels were observed in the U.S. population and lowest levels in the Japanese populations. The differences are likely related to differences in CC in Japan.

Percent decreases of urinary total NNAL levels from baseline to Day 5 in all four studies ranged from 48% to 61% for the THS 2.2 arms and 59% to 64% for the SA arms. On Day 5, total urinary NNAL concentrations were similar among all REX studies, with levels ranging from 33 to 57 pg/mg creatinine in the THS 2.2 arms and from 28 to 54 pg/mg creatinine in the SA arms. During the ambulatory periods, the decreases in total urinary NNAL concentrations continued: in ZRHM-REXA-07-JP, the maximum reduction was reached on Day 90; in ZRHM-REXA-08-US, the urinary NNAL concentrations were lowest on Day 60. On Day 90, percentage decreases from baseline were 69% to 67% (median value) in the THS 2.2 arms. On Day 90, the total urinary NNAL concentrations were higher for the U.S. population, with 23 and 47 pg/mg creatinine in the THS 2.2

54

arms and 13 and 48 pg/mg creatinine in the SA arms for ZRHM-REXA-07-JP and ZRHM-REXA-08-US, respectively.

While baseline levels of total urinary NNAL varied among studies, a steady decline in total urinary NNAL concentrations was observed in all studies. In smokers who switched to THS 2.2, systemic exposure to total NNAL was statistically significantly lower than those who continued CC smoking.

f. CYP1A2 Activity At baseline, mean CYP1A2 activity ranged from 70% to 121% across all four studies. By Day 5 across all four studies, CYP1A2 activity decreased in study participants randomized to switch to THS 2.2 (55% to 91%) and SA arms (52% to 94%). In contrast, CYP1A2 activity at Day 5 stayed similar to baseline levels in the CC arms (76% to 123%). In the two ambulatory studies, the changes from baseline ranged at Day 90 from decreases of 20.2% to 32% in THS 2.2 users and from 15.8% to 35.4% in the SA arms, respectively.

g. Urine Mutagenicity At baseline, median urinary levels of mutagenicity ranged from 12574 to 25823 rev/24h across all four studies. Overall, the relative change in urine mutagenicity levels from baseline during the 5-day confinement periods ranged from decreases of 42.8% to 72% for the THS 2.2 arms, decreases of 37% to 74% for the SA arms, and from a decrease or 24.4% to an increase of 40% for the CC arms. In the two ambulatory studies, the decrease from baseline ranged at Day 90 from 61.4% to 61.6% in THS 2.2 users and from 45.2% to 67.9% in the SA arms, respectively. In contrast, at Day 90, the relative change from baseline in the CC arms was an increase of 10.9% to 16.2%.

h. Summary The BCP review concludes that the reductions in systemic exposures to 15 BOE seen after switching to from CC smoking to THS 2.2 may lead to reduced likelihood of smoking-related diseases.

As TPL, I agree with the BCP conclusion that reduced BOE may lead to reduced risk of tobacco-related disease. The BOE chosen by the applicant are well established in peer-reviewed literature as measures of exposure to HPHCs. Biomarkers of some other particularly concerning chemicals found in CC smoke were not assessed in the clinical studies; e.g., acetaldehyde, formaldehyde, isoprene, ammonia, arsenic, cadmium, and lead. (There are not suitable biomarkers for acetaldehyde, formaldehyde, isoprene and ammonia; the applicant demonstrated low levels of arsenic, cadmium and lead in nonclinical studies.).

Although the applicant’s data show reductions in BOE during short-term exposures, these measures were not intended to evaluate long-term disease risk. In the reduced exposure studies, all but one of the measured BOE were consistently and substantially lower in the groups who switched completely from CC to THS 2.2. In the case of 3-HPMA, the applicant’s explanation for slightly higher levels in the THS users compared to those in the SA arm is reasonable and the level of 3-HPMA is decreased in the THS users compared to continued CC use.

There are some limitations to these trials: • The small sample sizes limit extrapolation of results to the entire U.S. population. Study ZRHM- REXA-08-US, the only REX study conducted in the U.S., enrolled 164 subjects. The size of the PK/PD studies limits analysis of sub-groups (e.g., youth, low socio-economic status, minorities).

55

• The studies were not designed as nationally representative of the U.S. smoking population. Participants were moderate smokers; therefore, data may not generalize to light or non-daily smokers. IQOS products, but not own-brand cigarettes, were provided free of charge for participants in the REX studies, which may affect product use rates. • The applicant did not estimate the percent change in BOE in the subset of participants who did not completely switch to IQOS and continued to use IQOS and CC (dual use). Dual use was particularly common in the Actual Use study and may account for a substantial proportion of IQOS users in a real-world setting. Whether this user population will achieve an exposure reduction when compared to exclusive CC use, and to what magnitude, is unclear. • The applicant compared CC to THS use to SA. Participants in the SA arms differ from never-users; they may have residual or continued exposure to HPHCs or other chemicals with longer half-lives. A comparison to never-users would have been helpful to determine to what extent THS users (i.e., switchers) are still exposed to HPHCs compared with never users. • No biomarker studies of secondhand exposure to these products were conducted by the applicant. This type of study could have helped to better understand potential risks to non-users. There were also no comparisons between IQOS and other tobacco products (e.g., e-cigarettes). Given that IQOS and e-cigarettes may both be considered by consumers to be a substitute for cigarettes, a comparison of the differences in exposure would be useful. However, the popularity of e-cigarettes in the U.S. has increased significantly in recent years; this change largely occurred during or after the time during which the applicant conducted the clinical trials submitted in these PMTAs.

Overall, the BOE reductions were statistically significant over five days and the decreases persisted up to three months. For those that switch completely from CC to THS 2.2, these reduced BOE exposures, which indicate reduced HPHC exposures, are likely to result in reduced risk of tobacco-related disease although that reduced risk has not been demonstrated in the studies submitted by the applicant.

2. Biomarkers of Potential Harm (BOPH) All REX studies included measurements of several BOPH as secondary or exploratory study endpoints to determine if THS 2.2 use resulted in biological changes that may indicate a change in long-term disease risk. Selection of BOPH was based on key mechanisms of three major smoking-associated diseases: cardiovascular disease (CVD), chronic obstructive pulmonary disease (COPD), and lung cancer. Markers of oxidative stress (8-iso-F2-isoprostane-alpha [8-epi-PGF2α] and thromboxane metabolites) and inflammation (WBC, c-reactive protein [CRP], soluble intercellular adhesion molecular 1 [sICAM-1], fibrinogen) were selected due to their role in the development and progression of three major smoking-related diseases. Additionally, some disease-specific endpoints were selected for CVD (carboxyhemoglobin, lipid profile and oxysterols, HbA1c, and blood pressure), lung function (FEV1), and cancer (selected BOE). The applicant selected these biomarkers based on changes shown in previous smoking cessation studies and the general acceptance in peer-review literature of association with health risks. The applicant specified six BOPH as representative of mechanisms underlying diseases of interest. Two markers were measured in all four REX studies: 8 epi-PGF2α and 11-dehydrothromboxane B2 (11-DTX-B2). Four additional markers were measured in the ambulatory studies: high-density lipoprotein (HDL) cholesterol; WBC; soluble intercellular adhesion molecule-1 (sICAM-1); and forced expiratory volume in the first second (FEV1).

The applicant provided literature review monographs for most of the clinical measures undertaken in the REX studies. FDA conducted an independent review of the relationship of these BOPH to diseases of interest; details are provided in the medical review. The applicant monographs concluded that only WBC count and HDL were useful clinical risk markers for the evaluation of health risks of THS 2.2. Neither of these

56

measures changed significantly in the U.S. study. Evidence regarding the remaining markers is insufficient to allow reliance upon them as surrogate predictors of either short- or long-term health effects from switching to THS 2.2.

a. Assessment of Inflammation Only WBC count and sICAM-1 demonstrated some differences in the two 90-day studies for the THS 2.2, CC, and SA arms. There was a reduction in WBC counts over the course of the studies. Reductions were generally largest in the SA arm, but there were consistent reductions approaching similar levels in THS 2.2 arms. The U.S. study results are difficult to interpret because of the small sample size. The study length is a limiting factor for interpretations because WBC reductions are optimally detected between six and ten months after smoking cessation.

In the two ambulatory REX studies, smokers in the THS 2.2 arm had lower sICAM-1 levels than participants in the CC arm after adjusting for baseline sICAM-1 levels, sex, and baseline CC consumption. The sICAM-1 levels in the THS 2.2 arm (approximately 8.5% to 10.5% reduced from baseline values) were similar to the SA arm.

b. Assessment of Oxidative Stress The markers for oxidative stress included 8-epi-PGF2α and 11-DTX-B2. Smokers who switched to THS 2.2 showed more than a 12% reduction in 8-epi-PGF2α levels compared with smokers who continued to smoke CC; however, these reductions were not conclusive due to high variability in the data. Although 11-DTX-B2 levels were reduced for smokers who switched to THS 2.2, as in the SA arm, the magnitude of the change was smaller than expected, especially in the U.S. study (ZRHM-REXA-08-US) – possibly due to poor compliance. The U.S. study findings were confounded by non-compliance with product use (in both THS 2.2 and SA arms), resulting in a reduced sample size and a greater-than-expected variability in 11-DTX-B2 results.

c. Assessment of Cardiovascular (CV) Risks CV risks were assessed by measurements of blood lipids, triglycerides, apolipoprotein B, and blood pressure in the REX studies with the ambulatory period. Except for HDL, CV risk biomarkers did not change significantly over the course of the ambulatory periods. Smokers who switched to THS 2.2 had higher HDL levels compared to smokers who continued to smoke CC. In the U.S. study (ZRHM-REXA-08-US), HDL levels in the CC arm were similar to those in the SA arm; however, only nine of the 40 participants randomized to the SA arm reported adherence to SA, making results of the U.S. study difficult to interpret. The BOPH related to CV risk did not significantly change, suggesting no improvement in CV risk during the relatively short study period.

d. Assessment of Lung Function Spirometry measurements were included primarily as a safety measure. In the Japanese study (ZRHM-REXA- 07-JP), smokers who switched to THS 2.2 had an increase of 1.91% of predicted value in FEV1 compared to smokers who continued to smoke CC, with no notable differences between THS 2.2 and SA arms. However, studies of a longer duration (at least 6-12 months) would be necessary to fully assess the impact of THS 2.2 use on FEV1. Additionally, deterioration in lung function associated with CC may not be reversible.

57

e. Assessment of Genotoxicity and Mutagenicity Several indirect measures evaluated cancer-specific endpoints. There may be reduction in cancer development risk related to reduced exposure to carcinogens; for example, HPHCs, reduction in CYP450 1A2 activity, and urine mutagenicity are additional indirect measures.

CYP450 1A2 activity at baseline was markedly different among the REX studies, ranging from 70% to 122%, with the lowest values in the Japanese populations and highest in the U.S. population (due to differences in genetic factors and smoking behavior). In participants who switched to THS 2.2, CYP450 1A2 activity was significantly lower (decreased by 30% to 36%) compared to participants who continued to smoke CC on Day 5; these reductions were similar or lower than levels during the ambulatory period on Day 90 (range of 21% to 32%), and the changes in the SA arms were of similar magnitude.

In the REX studies, the baseline urine mutagenicity levels were highly variable; however, there was a clear trend toward decreased urine mutagenicity values in the THS 2.2 arm on Day 5 (by 47% to 72%); these differences were sustained in the ambulatory period and were similar to changes in the SA arm. Data in the CC arms did not follow a consistent profile through exposure periods across studies. There was an unexpected decrease in mutagenic activity in the ZRHR-REXC-04-JP, similar to what was observed in the THS 2.2 and SA arms. The high variability of results may be due to test sensitivity of dietary mutagens and the complexity of this cellular test; data should be interpreted with caution. However, the results observed indicate a lower level overall of mutagenic compounds in the urine of THS 2.2 users compared to the CC arm.

f. Summary The applicant notes that many of these endpoints are more appropriate for longer-term studies as changes in these measures are expected to take months to years. Some BOPH had a desirable trend in improvements for THS 2.2 users compared to the CC arm during the 90-day exposure, specifically: HDL: lipid pathway; 8- epi-PGF2α: oxidative stress; 11-DTX-B2: platelet activation; sICAM-1: endothelial dysfunction; WBC (leukocytes) count: inflammation; FEV1: lung function parameter.

The medical, epidemiology, and BCP reviews concluded that: • The minor improvements in some BOPHs in the THS 2.2 arm relative to the CC arm may not be of clinical significance and it is unclear how predictive the chosen BOPH are for long-term tobacco- related disease risk. • While no deaths, CV disease, COPD, or lung cancer were reported during the clinical studies, these diseases have a long latency and are unlikely to be observed during studies of this type.

FDA medical reviewers conducted an independent review of the literature and concluded that while each of the six markers has data suggesting a relationship with one or more of the three identified diseases of interest, none were strong predictors of future health risks and many will take months to years before change can be measured. WBC had the most data to suggest utility; however, it is a non-specific marker that can change for numerous reasons independent of tobacco product use.

The epidemiology review notes that while the BOPH can be informative with respect to key mechanisms of smoking-related diseases, they are not necessarily replacements of clinical endpoints. In general, there are continued questions about the credibility of BOPH as surrogate endpoints.

58

The statistical reviewers evaluated the two 90-day studies and concluded that they were not designed to ascertain any effect associated with the “risk endpoints.” The BOPH were secondary endpoints and were not the basis for sample size/power calculations; it is not clear from a statistical perspective whether the data generated from the studies are clinically meaningful.

As TPL, I agree with the BCP, medical, epidemiology, and statistical reviews. Compared with the significant reductions in BOE, the changes in BOPH were less pronounced. One explanation is that none of the BOPH are specific to tobacco use. Changes in BOPH may be attributed to other factors (e.g., weight, diet, exercise). Also, biologic responses related to exposure to tobacco smoke and reversal of these harmful effects may take more time to manifest than the duration of the ambulatory periods of the current studies; many of the effects, e.g., effects of CC, may not be reversible. These factors limit the interpretation of results related to the effects of long-term exposures.

Overall, the studies conducted by the applicant have not demonstrated evidence of reduction in long-term disease risks. BOPH may be informative, however, for understanding potential effects on biological processes such as inflammation and oxidative stress. Long-term tobacco related diseases, e.g., cardiovascular disease, cancer, chronic lung disease, begin as inflammatory processes. Reduction of inflammation and oxidative stress may eventually lead to reduced disease risks. Use of THS appeared to reduce these processes to some degree during the studies, but, as noted, the data are not sufficient to show that these small changes are associated with long-term results.

3. Clinical Effects of IQOS The application included several types of health-related data and supporting information that aid in the evaluation of short-term health risks of IQOS. Safety data reports for THS 2.2, including cumulative safety summary information from the eight completed clinical studies, two ongoing clinical studies, premarket safety surveillance covering six market research studies and one perception and behavior study, as well as post-market surveillance studies outside the US, were submitted.

a. Analysis of Adverse Events in Clinical Studies In the eight clinical trials (four PK/PD and four REX studies), adverse events (AEs) associated with acute exposures to THS 2.2 were like those ordinarily encountered with CC use. A total of 717 AEs was reported. Most (>95%) were non-serious,68 mild to moderate in severity, expected,69 and temporary. They encompassed acute, short-term health effects including cardio-pulmonary, nasopharyngeal, neurologic, and laboratory abnormalities. The number of reported severe70 AEs was relatively low across the eight clinical studies (N=19). Of these 19 severe AEs, 16 occurred in the 90-day U.S.-based clinical trial. No deaths were reported for any subjects participating in the clinical trials. Although the applicant determined that most of the reported AEs were unrelated to product use, THS 2.2 exposure cannot be ruled out as contributing to or exacerbating those AEs typically associated with tobacco exposure (e.g., cough, headache, syncope).

Two serious AEs were reported in ZRHM-REXA-08-US; both occurred in Subject 1119, who was enrolled and exposed but not randomized. Both AEs were not related to IQOS or to the investigation and resolved with

68 Serious adverse event (SAE) = Event that results in death, life-threatening condition or event, persistent or substantial disability or incapacitation, hospitalization or prolonging of current hospitalization, or congenital anomaly or birth defect. 69 Unexpected adverse event = The nature or severity of the event is not consistent with information currently known about the product and/or has not been previously observed or described in the investigator’s brochure. 70 SAE = Event that interferes with most daily activities. NOTE: these definitions were used by the applicant.

59

treatment. Twelve severe AEs were reported during the ambulatory period; all were due to abnormal clinical laboratory results and were unrelated to product use.

Notably, 56 subjects in study ZRHM-REXA-07-JP were discontinued by the applicant due to non-compliance with International Conference on Harmonization Good Clinical Practices (ICH GCP) at the Seishukai Tokyo Clinic. The 56 subjects were exposed to IQOS. Their subject-level data were excluded in the data analysis in the full analysis set/per protocol and the safety population, so no further safety analysis was possible; however, identification and removal of 56 subjects’ data prevented a potential compromise of data validity and integrity.

In the Actual Use study (THS-PBA-07-US), a single prospective multi-center study that exposed 1,158 daily smokers of CC to THS 2.2 in a naturalistic, close to real-world setting, 121 AEs were reported by 48 participants. Most were expected; 102 AEs were non-serious and 19 were serious. Headache was the most frequently reported non-serious AE. Severity was not reported in 50% of the cases.

b. Review of Published Clinical Literature Post-marketing AE reports about IQOS have been sparse, despite increasingly widespread international marketing since its commercial introduction in Japan and Italy in 2014. A Safety Update Report (PMI-SURV- 2016_SUR01), published in April 2016, reported two serious AEs (nervous system disorders/syncope). The serious AEs involved “THS 2.2 and 2.4/All variants” and were reported from an unspecified “spontaneous source.” A review of published clinical literature provided by PMI at the time of the application found one case report of a serious AE from THS 2.2. A young adult Japanese male developed acute eosinophilic pneumonia after increasing his consumption of Heatsticks. This disease has a known association with tobacco products and is not unique to THS 2.2.71

FDA conducted an independent clinical literature review to ascertain whether any serious adverse experience had been reported from products referred to in the literature as tobacco heating systems. A search was completed on 1/31/17 and repeated on 12/5/17; details are in the medical review. Only the case described above pertained to a new generation tobacco heating system. Although the device in this case report was not explicitly named, the description was consistent with IQOS. The remaining articles described small studies of earlier generations of heated tobacco products and showed “minor but favorable” changes in some cardiopulmonary and inflammatory parameters for those using the “heat-not-burn” product compared to a conventional cigarette. Two studies of the product, which uses a carbon-burning heat source, showed increases in CO measurements.

c. Summary The medical review noted additional limitations related to information about health effects of IQOS: • The eight clinical studies did not specifically evaluate the possible risks or benefits of dual/poly tobacco product use. • The reported AEs and compliance rates in a controlled clinical setting and small sample population may not be reflective of general use. Detection of other more clinically significant, serious, or severe AEs may occur with use of the product by a diverse population, especially for chronic and exclusive users.

71 Kamada, T., Yamashita, Y., & Tomioka, H. (2016). Acute eosinophilic pneumonia following heat‐not‐burn cigarette smoking. Respirology Case Reports, 4(6).

60

As TPL, I agree with the medical review. There are limited data related to short-term health effects of IQOS and even less data for longer-term effects. IQOS has been available in other countries for several years; no health-related short-term issues uniquely related to IQOS were identified in three searches of published clinical literature. There are limitations to the clinical studies conducted by the applicant; however, there are practical limitations to the number, size, and nature/design of clinical studies that can realistically be completed during new product development. Although limited, the data available in the clinical studies completed by the applicant do not raise concerns or identify specific health-related issues uniquely related to IQOS.

4. Likelihood of Product Misuse or Malfunction The Actual Use study assessed self-reported misuse of the THS 2.2. Of 985 participants, 47 (4.8%) reported using the Heatstick without the IQOS device; the majority (97.9%) lit the Heatstick like a CC, and one participant chewed the Heatstick on one occasion. Only two participants (0.2%) reported using the IQOS device without Heatsticks; one participant used the IQOS device with marijuana on one occasion and one participant used it with CC on more than ten occasions.

The applicant reports device events or malfunctions in several of the PK/PD and REX studies. Device events, malfunctions, and misuse events were relatively minor or easily correctable (e.g. device inoperable/does not charge, battery malfunction, heater broken) and did not impact subject safety. There were no battery explosions or subject burns resulting from device malfunctions. One major device event (device inoperability) was reported in the pre-randomization period and did not impact subject safety. There were no subject discontinuations resulting from a device event.

Table 4: Summary of Device Issues in IQOS Clinical Trials Trial Number Device Event/Malfunction/Misuse Consequence ZRHR-PK-01-EU None None ZRHR-PK-02-JP 1 broken heater, 1 charging issue None 12 subjects reported 19 device problems; ZRHR-REXC-03-EU No AEs Charging issues, inoperable, stops intermittently ZRHR-REXC-04-JP 4 subjects reported 5 events No AEs ZRHM-PK-05-JP None None ZRHM-PK-06-US 3 subjects reported 6 minor events No AEs 189 device events; inoperable, won’t charge, battery ZRHM-REXA-07-JP No AEs malfunction, heater broken 149 events reported by 55 subjects; most minor (inoperable, ZRHM-REXA-08-US No AEs won’t charge, battery malfunction, heater broken)

The applicant evaluated the potential for consumers to attempt to re-use Heatsticks. (b) (4)

The applicant also evaluated (b) (4)

The applicant did not provide data on consumer misuse of the Holder by attempting to use a conventional combusted tobacco product (e.g., , cigarette). However, heating tobacco will only generate an aerosol if

61

there is enough of an “aerosol forming agent,” such as glycerin.72 Although glycerin is used in conventional tobacco filler as a humectant, the level is below that required to generate a nicotine-containing aerosol. In addition, the tobacco of any conventional product inserted into the IQOS Holder would be heated only to a maximum of 350 °C - the maximum temperature of the heating blade. This temperature is much lower than the combustion threshold of tobacco (>400 °C).73, 74 Therefore, the levels of any HPHCs generated by using a conventional combusted tobacco product with the IQOS Holder would be lower than the HPHCs generated with usual use of the conventional combusted tobacco product. Furthermore, only products with a circumference of 22.9 mm or less would fit inside an IQOS Holder, which excludes most conventional U.S. cigarettes.75

Overall, although self-report data has limitations, the Actual Use study suggests that consumer misuse of the IQOS device and Heatsticks is uncommon. Studies of this type are generally accepted by FDA for evaluating how consumers will actually use a product. The applicant has considered how consumers could re-use a Heatstick or inappropriately use other conventional tobacco products with the IQOS Holder. (b) (4)

5. Bioresearch Monitoring (BIMO) Inspection Bioresearch monitoring (BIMO) inspections of two clinical investigators were conducted in support of the applications. Protocol ZRHM-REXA-08-US had the largest number of reported AEs (N=301). In addition, overall in the safety population at post-randomization, eight AEs were classified as related to THS 2.2 exposure and 12 severe AEs occurred. Protocol ZRHM-REXA-08-US was conducted at two clinical sites in the U.S. – Daytona Beach, FL and Dallas, TX. No significant issues were identified during inspection of the Daytona Beach site; final classification was No Action Indicated (NAI). There were issues related to documentation of study records identified at the Dallas site. These were discussed with the study sponsor, monitor, and IRB. The appropriate samples were removed from final sample data analysis. Final classification was Voluntary Action Indicated (VAI).

Overall, BIMO inspection findings indicate the conduct of the U.S.-based study at the two clinical sites generally complied with study-related procedures, documented and monitored AEs, and followed procedures to ensure informed consent and human subject protection. The inspection revealed no major BIMO issues or clinically-significant protocol deviations that would compromise data validity and integrity.

6. 2017 Safety Update Report In a letter dated 5/16/2018, FDA received a PMI Safety Update Report (SUR). This report summarized safety information on THS for the period covering 1/1/2017 thru 12/31/2017. The SUR identified previously unrecognized short-term health risks associated with THS including hypersensitivity reactions, an accidental

72 Nordlund, Markus; Belka, Miloslav; Kuczaj, Arkadiusz K; et al., Multicomponent aerosol particle deposition in a realistic cast of the human upper respiratory tract. Inhalation toxicology. 2017, Vol.29(3), p.113-125. 73 Barontini, Tugnoli, Tetteh, et al., Volatile products formed in the thermal decomposition of a tobacco substrate. Industrial and Research Chemistry. 2013, Vol 52, p 14984-14997. 74 Senneca, Chirone, Salatino, and Nappi. Patterns and kinetics of pyrolysis of tobacco under inert and oxidative conditions. Journal of Anal. Appl. Pyrolysis. 2007, Vol 79, p. 227-233. 75 Agnew-Heard KA, Lancaster VA, Bravo R, Watson C, Walters MJ, Holman MR., Multivariate Statistical Analysis of Cigarette Design Feature Influence on ISO TNCO Yields. Chem Res Toxicol. 2016 Jun 20;29(6):1051-63.

62

child exposure, and a reported weather-related (heat and humidity) “burning sensation.” No AEs were reported as actual consumer burns necessitating clinical treatment; however, consumers did report Heatstick discomfort with the “perception of hot aerosol causing burning sensation and thermal burns.”

The data showed that hot aerosol AEs were reported more frequently during the summer months. The results of the initial assessment indicate that (b) (4)

. The minimum time to sense pain and react to it at any temperature is 0.3 seconds in average adults. The thermal threshold for pain in the oral cavity varies between individuals but is normally around 46 - 47°C.76, 77 To sustain a burn in humans, the skin needs to reach and remain at a temperature of 50°C for over 100 seconds – an unlikely occurrence. The applicant has made no product changes but has provided Customer Care agents with a consumer communication script, reminding product users not to expose HeatSticks to high temperature and humidity but to keep them in a dry environment, especially during summer months. This consumer communication was initiated in November 2017 and the applicant plans to use it in the U.S. following IQOS commercialization. Further information and instructions on how to handle the product, aiming at minimizing and preventing the risk of hot aerosol sensation, are provided in the IQOS User Guide supplied with each THS pack. Users are advised not to use the product if “it has been exposed to excessive heat or moisture” or if “it becomes wet or is immersed in any liquid” and to “store the product in a clean, dry, cool place.”

The safety update also reported battery leakage due to short circuiting and THS-potentiated risks of thermal burns. In response to FDA’s request for additional information, the applicant clarified that the device malfunction was the result of a short circuit (b) (4)

.

Based on the root cause analysis performed, PMI concluded that the battery electrolyte leakage was a result (b) (4)

The applicant reports that (b) (4) (b) (4) . These improvements/modifications are expected to decrease the occurrence rate of AEs.

The applicant states that this “improved” THD 2.4 is the product that is the subject of these applications and intended for U.S. marketing (if authorized). This product reflects modifications have been made as part of continuous product improvement processes. These modifications include: (b) (4)

76 Green B.G. Heat pain thresholds in the oral-facial region. Perception and Psychophysics. 1985, Vol 38(2), p. 110-114. 77 Dirler, Winkler, Lachenmeier. What temperature of coffee exceeds the pain threshold? Foods. 2018, 7, 83.

63

The applicant confirmed that all device modifications were implemented prior to the September 8, 2017 amendment submitted to FDA. A CB Test Certificate was provided, dated 12/28/2017, for THD 2.4, stating that it is compliant with IEC 60335-1:2010 (safety of electrical appliances for household and similar purposes) and Test Report IEC 60335, dated December 1, 2017 for THD 2.4. The applicant reports no changes have been made to the design of HeatSticks since submission of the PMTAs. The engineering review provides details of the risk analysis and the subsequent modifications.

The applicant reports that new post-market safety surveillance data and published literature reports of acne, chest discomfort and rash resulted in heightened awareness and monitoring for these potential safety signals. Serious adverse events (SAEs) were reported in the SUR Supplement 1. Definitive conclusions could not be determined on the 14 SAEs reported as cardiac disorders (e.g. angina pectoris, arrhythmia, myocardial infarction); the summaries provided were incomplete and anecdotal and many cases lacked a verifiable consumer medical history. THS use habits were inconsistent, not reported or unknown; SAE seriousness, severity, and outcomes were inconsistently reported.

The SUR included report of one death in an 88-year-old Caucasian male with a medical history of hypertension, gastroesophageal reflux disease, and cigarette smoking (1 pack/day) since age 13. On 12/28/2015, the subject was enrolled in the ZRHR-ERS-09-EXT-US clinical study and on 1/4/2016 he received THS 2.2 for one week prior to randomization to the CC arm. After randomization, the subject continued smoking 1 pack/day. The subject was lost to follow-up for a month before the investigator learned of his death from an obituary. The MedWatch Report stated the subject died of unknown causes on 8/14/2016; no autopsy was performed. The investigator reported that the SAE was unrelated to THS, conventional cigarette use, or any study procedures.

The new AE safety information, including the unexpected death of the study participant, does not change the conclusion that short-term risk of THS use is no greater than the risk of CC smoking.

7. Summary of Individual Health Findings from Clinical Studies, Literature, Adverse Experience Reports, and Safety Updates The medical review concludes that THS 2.2 has the potential to benefit certain individuals seeking to reduce their HPHC exposure by completely switching from CC. The review concludes that short-term health effects data from the clinical studies and additional longer-term information from published literature provided in the applications do not raise unique or additional health concerns or identify unique, specific health-related risks for the IQOS system. The following rationale for this conclusion is provided: • Reducing exposures to HPHCs in THS 2.2 through complete switching can potentially reduce the risk of adverse health effects compared to CC • Data about BOPH are insufficient to draw meaningful conclusions about the ability of THS 2.2 to impact disease risk • Clinical trial data about AEs related to THS 2.2 are limited but suggest that the short-term risk is no greater than risk from CC • The relatively low incidence of serious and severe AEs in the international post-marketing surveillance SUR and the published literature suggest that switching to THS 2.2 may not increase the incidence of short-term adverse health effects for U.S. users relative to CC. However, the short-term AE data do not demonstrate a reduction in long-term health risk relative to CC.

As TPL, I agree with the medical review overall conclusions. The 5-day studies demonstrate improved BOE in those that completely switched to THS 2.2, which indicates reduced HPHC exposures. These improvement

64

trends persisted in the 90-day studies despite reduced compliance and use of other tobacco products. The currently available evidence indicates CC smokers who switch completely to IQOS will have reduced toxic exposures and, consequently, although not demonstrated in the studies in the application, are less likely to be at risk of tobacco-related diseases.

Additional health effects information was included in a late amendment to the submissions (see section II.F.6 below). MR0000117 included results of a randomized, controlled, open-label, 2-arm, parallel group, multi-center clinical study of six months of ad libitum use of the non-menthol THS 2.2 compared to continued CC users in an ambulatory setting. In this study, CC smokers who use IQOS while continuing to smoke (dual use) do not appear to have increased HPHC exposures; the limited available information shows trends, although not statistically significant, toward reduced HPHC exposures in this population.

Experience with IQOS is limited, even when considering data from other countries. There have been no specific, health-related or product quality issues unique to IQOS products found in the clinical studies, the current world-wide markets, or the published literature. Misuse of IQOS appears to be uncommon and the product design makes it unlikely users will have a satisfactory experience (e.g., no significant nicotine is delivered with reusing a Heatstick and use of CC in the Holder is not effective).

F. Population Health The applicant includes the following studies for evaluation of IQOS and population health effects: • PBA-05-NOC: a U.S. study assessing perceptions and intentions regarding IQOS use in current smokers with and without intent to quit, former smokers, never smokers, and young adult never smokers • Four observational studies (Table 5 summarizes design features) o PBA-07: Actual Use pre-market study conducted in the U.S. o Whole Offer Test (WOT): Post-market study conducted in five countries in Asia and Europe o Two post-marketing surveys conducted in Japan • Population Health Impact Model (PHIM)

65

Although the PBA-07 and WOT studies are not generalizable to U.S. cigarette smokers, the information gained from these studies provides useful trends for consideration in review of these applications for marketing in the U.S. The sample for PBA-07 was a non-probability sample recruited from marketing research databases. Furthermore, the dataset used in the analyses only included participants who used at least one cigarette during the baseline period and at least one Heatstick during the observational period, thereby potentially overestimating the prevalence of IQOS use after excluding 230 participants who did not meet these criteria. Similarly, the WOT analysis only included participants who completed at least 26 of the 28 days during the observational period. If participants who did not complete the observational period were less likely to use Heatsticks, the estimates for initiation of IQOS and switching to IQOS would be overestimated, while the estimates for switching from IQOS back to CC (e.g., 0% in Japan) would be underestimated. The results from the international actual use studies pose additional challenges with respect to generalizability to cigarette smokers in the U.S. population due to different cultural contexts and differences in the availability of e-cigarettes or other heated tobacco products.

The FDA statistical reviewers evaluated four studies submitted by the applicant in support of the PMTAs and their benefit to the health of the population as a whole: THS-PBA-05-NOC-US, THS-PBA-06-US, THS-PBA-07- US, and WOT. In general, the statistical reviewers found the information for these studies to be descriptive in nature: computation of summary statistic (proportions and means) and standard deviation. Statistical inference was not part of the conclusion-making process in relation to these studies.

1. Likelihood of IQOS Use by Current Cigarette Smokers PBA-05-NOC This study was designed to assess the likelihood of use, comprehension, and risk perception among current smokers with and without intent to quit, former smokers, never smokers, and young adult never smokers. The objective of this study was to assess perception associated with exposure to THS 2.2 label, labelling, and marketing materials. A total of 1,817 participants were recruited from multiple consumer databases across the U.S. Enrolled participants were randomized into one of four arms - 384 individuals per arm (96 smokers intending to quit, 96 smokers not intending to quit, 96 former smokers, 96 never smokers). Each arm was presented with a separate label, labeling or advertising material: x Arm 1: THS 2.2 brochure with product information and a Surgeon General’s Warning x Arm 2: THS 2.2 HeatSticks pack with a Surgeon General’s Warning and a statement that THS 2.2 heats but does not burn tobacco x Arm 3: THS 2.2 HeatSticks pack with a Surgeon General’s Warning but without the statement that THS 2.2 heats but does not burn tobacco x Arm 4: THS 2.2 direct mail with product information and a Surgeon General’s Warning All four Surgeon General’s warnings were used but each participant saw only one of the statements on the information received.

The applicant developed two items assessing intentions to try IQOS and one item assessing intentions to use IQOS regularly, if one tries it and likes it. Table 6 shows results from PBA-05-NOC across the study arms; in addition to the top two categories (Definitely or Very likely), FDA reviewers included those who responded Somewhat likely. Considerable proportions of current smokers also reported that, if they tried IQOS and liked it, they would Definitely or Very Likely use it regularly, on an ongoing basis. As the applicant acknowledges, self-reported intentions to use products are limited in terms of predicting behavior and can overestimate the likelihood of purchase, particularly when participants’ responses have no consequences. For example, although participants viewed price information about IQOS and Heatsticks, they were not asked to make a choice between the product and money.

68

Figure 9: Likelihood of Heatstick Purchase among Current Smokers who used at least 100 Heatsticks in PBA-07 Actual Use Study Note. FDA Reviewer drafted this figure using data from the application. Data Source: ths-pba-07-us- study-report-add-HSType_fdafixed.pdf.

The PBA-05-NOC study distinguished between smokers with and without an intention to quit smoking to evaluate whether marketing IQOS would have negative effects on smokers who intend to quit, such as causing them to delay their quit attempts. The applicant considered a change in quit intention from any intention to quit smoking in the next 30 days or 6 months pre-exposure to no longer intending to quit after exposure to the LLA materials; the FDA social science reviewer also included those changing from intending to quit in the next 30 days to an intention to quit in the next 6 months, as well as a change from any quit intention to no intention to quit. When considering the social science reviewer’s broader definition of change in quit intentions, 4-19% of current smokers delayed a quit intention or no longer held an intention to quit cigarettes and 4-10% of current smokers with an intention to quit all tobacco had a lower or no intention to quit all tobacco after exposure to LLA materials.

The applicant explained in the September 2017 amendment that the decision was made to prioritize specificity over sensitivity in order to minimize false positive predictions of actual (post-market) trial/use. As there is no consensus in the literature on how intention should be measured, it is difficult to know what kind of assessment most accurately predicts actions that will be taken by the respondents in the future. Additionally, most cigarette smokers express intention to quit at some point; however, the success rate for any given quit attempt is low. The uncertainty of smokers’ intentions in this study is consistent with our current knowledge. Thus, the broader approach used by the social science reviewer may not predict actual behaviors.

Overall, pre- and post-market observational studies found that IQOS use patterns varied across the U.S., Asia, and Europe. In the U.S., 34% of cigarette smokers in the Actual Use study initiated IQOS use, defined as using at least 100 Heatsticks. In the WOT, conducted in multiple countries where IQOS is currently marketed, the prevalence of initiating IQOS use ranged from 36% in Italy to 76% in South Korea. The findings suggest that some smokers will find IQOS appealing and acceptable enough to initiate use of the product. In the U.S. Actual Use study, daily cigarette consumption decreased between baseline and the observational period for all IQOS use groups, with the largest decrease occurring in participants who were predominant

71

Despite the high proportion of incomplete switching or dual use in the study, the applicant notes that there was a reduction in average daily cigarette consumption across all IQOS use groups in the PBA-07 study, even among the group of participants who were predominantly using cigarettes at Week 6. However, when FDA considered baseline cigarette use compared to each of three follow-up visits (as opposed to compared to the 6-week average daily use highlighted by the applicant), participants appear to reduce by about 1 CPD over the entire study period and add about 2-4 Heatsticks per day. Accordingly, even though the average daily total appears relatively stable, participants may be using more total units of tobacco products when both cigarettes and Heatsticks are considered. Heatsticks were provided free of charge; the pattern of use during this study may not accurately reflect the use pattern of marketed product.

In PBA-07, the use of NRT products remained stable from baseline to Week 6 (2.4%), while the use of e- cigarettes increased from baseline (14.2%) to Week 6 (20.6%) and the use of other tobacco products such as , and smokeless tobacco products decreased from baseline (38.7%) to week 6 (26.9%). Those reporting no other tobacco use was 54.5% at baseline and 60.7% at week 6.

The applicant provided data from two Japanese on-line post-marketing surveys. In a 2016 Japanese online cross-sectional survey of 2000 adult smokers and nonsmokers, 3.7% of respondents reported using “heat- not-burn” (heated) tobacco products. The prevalence of heated tobacco product use was higher among those aged 20-39 (~4 %) than those aged > 40 (~1 - 1.5%) and most (96.3%) were using “Marlboro Heatsticks with IQOS device.” Among respondents currently using heated tobacco products, 84.9% also smoked cigarettes, most of them daily. In total, 91.8% of heated tobacco product users reported dual use with at least one other tobacco product. For most respondents, heated tobacco products comprised less than 30% of their average total daily tobacco consumption. A total of 15.5% of heated tobacco product users were ĐŽŶƐŝĚĞƌĞĚ ĞdžĐůƵƐŝǀĞ ƵƐĞƌƐ ;шϵϱй use). All respondents in the exclusive use group were not current cigarette smokers. Although the applicant describes data from Japan, use of nicotine-containing e-cigarettes requires a prescription in Japan, which may limit generalizability of the data to the U.S. population.

In the second Japanese marketing survey, data on self-reported use of IQOS and cigarettes were also collected from 14,999 adult IQOS purchasers who registered their device in an online market research database. The proportion of IQOS purchasers who ǁĞƌĞ ͞ĞdžĐůƵƐŝǀĞůLJ͟ ƵƐŝŶŐ /YK^ ;шϵϱйͿ ŝŶĐƌĞĂƐĞĚ ĨƌŽŵ ϱϮй in January 2016 to 65% in July 2016. The applicant suggests that the difference between these survey results and those of other studies reflects increasing popularity and awareness of IQOS. Of note, purchasing IQOS and registering the device was a requirement for inclusion in the larger and more recent survey; this may not be a representative sample of all users.

Dual use of CC and IQOS appears likely. There is concern about the effects that dual use of IQOS and CC (compared to complete switching) will have on long-term reduction of HPHC exposures and the health risks for tobacco-related diseases. While results from the PBA-07 study showed that IQOS use was associated with reduction in cigarette consumption, the health benefits of reducing cigarette consumption instead of quitting completely are unclear.

3. Use of IQOS by Former or Never Smokers and Youth In PBA-05-NOC, the applicant assessed perceptions and intention to use IQOS among former smokers, young adult never smokers (aged 18-25 years), and other never smokers. The applicant also conducted research on non-smokers’ use of heated tobacco products in Japan, where IQOS is on the market. The epidemiology review describes results from two published studies from Japan and Italy that reported the prevalence of IQOS use in never and former smokers after IQOS was marketed in these countries.

73

For PBA-05-NOC the applicant developed LLA materials, including an IQOS brochure, Heatsticks pack, and direct mail communication. Measures were developed for assessing intentions to try and use IQOS. The applicant examined the percentages of participants who reported that they will Definitely or Very likely use IQOS, which were the top two categories on a six-point response scale ranging from Definitely not to Definitely. The LLA materials provided information to distinguish IQOS from e-cigarettes including statements about “real tobacco” and an appearance similar to CC; they contain a tobacco plug wrapped in paper. In Arm 3, which presented a photo of the IQOS device with a Surgeon General’s warning but no additional information about IQOS, the applicant notes, “It is likely that Risk Perceptions are primarily based on factors such as the appearance of the THS device (which is similar to some e-cigarettes) and the impression that the product is innovative and new.” These perceptions related to similarities between IQOS and e-cigarettes may be important when considering potential appeal among people who do not currently smoke. The brochure included a statement that the product is intended for smokers who want to continue using tobacco and is not intended for use by non-smokers. Heatsticks would be marketed under the Marlboro brand name, which consumers may associate with CC. As a cigarette product, Heatsticks cannot be marketed with characterizing flavors aside from tobacco or menthol; the availability of different flavors is a commonly-cited reason for never smokers’ use of e-cigarettes.78 79 These characteristics may reduce the appeal to nonsmokers.

FDA questioned the applicant’s decision to define ‘positive intent’ as those responding Definitely or Very likely but excluding Somewhat likely. The applicant explained in the September 2017 amendment that the decision was made to prioritize specificity over sensitivity, i.e., to minimize false positive predictions of actual (post-market) trial/use. The applicant accurately notes there is no consensus in the literature on how intention should be measured or what kind of assessment most accurately predicts uptake.

In response to an inquiry from FDA, the applicant submitted information on intention to try and use IQOS among current and former smokers in Arm 2 and Arm 3 of PBA-05-NOC based on whether the participant viewed the Regular, Smooth Menthol, or Fresh Menthol Heatsticks pack. Former smokers were assigned to view a menthol variant. Although the analyses were descriptive and sample sizes were small, a slightly greater proportion of former smokers randomized to view Smooth or Fresh Menthol variants intended to try or use IQOS compared to those that saw the Regular Heatsticks pack (7-8.1% vs. 4.8% to try IQOS; 5.4% vs. 2.9% to use IQOS). This could indicate that former smokers who used menthol cigarettes may be more likely to try and use IQOS than former regular cigarette smokers.

Never smokers in PBA-05-NOC, including young adults (legal age to 25 years), were only exposed to the Regular Heatsticks pack in Arm 2 and Arm 3. This is a study limitation since menthol cigarette smokers comprise one-third of the U.S. market. The 2011 NSDUH Report noted that while rates of menthol cigarette use among 12-17-year-olds were stable between 2004 and 2010, more than half (51.7%) of those who had smoked a cigarette for the first time in the prior 12 months smoked menthol cigarettes.80

Among never smokers and young adult never smokers, only 0-1% who viewed the LLA materials indicated they would Definitely or Very Likely try IQOS. The results for former smokers were slightly higher; of those

78 Berg, C.J. (2016). Preferred flavors and reasons for e-cigarette use and discontinued use among never, current, and former smokers. International Journal of Public Health, 61(2), 225-236. 79 Kong, G., Morean, M.E., Cavallo, D.A., et al. (2015). Reasons for experimentation and discontinuation among adolescents and young adults. Nicotine & Tobacco Research, 17(1), 847-854. 80 Substance Abuse and Mental Health Services Administration, Center for Behavioral Health Statistics and Quality. (November 18, 2011). The NSDUH Report: Recent Trends in Menthol Cigarette Use. Rockville, MD (Amendment December 17, pg. 42).

74

who viewed LLA materials, 5-6% indicated they would Definitely or Very Likely use IQOS. When never smokers and young adult never smokers were asked about intent to use if offered by a friend, the positive responses were minimally higher; 2-7% indicated they would Definitely or Very Likely try IQOS. Former smokers indicated positive intent to try IQOS at 15-24% if offered by a friend.

When responses of Somewhat likely are included in estimates of positive intention to use IQOS, 4-7% of never smokers and 7-11% of young adult never smokers reported an intention to try IQOS. Among former smokers, 17-25% reported an intention to try IQOS, with 7-14% reporting an intention to use regularly if they tried IQOS and liked it. If offered by a friend, 33-42% of former smokers responded they would Definitely, Very Likely or be Somewhat likely to try IQOS.

For comparison, the applicant also asked former smokers about their intention to use e-cigarettes regularly and asked never smokers about their intention to try e-cigarettes. Former smokers’ intention to use and never smokers’ intention to try IQOS appeared to be similar to or somewhat lower than their intention to use or try e-cigarettes, although the applicant provided no statistical analysis of these differences.

In the U.S., most cigarette smokers begin trial and progression to regular use before age 18 (USDHHS 2012; USDHHS 2014). While oversampling of young adult never smokers in PBA-05-NOC is a strength of the study, the applicant did not submit any information or bridging study data to youth under age 18 and did not stratify information submitted in the PBA-05-NOC and PBA-07 studies by age beyond the 18-25-year-old age group. Excluding current smokers who started smoking cigarettes in the prior 30 days is also a limitation; this may have limited the inclusion of young adult cigarette smokers in the adult smoker groups since young adult cigarette smokers are often smokers or in a recent state of transition.81 82 As the applicant notes, FDA clarified during a September 5, 2013 meeting (TC0000737) that studies in youth were not expected; however, the applicant did not include bridging information on youth use of other products (e.g., cigarettes, e-cigarettes). This might have helped FDA better understand youth intentions and perceptions with respect to IQOS.

The applicant conducted cross-sectional studies to monitor the prevalence of heated tobacco product use by adult non-smokers (age 20 or older) in Japan. During the first one to two years after IQOS went on the Japanese market in 2014, use by adult former and never smokers was low (1.5% among former smokers and 1.2% among never smokers). Additional internet surveys were conducted in 2015, 2016, and 2017. Panelists aged 15-69 years from a major Japanese internet research agency provided information on current use (i.e., any use in the previous 30 days) of IQOS, other heated tobacco products, e-cigarettes, and combustible cigarettes.83 Among survey responders in 2017, there were 3.6% current IQOS users and 2.0% of those aged 15-19 years reported current use of IQOS in 2017. Of the 2017 survey responders 1.3% were never smokers, 2.1% were former smokers, 18.8% were current smokers with intention to quit, and 10.3% were current smokers with no intention to quit.

81 Lenk, K.M., Chen, V., Bernat D.H., Forster J.L., Rode, P.A. (2009). Characterizing and comparing young adult intermittent and daily smokers Subst. Use Misuse, 44, 2128-2140. 82 U.S. Department of Health and Human Services (USDHHS). (2014). The Health Consequences of Smoking—50 Years of Progress: A Report of the Surgeon General. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. 83 Tabuchi, T., Gallus, S., Shinozaki, T., Nakaya, T., Kunugita, N., & Colwell, B. (2017). Heat-not-burn tobacco product use in Japan: its prevalence, predictors and perceived symptoms from exposure to secondhand heat-not-burn tobacco aerosol. Tob Control, [Epub ahead of print].

75

In 2017, a face-to-face survey was conducted in Italy among 3000+ ƉĂƌƚŝĐŝƉĂŶƚƐ ĂŐĞĚ шϭϱ LJĞĂƌƐ ƐĞůĞĐƚĞĚ from the general Italian population.84 Based on cigarette smoking status, 1.0% of never smokers, 0.8% of former smokers, and 3.1% of current smokers reported having ever tried IQOS. Among participants who reported that they had never tried IQOS but were intending to try it, 1.7% were never smokers, 0.5% were former smokers, and 5.0% were current smokers.

The post-market surveys conducted in Japan and Italy also measured IQOS use among youth. In Japan, 2.0% of those aged 15-19 years reported current use of IQOS in 2017. The prevalence of current use was lower in youth than in those aged 20-29 years (5.8%), 30-39 years (5.4%), 40-49 years (3.9%), and 50-59 years (3.7%). In the Italian study, 0.9% of those aged 15-24 years reported having ever tried IQOS, compared to 1.0% of those aged 25-44 years, 2.4% of those aged 45-ϲϰ LJĞĂƌƐ͕ ĂŶĚ ϭ͘Ϭй ŽĨ ƚŚŽƐĞ ĂŐĞĚ шϲϱ LJears. Youth and young adults aged 15-24 years also had a slightly lower prevalence of participants reporting that they had never tried IQOS but were intending to try it (1.9%), compared to those aged 25-44 years (2.9%) and 45-64 years (2.5%) who said they were intending to try IQOS.

Overall, the available information suggests the prevalence of IQOS use is lower in never smokers compared to current smokers and that fewer youth than adults currently use IQOS in Japan or Italy. The data from countries where IQOS is marketed, Italy and Japan, show low uptake by youth and current nonsmokers. These two published survey studies are the only data currently available on the prevalence of IQOS use in youth.

The PBA-05 study also suggests a low prevalence of intention to use IQOS among never smokers. The likelihood is slightly higher in former smokers, but still low. As noted by the applicant, these data may not be as sensitive for less decisive responses, e.g., Somewhat likely. There is no agreed-upon method for conducting these types of studies where theoretical choices are being made that have no true consequence. Introducing additional conditions to the study scenario, e.g., intent to try or use IQOS if offered by a friend, makes interpretation of the data even more uncertain. These studies, while providing an indication of intent among smokers, nonsmokers, and former smokers, cannot be considered as absolute indicators of behaviors when/if IQOS is a marketed product.

Certainly, the potential for rapid uptake of a novel tobacco product among youth exists. In the decade since e-cigarettes were introduced to the U.S. market, youth use rose rapidly but the limited flavor choices may reduce IQOS’ appeal to youth. The limited options in terms of flavor choice and the price of the IQOS device may reduce the appeal to youth. Given that IQOS is still a relatively new product to Italy and Japan, the extent to which youth will initiate and use IQOS in these markets, or any other market that may start selling IQOS, is unknown though the trend from other countries indicates that this is uncommon. Overall, the current evidence indicates IQOS uptake by youth and nonsmokers will be low.

4. Likelihood of IQOS leading to Conventional Cigarette Smoking Cessation Both the PBA-07 and WOT studies evaluated the likelihood of smokers switching to IQOS. During the six- week observational period of the PBA-07 study, 33.8% of current smokers initiated use of Heatsticks (defined as consuming шϭϬϬ Heatsticks). Among those who started using Heatsticks, 16.3% were exclusively using Heatsticks ;шϵϱй Heatstick use) during Week 6. Of those who switched to Heatsticks in an earlier

84 Liu, X., Lugo, A., Spizzichino, L., Tabuchi, T., Pacifici, R., Gallus, S. (2018). Heat-not-burn tobacco products: concerns from the Italian experience. Tob Control, [Epub ahead of print].

76

week, 15.5% had reverted to predominantly using cigarettes (i.e., Heatsticks ǁĞƌĞ чϯϬй ŽĨ ƚŽƚĂl cigarettes + Heatsticks consumed in a week) by the last week.

In the WOT, the prevalence of using Heatsticks varied by country. Exclusive Heatstick use among those who had used at least 100 Heatsticks ranged from 7.8% in Switzerland to 21.5% in Japan. The proportion of Heatstick initiators who switched from Heatsticks back to cigarettes ranged from 0% in Japan to 10.3% in Italy. Exclusive and predominant IQOS use was most common in Asia where these outcomes were observed in 14% and 16%, respectively, of Japanese smokers and 16% and 22%, respectively, of South Korean smokers. Although more IQOS users may quit smoking over time, data from the PBA-07 study and the WOT study show that most smokers become dual users or at least go through a dual use phase before quitting.

In the Japanese post-market study of IQOS purchasers who registered their device in an online database, 52%-65% of IQOS purchasers were considered exclusive IQOS users. However, those who take the initiative to register their device are likely to be a non-representative sample of all Japanese IQOS users and may be more motivated to become exclusive IQOS users. Also, nicotine containing e-liquid is categorized as a pharmaceutical ingredient in Japan and nicotine-containing e-cigarettes are not as readily available in Japan as they are in the U.S.

Although less than 10% of cigarette smokers in the U.S. PBA-07 study switched to exclusive IQOS use, the proportion of exclusive IQOS users remained steady during the 6-week observational period. This suggests that individuals who initiate IQOS and use Heatsticks for at least 95% of their tobacco intake are able to maintain exclusive IQOS use over time and potentially replace their use of CC with Heatsticks long-term.

5. Population Modeling The applicant presented results from a Population Health Impact Model (PHIM) to assess the possible effects of the proposed new products on population health in the U.S. This is a computational and simulation model that tracks tobacco prevalence and deaths from four specific smoking-related diseases: lung cancer, ischemic heart disease, stroke, and COPD on a hypothetical population exposed to two tobacco products - CC and THS 2.2. The model consists of two quantitative components: The “Prevalence Component” (P-Component) in which individual smoking histories are simulated over a follow-up period from 1990 to 2009, and the “Epidemiological Risk Component” (E-Component) in which the smoking histories produced by the P-component are used to estimate smoking-related deaths for each morbidity.

The initial population in the scenarios is representative of the U.S. in 1990, and the scenarios are modeled for a twenty-year period. x Null Scenario: THS 2.2 is not introduced into the market. This scenario considers three possible tobacco use statuses: never smoker (N), current smoker (C), former smoker (F). x THS Scenario: THS 2.2 is introduced into the U.S. market. This scenario considers five possible use statuses: N, F, current cigarette smoker (CC), current THS user (CT) and current dual user (CD).

The applicant presents results from a THS scenario called the “business case” that uses a specified set of input values and assumptions. The applicant’s findings are dependent on the following basic assumptions; additional details are in the epidemiology review: x Within 10 years of being on the U.S. market the new tobacco products will be used by 17% of U.S. smokers. Approximately 15% of users will be exclusive users and 2% will be dual users with cigarettes. x Over a twenty-year period approximately 30% of smokers will be users of the new products.

77

x Most current cigarette smokers transitioning to the new products will be middle aged; younger people are less likely due to cost and older people are generally less likely to switch. x The new products would not change the combined initiation, re-initiation, or cessation rates for cigarette smoking but would change the distribution of use of these products in the THS scenario with the introduction of new product and dual use. The applicant concludes, “Overall, based on the modeling results and scenario specifications, introducing THS into the US population appears to lead to a sizeable public health benefit in terms of reduced cigarette smoking and tobacco-related mortality. Variation in the model parameter estimates within reasonable ranges would not materially change these conclusions.”

FDA evaluated the applicant’s approach. The model considers deaths from four conditions (lung cancer, COPD, ischemic heart disease, and stroke). According to the 2014 U.S. Surgeon General’s Report,85 these causes account for approximately 336,000 of 437,000 deaths directly attributable to cigarette smoking among U.S. adults. The model does not account for changes in the number of deaths from environmental tobacco smoke exposure due to use of the proposed new tobacco products. The prevalence estimates used by the applicant are higher than those observed in more recent years, with U.S. adult smoking prevalence having been 20.6% in 200986 and having since declined to levels around 15%.87 As such, model estimates may tend to overestimate the number of current smokers in the baseline case compared with the present population and could overestimate any population health impact of smokers switching to another tobacco product.

The business case projects that the proposed new tobacco products will come to represent a substantial proportion of the smoking market in the U.S., accounting for ~17% of users in 10 and 30% of users in 20 years, with most being exclusive users. The modeling section does not present empirical evidence to support this forecast. If uptake of the products by consumers is lower, takes more time, or is more likely to occur as part of dual use, then the magnitude of any population health effects would be expected to be reduced. The assumption that relative exposure for dual users is the mean of relative exposure for smokers and proposed new product users may underestimate risk; exposure among dual users may not be the average of exposure of exclusive cigarette and THS users. In addition, individual harm from exposure to exclusive or dual use with the proposed products may not follow a linear dose-response relationship.

The applicant included a series of simulation results to describe the potential impact of IQOS marketing of the health of the population: x 20 years of cessation - This simulation assumes all current smokers would stop smoking immediately. Under this scenario, the smoking prevalence will be zero during the 20-year follow-up period and the risks associated with smoking-related diseases diminish over time. As a result, over the 20-year simulation period (1990-2009), the 100% cessation assumption would result in 934,947 fewer smoking-attributable deaths. x 20 years of THS with no cigarettes - This simulation assumes that all current smokers in 1990 transitioned immediately to THS rather than quit smoking. Also, it is assumed that initiation and relapse rates changed, and future smokers will use THS only. Under this scenario, two relative exposure (f-values) were considered: f-value=0.1 (THS preserved the effects of cessation by 90%)

85 US Department of Health and Human Services. The Health Consequences of Smoking – 50 Years of Progress: A Report of the Surgeon General, 2014. Atlanta, GA 2014. 86 Centers for Disease Control and Prevention. Cigarette smoking among adults aged >= 18 years – United States, 2009. Morbidity and Mortality Weekly Report. 2010;59(35)1135-1140. 87 Phillips E, Wang TW, Husten CG, et al. Tobacco Product Use Among Adults – United States, 2015. Morbidity and Mortality Weekly Report. 2017;66(44):1209-1215.

78

and f-value=0.3 (THS preserved the effects of cessation by 70%). Based on these assumptions, over the 20-year simulation period, the introduction of THS would result in 780,433 (if f-value=0.1) or 516,944 (if f-value=0.3) fewer smoking-attributable deaths. x World Health Organization 2025 target and projection: The 2015 WHO Report targets a 30% reduction in smoking prevalence from 22.1% in 2010 to 15.4% in 2025, with a revised projection of 18.9% in 2025 representing only a 14% reduction. In this simulation, the PHIM was used to estimate the impact of reducing smoking prevalence by 30% (WHO 2025) and 14% (WHO revised) over a 15- year period (1990-2005). A null scenario (no THS into the market) was also used to compare the projected smoking prevalence assuming 30% and 14% reduction in prevalence. The results indicate that, under the null scenario, smoking prevalence remained somewhat constant over the 15-year simulation, with 27% and 24% smoking prevalence for males and females, respectively. Under the WHO 2025 scenario, in 2005 the smoking prevalence was 19% for males (29.6% prevalence reduction) and 16% for females (27.3% prevalence reduction), resulting in 172,458 fewer smoking- attributable deaths cumulatively over 1990-2009. Under the WHO revised scenario, in 2005 the smoking prevalence was 22% for males (18.5% prevalence reduction) and 18% for females (18.2% prevalence reduction), resulting in 111,102 fewer smoking-attributable deaths cumulatively over 1990-2009.

There are no major concerns with the statistical and computational aspects of the PHIM. Overall, the simulation results suggest that the introduction of THS 2.2 into the commercial market will reduce the overall morbidity and mortality from tobacco products. However, there are limitations to the PHIM modeling assumptions, input data construction, and inference procedures. The model only considers two products – cigarettes and IQOS; other tobacco products were not considered in the simulations. Furthermore, the population size does not change over time. There is also a question as to whether the general approach for modeling risk reduction with the proposed products, which is based on reduction in risk based on the time since complete smoking cessation, is appropriate when used to represent risk caused by continuing use of a tobacco product. The applicant provides very little justification and no specific empirical evidence to support the assumptions that individuals who do not currently smoke cigarettes would not be interested in using the proposed products or that young people would not find them appealing. Finally, the relatively short projection period of 20 years and use of mortality as a health outcome does not allow for adequate consideration of the long-term health effects of tobacco use initiation among youth and youth adults. The projected population health effects of the proposed new tobacco products may be overstated if specific assumptions about tobacco use behavior and risks are not realized in the actual population. Although the model is statistically valid, the overall analysis of the population model does not provide evidence to support the application.

6. Amendment MR0000117; Study ZRHR-ERS-09-US On June 11, 2018, FDA received amendment MR0000117 to MR0000059-61, which includes the final study report for ZRHR-ERS-09-US. Although this amendment was directed to the MRTPAs, since the amendment included information with respect to the products that are the subject of the PMTAs, the amendment is considered here.

Study Title: Evaluation of Biological and Functional Changes in Healthy Smokers After Switching to THS 2.2 for 26 weeks

Study Design: This was a randomized, controlled, open-label, 2-arm, parallel group, multi-center clinical study of six months of ad libitum use of the non-menthol THS 2.2 compared to continued CC users in an

79

ambulatory setting. Participants were healthy adult (age > 29 years) non-menthol CC smokers not interested in quitting within the next six months. This study was conducted in 20 clinical sites across the continental U.S. All participants used THS 2.2 for a one-week run-in period and those who were willing were considered for randomization after this period.

Study Population: Of 984 subjects, 488 were randomized to THS 2.2 and 496 to CC. Study participants had a mean age of 44.6 years, 58.8% male, 79.2% white and 17.6% African American. Most (62%) had high school education and 31.9% had a college education or higher. Smoking duration average was 26.2 years. Mean CPD for the past year was 19.3 and most were moderately (45%) or severely (39.3%) dependent. There were insufficient data for analysis of 127 subjects, leaving 857 in the analysis population. The group descriptions and numbers are shown in Table 9.

Table 9: Study Population ZRHR-ERS-09-US Number (Percent) Analyzed Category Description THS Group CC Group >70% THS use over entire analysis period and THS use 245 (51.4%) >70% THS use on >50% of the days in the analysis period 1% < THS < 70% over the entire analysis period or Dual use 142 (29.8%) THS-use and CC-use on <50% of the days <1% < THS over the entire analysis period and CC use 3 (0.6%) 425 (86%) <1% THS on >50% of the days in the analysis period General category encompassing subjects with missing product use, those using e-cigarettes or other tobacco Other use products, those who quit, or subjects who switched across 24 (4.9%) 18 (4.2%) different use patterns between consecutive analysis periods Table created based on information in amendment MR0000117 Overview

Primary Study Objective: To demonstrate favorable changes after six months across eight co-primary clinical risk endpoints (referred to by FDA as BOPH) for those switching from CC to THS as compared to continued CC use. The co-primary endpoints are: HDL-C, sICAM-1, total WBC, COHb, 11-DTX-B2, 8-epi-W'&^Ϯɲ͕ &sϭ͕ ĂŶĚ ƚŽƚĂů NNAL. The applicant defined success as: 1. Statistically significant improvements in at least five of the eight endpoints 2. All endpoints changing in the direction observed with smoking cessation

Primary Study Results: Five of the eight endpoints showed a statistically significant change in smokers who switched from cigarette smoking to THS use. All BOPH shifted in the direction seen when smokers quit, as described in literature. Results are summarized in Table 10.

80

Table 10: Primary Analysis of CREs between THS Use and CC Use at Six Months for ZRHR-ERS-09-US LS Mean Difference or 1-sided p-value Endpoint Change from CC-use Relative Reduction *=statistically significant HDL Difference 3.09 mg/dL <0.001* WBC count Difference -0.420 GI/L 0.001* sICAM-1 % Reduction 2.86% 0.030 11-DTX-B2 % Reduction 4.74% 0.193 8-epi-W'&Ϯɲ % Reduction 6.80% 0.18 COHb % Reduction 32.3% <0.001* FEV1 %predicted Difference 1.28% predicted 0.008* Total NNAL % Reduction 43.5% <0.001* Table created based on Table 4 in MR0000117 Overview

Secondary Study Objectives: x Evaluate self-reported product use (THS and/or CC) and nicotine exposure levels x Evaluate exposure reduction to selected HPHCs (BOE) in dual use and THS use groups

Significant reductions in exposure levels in the THS group and the dual use group are noted in Figure 10.

Figure 10: THS:CC (light grey) and Dual Use:CC (dark grey) Ratios (%) and Difference (ppm) and 95% CI at Month 6

Source: Figure 2, MR0000117 Overview

Baseline nicotine equivalents (NEQ) levels ranged from 9.2-10.3 mg nicotine/g creatinine across the category groups. At Month 6, the geometric least square mean values were almost identical in the THS and CC groups at 8.92 and 8.86 mg nicotine/g creatinine, respectively. In the dual use group, the LS NEQ values were slightly lower at 8.34 mg/g creatinine. The applicant believes these results confirm that THS 2.2 can deliver nicotine at levels comparable to CC and that adult smokers can accept THS as an alternative to CC.

81

Additional Analysis: Concomitant CC Use The applicant notes this study assessed the effect of THS as “actually used” considering that a significant amount of concomitant use (up to 30%) may occur and this could reduce the risk reduction potential of THS. The applicant believes the level of concomitant use seen in this study is consistent with the experience in markets where THS is commercialized. In these markets, the applicant has observed that 70-90% of THS users use THS in >70% of all tobacco use experiences. The applicant also notes that the primary study objective was met and HPHC exposures (BOE were secondary endpoint) were reduced, even with the concomitant use pattern.

Adverse Events (AEs) Associated with the Study There were 19 serious AEs reported by 13 subjects: 8 events in 6 subjects in the THS arm, and 11 events in 7 subjects in the CC arm. None of the serious AEs was believed related to THS or CC by the applicant. There were two deaths: Subject 04-384 was randomized to the THS 2.2 arm on Oct 7, 2015. On Dec 16, 2015 the subject was found deceased in the bathtub at his residence; cause of death was acute and chronic alcohol abuse. Subject 14-101 was randomized to the THS 2.2 arm on Jan 11, 2016. He died of a self-inflicted gunshot wound on Jan 25, 2016.

During the exposure period, 415 subjects reported 758 AEs: 358 events in the THS arm and 400 in the CC arm. Most AEs were mild or moderate in severity. Three subjects discontinued the study from the THS arm and two from the CC arm. The most common AEs were upper respiratory tract infections (4.3% in THS, 4% in dual use, and 6.2% in CC).

Applicant’s Conclusions x Among subjects randomized to use THS, 34% used THS exclusively (defined as >95% use). Another 34% dual-used THS and CC. The applicant believes these results show the product was well accepted considering that before switching, subjects were naïve to the product. x Overall, all the clinical risk endpoints (BOPH) evaluated in those switching from CC to THS followed the same direction as seen following smoking cessation. The changes were statistically significant in five of the eight BOPH measured. x In addition to NNAL and COHb measured as BOPH, eight BOE were assessed. In all cases, there was significant reduction in THS users compared to CC users. x Exposure to nicotine was comparable between THS and CC users. x With respect to dual use (defined as subjects whose THS use was 1-70%): o BOPH showed a shift (although minor and not generally significant) in the favorable direction at six months compared to CC use. o BOE showed slight reductions compared to CC at six months (2.6-13%)

FDA Statistical Analysis There are limitations to the applicant’s statistical approach: x Although the report describes this as a controlled clinical trial, the study design is ambulatory. The study staff did not control participants’ exposure to the products. This is part of the study design and does not affect data validity. x While the initial randomization scheme is acceptable, the applicant used modified groups for the primary analysis. In the newly-defined post-randomization “Actual Product Use Categories,” only 245 of the randomized 488 participants in the THS 2.2 group were included in the primary analysis.

82

Eliminating about 50% of participants breaks the initial balance between the THS 2.2 and CC study arms obtained via randomization. Therefore, significant differences between the THS 2.2 and CC study arms may be due to factors unrelated the exposure. x The data analyses assumed that the individual outcomes (changes in BOPH) are independent for each of the eight primary biomarkers. This is unlikely. The BOPH selected are affected by multiple factors, including general health, other medical conditions, infections or other inflammatory processes, genetics, age, diet, exercise, and medications. It is difficult to consider these biomarkers as individual measures of tobacco-related disease and the levels are unlikely to change independently. Nonetheless, the measured changes in BOPH are valid, even if not independent. x The applicant provided no scientific justification suggesting that BOE and BOPH related to CVD, cancer, and lung function are appropriate to combine as an overall metric of clinical significance. x Because the study arms became imbalanced at the six-month time point, the use of multiple comparisons performed with the Halperin-Ruger statistical method is not justified.

Due to limitations in design and statistical analysis, no definite conclusions can be made based on this study. However, despite these limitations, the study provides evidence of reduced exposures associated with switching completely from CC to THS. Additionally, those with self-reported dual use had no evidence for increased toxin exposures. The rate of self-reported dual use during this longer study was lower (~34%) than the considerably higher dual use rates in previous studies (~58% in the Actual Use study). There were no unexpected safety signals identified during the study and the rate of adverse events was similar for those exposed to THS and CC. Nicotine exposure levels were also comparable between THS and CC.

7. Summary of Population Health Findings The social science review concludes that based on the information submitted by the applicant, we have concerns with respect to: the lack of information about youth under age 18, as well as the lack of a discussion of submitted data’s applicability to youth and the lack of presentation of the data in stratified categories that would allow us to make inferences about youth, the potential for initiation among young adult never smokers, and the potential for dual use among current smokers with only a one cigarette per day decrease in use frequency. Philip Morris Products S.A.’s premarket tobacco product applications do not contain sufficient information to address these concerns from a Social Science perspective.

As TPL, I do not agree with these social science conclusions. I agree there are limited data regarding use and possible uptake of IQOS in youth. However, I disagree that there is no data. The applicant provided data a Japanese internet research agency which included panelists ages 15-69 years; they found 2.0% of those aged 15-19 years reported current IQOS use in 2017. Additionally, the Epidemiology review team reports a face-to-face survey conducted in Italy with >3000 participants, age >15 years found 0.9% of those aged 15-24 reported ever trying IQOS. The data from countries where IQOS is marketed, specifically Italy and Japan, show low uptake by youth and current nonsmokers. Overall, the current evidence indicates low IQOS uptake by youth.

I agree there are concerns about dual-use. There is evidence that U.S. cigarette smokers are interested in IQOS, but limited data for use of IQOS to achieve CC smoking cessation. The company states they intend to market IQOS ‘for adult smokers who wish to completely switch.’ The limited data available indicates that a dual-use period is common during the switching period, but those who switch ‘quickly and completely’ were more likely to successfully remain off conventional cigarettes. There are data that HPHC exposures are not increased in those who dual-use IQOS and cigarettes. In fact, HPHC reductions continue through the 90 extended exposure studies even though during the last 85 days of these studies participants were not in

83

controlled environments and dual-use was likely. Additionally, although the changes were not statistically significant, the six-month study showed decreases in BOE for dual users as compared to exclusive CC users. The studies conducted by the applicant have not demonstrated reduction in long-term disease risk; however, the reduced exposures combined with the other available information, lead me to conclude IQOS is appropriate for protection of public health, even if there is some dual-use among smokers as they potentially transition to the product.

The epidemiology review concludes the applicant has demonstrated that the exclusive use of the products that are the subject of these applications exposes users to substantially lower exposure to many HPHCs compared to conventional cigarette smoking. The review also recommends any marketing authorization granted in response to these PMTAs be accompanied by requests for information, collected under a real- world context, on the differences in BOE in CC smokers that completely switch to the products that are the subject of these applications compared to those who dual use the products with CC. Additional clinical evaluation of the 53-62 compounds found at higher levels in the aerosol of the products that are the subject of these applications compared with cigarette smoke would also be helpful for supporting continued marketing of the products as appropriate for the protection of public health. Finally, long-term evaluation that assess changes in BOPH as well as clinical endpoints associated with complete and incomplete switching to the products that are the subject of these applications would also provide support for the continued marketing of the products as appropriate for protection of public health.

As TPL, I agree with the epidemiology review conclusions. I also agree that continued information regarding toxic exposures as the products are actually used, including both CC who switch completely and those who use multiple tobacco products, will be helpful information for supporting the continued marketing of these products as appropriate for the protection of public health. Additionally, continued information regarding long-term health effects, such as may be obtained with additional BOPH studies of longer duration, may also provide support for the continued marketing of the products as appropriate for protection of public health.

84

III. Product Labeling, Consumer Comprehension, and Marketing Plan A. Proposed PMTA Labeling The following sample labeling materials were provided: x Heatstick pack labeling for Regular, Smooth Menthol, and Fresh Menthol x Heatstick carton labeling Regular, Smooth Menthol, and Fresh Menthol x IQOS device package (black kit and white kit) x IQOS printed film for Regular, Smooth Menthol, and Fresh Menthol The proposed labeling has been evaluated by CTP Office of Compliance and Enforcement, Division of Promotion, Advertising, and Labeling (OCE DPAL) and they conclude there is no evidence to suggest the planned labeling (other than discussed below) is false or misleading.

B. Consumer Comprehension The submission included a copy of the IQOS Tobacco Heating System User Guide and the IQOS Quick Start Guide. The User Guide provides comprehensive instructions for use including information for device storage, cleaning, charging, and disposal. The Quick Start Guide provides the basic information needed to use the IQOS system and a high-level explanation of the device indicator lights, buttons, and accessories.

Study PBA-06-US was designed to describe the ability of prospective consumers to correctly understand and comply with THS Instructions for Use. Adult smokers (N=258) reviewed the provided instructions and were asked to perform nine tasks and answer three comprehension questions related to the materials. No product was administered during this study.

Based on the range of the proportions of subjects who correctly demonstrated or comprehended the tasks and instructions, the applicant concluded that “a relatively large majority of subjects” were able to correctly demonstrate the following tasks: x Task 1 (How to charge the THS 2.2 Holder and Pocket Charger Simultaneously) x Task 2 (How to Insert a THS 2.2 Tobacco Stick into the THS 2.2 Holder) x Task 3 (How to Heat and Consume a THS 2.2 Tobacco Stick) x Task 4 (How to Know When a THS 2.2 Tobacco Stick Has Been Consumed)

The applicant concludes that subjects “found the following tasks more complicated:” x Task 5 (How to Remove a THS 2.2 Tobacco Stick from the THS 2.2 Holder) x Task 6 (How to “Heat Clean” the THS 2.2 Holder) x Task 7 (How to Clean the THS 2.2 Holder with the THS 2.2 Cleaning Tool) x Task 8 (How to Remove a THS 2.2 Tobacco Stick Stuck from the THS 2.2 Holder Cap) x Task 9 (How to Re-attach the THS 2.2 Holder Cap to the THS 2.2 Holder Body)

In addition, the applicant concluded that a majority of the subjects understood that the THS 2.2 Holder is to be used only with the Tobacco Sticks (85.3% “correct” or “acceptable”) and that the Tobacco Sticks should not be lit with a lighter (94.6% “correct” or “acceptable”). However, more subjects had difficulty understanding that the THS 2.2 Holder needed to be fully charged before it could be heat cleaned (66.7% “correct” or “acceptable”).

85

FDA statistical reviewers conducted an analysis based on the percentages of participants providing a “correct” or “acceptable” response to all steps for each task. The reviewers reported the following results: x Task 1 (How to Charge the THS 2.2 Holder and Pocket Charger Simultaneously): 63% x Task 2 (How to Insert a THS 2.2 Tobacco Stick into the THS 2.2 Holder): 52% x Task 3 (How to Heat and Consume a THS 2.2 Tobacco Stick): 68% x Task 4 (How to Know When a THS 2.2 Tobacco Stick Has Been Consumed): 81% x Task 5 (How to Remove a THS 2.2 Tobacco Stick from the THS 2.2 Holder): 34% x Task 6 (How to “Heat Clean” the THS 2.2 Holder): 29% x Task 7 (How to Clean the THS 2.2 Holder with the THS 2.2 Cleaning Tool): 41% x Task 8 (How to Remove a THS 2.2 Tobacco Stick Stuck from the THS 2.2 Holder Cap): 68% x Task 9 (How to Re-attach the THS 2.2 Holder Cap to the THS 2.2 Holder Body): 73%

Statistical inference was not the basis for informing the conclusion-making process in this study, therefore the results are not generalizable to the U.S. population. Tasks were demonstrated in a structured, monitored setting, and may not be representative of performance in a real-life setting. Also, performances were scored based on a participant’s first attempt at the use tasks. Consequently, results may not be indicative of performance after users become familiarized with the product through repeated attempts. Overall, the results demonstrate sufficient consumer understanding of the products and their use. Additionally, the applicant has stated their intent to (b) (4) (b) (4) . The additional support along with the instructions that are included with the IQOS device should resolve most consumer issues related to product use.

C. Marketing Plan At the request of FDA, the applicant provided a summary of their plans for marketing of IQOS in the U.S., assuming marketing authorization is granted. The marketing plan encompassed the following main concepts: (b) (4)

86

(b) (4)

C. Nicotine is Addictive Labeling As cigarettes, if authorized without changes, IQOS Heatsticks would bear the rotating Surgeon General’s warnings required under the Federal Cigarette Labeling and Advertising Act (FCLAA). These warnings do not currently include a warning related to nicotine and addiction. This raises concerns because studies suggest that people do not accurately perceive the risk of addiction associated with IQOS use. This, in turn, could have negative consequences to public health in terms of increased initiation among nonusers and decreased cessation among tobacco users.

As discussed in more detail in section II.D, II.E, and II.F above, the applicant conducted multiple studies to evaluate the nicotine delivery, addiction potential, and abuse liability of IQOS. These included four single use PK/PD studies, four reduced exposure 5-day and 90-day studies, and the U.S. actual use study, evaluating the use of IQOS in an “almost real world” environment. Systemic nicotine exposure was similar after single and multiple uses of IQOS (both regular and menthol Heatsticks) and CC. In addition, self-report questionnaires found that IQOS produced reinforcing effects close to those of CC. Overall, the data from these studies show that IQOS is addictive and has nicotine delivery, addiction potential, and abuse liability similar to combusted cigarettes.

However, study data show that consumers do not accurately perceive and tend to underestimate the addiction risk of IQOS. In the U.S. consumer perception study PBA-05-NOC, the applicant assessed the perceived addiction risk of using IQOS, combusted cigarettes, e-cigarettes, NRTs, and cessation among 1,829 current smokers, former smokers, and never smokers (including, importantly, 18-25 year old never smokers) after exposure to various IQOS label, labeling, and advertising materials containing the Surgeon General’s warnings, including what the applicant described as the Heatstick pack intended for commercialization.88 Perceived addiction risk scores for each product type were transformed and reported on a 100 point scale, with 0 = No Risk, and 100 = Very High Risk. After viewing the IQOS LLA materials with the Surgeon General’s warnings, study participants rated IQOS as 10-20 points less addictive than combusted cigarettes. This was true across all study arms for adult current smokers, former smokers, and never smokers, including young adult never smokers. See Figure 11, showing the results for study arm 2, the Heatstick pack intended for commercialization. As shown in the figure, there was only a small overlap in the 95% confidence intervals in the adult former smoker group and no overlap in the confidence intervals in any other group, indicating a statistically significant difference in the perception of addiction risk between IQOS and combusted

88 For more details on the study design, see section II.F.1 above and the discussion in the Social Science review.

87

cigarettes. This difference persisted when the applicant, prompted by FDA, conducted analyses adjusting estimates of perceived risk for age, sex, race, education, and employment status (September 2017 amendment MR0000096). The lack of understanding of the addiction risks across different population groups is further evidence of the likelihood of consumer misperception if appropriate warning language is not included on the products.

Figure 11: Perceived Addiction Risk for Heatsticks Pack Arm 2 Source: Figure 11-8: THS PMTA-05-NOC Report v.1.0

These findings raise concerns because they indicate that consumers, including young adult never smokers, do not fully comprehend the addiction risk of IQOS based on the currently proposed labeling, which does not include any information about nicotine or addiction. Of further concern is that consumers, including young adult never smokers, who mistakenly believe IQOS to be less addictive, may start using it when they would not have otherwise initiated tobacco use.

To mitigate the potential for consumer misperception of the addiction risk of IQOS, I recommend the inclusion of the following warning on all IQOS Heatsticks labels and in all IQOS advertising: “WARNING: This product contains nicotine. Nicotine is an addictive chemical.” Smokers exposed to tobacco product warnings generally report greater knowledge of the risks associated with use of the products. Evidence indicates warnings that are larger and more comprehensive are more effective in communicating the health risks of smoking.89

I further note that pursuant to deeming rule, all ENDS that are made or derived from tobacco are generally required to bear the warning statement: “WARNING: This product contains nicotine. Nicotine is an addictive chemical.” As FDA previously explained, this warning is necessary given consumers’ erroneous and

89 Hammond D, Fong, G, McNeill A, Borland R, Cummings K; Effectiveness of cigarette warning labels in informing smokers about the risks of smoking: findings from the International Tobacco Control Four Country Survey. Tobacco Control. 2006; 15(Suppl III): iii19- iii25.

88 unsubstantiated beliefs that tobacco products other than conventional cigarettes are either less addictive than cigarettes or not addictive at all. See 79 FR 23141 at 23166. Numerous studies demonstrate that consumers tend to perceive IQOS as similar to e-cigarettes in terms of risk, including addiction risk in particular (this trend was demonstrated in PBA-05-NOC, PBA-05-RRC, PBA-05-RRC2, and PBA-05-REC). The absence of the addiction risk warning on IQOS, when e-cigarettes generally must bear such a warning, could reinforce existing false beliefs about the addiction risk of IQOS as compared to conventional cigarettes.

In conclusion, the lack of a nicotine addiction warning on IQOS labels and advertising raises significant concerns because study data show that in the absence of such a warning, consumers, including young adult never smokers, hold erroneous beliefs about the addiction risk of IQOS, particularly the relative addiction risk of IQOS compared to combusted cigarettes. This, in turn, could have negative consequences for public health in terms of increased tobacco use initiation among nonusers and decreased cessation among users. Accordingly, in order to find the marketing of the products appropriate for the protection of the public health, I recommend the following changes to the product labels and advertising for IQOS:

Inclusion of the warning: “WARNING: This product contains nicotine. Nicotine is an addictive chemical.” on all IQOS Heatstick (and Heatstick-containing kit) package labels and in all advertisements.

I recommend that the warning be subject to the same format requirements as those currently required for the nicotine warning on covered ENDS products under the deeming rule. See 21 CFR 1143.3. This includes, among other requirements: (1) occupying at least 30% of each of the two principal display panels of every Heatstick (or Heatstick-containing kit) package, and (2) occupying at least 20% of the area of every print or other advertisement with a visual component (e.g., Internet web pages). I recommend these size requirements because, as explained in more details in the preamble to the deeming rule, users are more likely to notice, pay attention to, and recall warnings that are in a larger size and that appear on the front/major surfaces of packages. This in turn directly affects the likelihood that a consumer will understand and appreciate the risks being warned against. See 81 FR 28973, 28988-89. See also 79 FR 23141 at 23164- 65.

D. Carbon Monoxide (CO) Warning CO is a highly toxic gas produced by incomplete combustion of hydrocarbons. Common sources include motor vehicle exhaust gases and combustion appliances (e.g., heating units) in which partial combustion of oils, coal, wood, kerosene and other fuels generate CO. Patients with underlying cardiac conditions are at risk for death from arrhythmias and fatal heart attacks can occur; however, CO exposure can cause chest pain and increase the risk of cardiovascular injury independent of previous cardiac disease. Carboxyhemoglobin (COHb) is the most accurate method of assessing CO exposure in humans. Normal level for non-smokers is < 2% and for smokers is 5-13%.90 COHb levels will vary depending on duration and extent of CO exposure, ventilation, and underlying medical conditions.

As cigarettes, IQOS Heatsticks product packages and advertisements would be required to bear the rotating SG Warnings, one of which states, “SURGEON GENERAL’S WARNING: Cigarette Smoke Contains Carbon Monoxide.”

90 Prockop, Leon D, Chichkova, Rossitza I; Carbon monoxide intoxication: an updated review, Journal of the neurological sciences. 2007, Vol.262(1-2), p.122-130.

89

The applicant has provided multiple lines of evidence that although the use of Heatsticks in the IQOS device does produce CO, the exposure to CO from IQOS use is comparable to environmental exposure to CO. The applicant uses CO as one of their product specifications as shown in the table below.

Table 11: CO Acceptance Criteria for Heatsticks Heatstick Acceptance Criteria Batch Analysis Results (b) (4) Marlboro Heatstick Smooth Menthol Heatstick Fresh Menthol Heatstick Source: FDA created table based o

During inspection of the manufacturing facilities in Bologna, Italy, the results were found to meet specifications. In the mouse switching study, COHb levels for mice “switched” to IQOS, those that “quit,” and sham controls were ~5%. (Statistical analysis was not provided.) In the 90-day inhalation study conducted in rats, those inhaling IQOS had COHb levels in the same range as sham control.

In the clinical studies assessing exposure, baseline COHb levels ranged from 4.65-6.66%. By Day 5 across all four studies, COHb in participants who switched to IQOS fell to 1.06-2.48%. For the smoking abstinence group, COHb was 0.99-2.5%. By Day 90 of the extended REX studies, COHb levels in the IQOS arm were 2.66- 2.97% and in the abstinence arm levels were 2.84-3.04%.91 After accounting for standard deviations, the abstinence arm groups and the IQOS groups were identical.

In the 6-month ad libitum use study, baseline exhaled CO in parts per million (ppm) for the CC group was 23.6 and the THS group was 21.9. After 6 months, the CC group CO was 25.3 ppm compared to 17.2 ppm for the THS group. (Note: Exhaled CO in nonsmokers is generally <6ppm.92 The results of this study are consistent with a significant dual-use population (including in the THS arm) as described by the applicant. Despite the high number of dual-users and the highly variable use patterns, the exhaled CO level is decreased in the THS group.)

Based on the above evidence, although IQOS Heatsticks produce CO, the CO exposure is comparable to environmental CO exposure. Use of Marlboro, Smooth Menthol Heatsticks, and Fresh Menthol Heatsticks in the IQOS device does not pose any CO-related risks. Accordingly, the required CO warning is misleading with respect to IQOS products. This warning should not be required on IQOS packaging or advertising.

IV. Conclusions and Recommendations

In its applications for the IQOS THS with Marlboro Heatsticks, Smooth Menthol Heatsticks, and Fresh Menthol Heatsticks, the applicant provided detailed information for the manufacturing process for the Marlboro, Smooth Menthol, and Fresh Menthol Heatsticks, the Holder and the Charger. The provided information includes adequate process controls and quality assurance procedures to help ensure the three Heatstick products are manufactured consistently to meet the applicant’s specifications. To verify chemical and physical data, confirmatory testing was conducted at FDA’s Southeast Tobacco Laboratory in October

91 CO levels would be expected to be higher in a community environment than in a confined laboratory setting. The levels of CO were not statistically different between THS and SA arms and were well within the range for normal environmental CO. 92 Sandberg, AnnSofi; Skold, CM, Grunewalk, J, Eklund, A, Wheelock, AM; Assessing recent smoking status by measuring exhaled carbon monoxide levels, PLoS One. 2011:6(12):e28864.

90

2017. Although there were some methodological differences between the applicant’s testing and the FDA testing, the results were similar. FDA also conducted a review of peer-reviewed literature describing chemical analysis of heated tobacco products. The information in published literature generally supported the data in the applications.

Product stability can be a concern for tobacco products as bacterial communities and constituents in tobacco products change as a function of storage time. The applicant provided complete stability testing data for all three Heatsticks over a period of (b) (4) of product storage and concluded a shelf life of (b) (4) (b) (4) is acceptable at(b) (4) ) and (b) (4) and a shelf life of(b) (4) is acceptable at (b) (4) for Marlboro, Smooth Menthol, and Fresh Menthol Heatsticks. The applicant has observed (b) (4) , but the changes are not related to product safety, performance, or type of Heatstick. The applicant has described additional testing for high humidity stability studies with plans for further microbiological testing if needed. However, even without that additional testing, the applicant has addressed factors that could affect microbial stability and provided adequate quality control information.

The applicant submitted part-by-part and sub-assembly details for the assembly and manufacturing processes for the Holder and Charger. A detailed summary of the testing method for the heating blade was provided; (b) (4) . (b) (4) The Holder and Charger contain microcontrollers and firmware (b) (4) . (b) (4)

.

The product is designed to use interchangeable batteries. The applicant provided the supplier manufacturing specifications, which are aligned with the product battery specifications for the Holder and the Charger. The applicant submitted battery samples for testing to Winchester Engineering and Analytical Center in September 2017. No individual data points were out of specification. FDA inspections of the applicant’s research and manufacturing sites were performed. Minor deviations were found during inspection of (b) (4) located in(b) (4) . During the inspection, the CTP OS subject matter expert observed that (b) (4) .

The toxicological assessment included measurement of HPHCs in the Marlboro, Smooth Menthol, and Fresh Menthol Heatstick aerosols and comparison to 3R4F reference cigarettes as well as comparison to the mean in the smoke of 31 CC. In the PMI-58 study, the 54 HPHCs measured in all three Heatstick aerosols were reduced were reduced by 54.4-99.9% on a per stick basis when compared to 3R4F smoke. Machine- generated nicotine yields were reduced 35.9-39.4%, but clinical data indicates human CC smokers and Heatstick users absorb similar amounts of nicotine. For 18 of these compounds, the applicant determined that yields in Heatstick aerosols were reduced by 40-99.8% when compared to the mean of 31 CC commercially available in the U.S. Side stream aerosol from all three Heatsticks does emit detectable levels of some HPHCs, but levels are significantly lower than emissions from CC. There are potentially concerning chemicals in the Heatstick aerosols. The applicant conducted a non-targeted differential screening assay, which found the three Heatstick aerosols contain higher levels of some chemicals than 3R4F smoke – four of these are possible or probable carcinogens and 15 others are possibly genotoxic. However, based on current

91

knowledge, the toxic exposures from all three Heatstick aerosols are reduced compared to CC, and many of the known HPHCs found in CC smoke are very low or undetectable in Heatstick aerosols.

The applicant conducted in vitro testing, including Ames assay, mouse lymphoma assay (MLA), and nuclear red uptake assay (NRU). Limitations of these assays, caused in part by methodological issues as noted by the reviewer, affect the conclusions that can be drawn from these in vitro tests. Notwithstanding such limitations, overall the in vitro studies show decreased cytotoxicity and mutagenicity from exposure to TPM and GVP of Marlboro, Smooth Menthol, and Fresh Menthol Heatsticks as compared to TPM and GVP of 3R4F cigarettes, which are consistent with expected results from aerosol containing the amount of HPHCs identified in the studies discussed above.

In vivo studies included two 90-day nose-only inhalation studies in rats, an 18-month carcinogenicity study in mice, a nicotine pharmacokinetic study in rats, systems toxicology studies with acute and repeated exposures to human organotypic tissues, and a mouse “switching” study. The 90-day inhalation studies showed changes from Heatstick aerosol exposures were not observed or much less severe than changes due to 3R4F. The interim report of an 18-month carcinogenicity study shows the incidence of neoplastic lesions to be higher in groups exposed to either Heatstick aerosol or CC when compared to sham control; however, in the final study report the applicant concludes this long-term study demonstrated no increase in lung cancer risk due to THS 2.2 aerosol exposure compared to sham group. Per the applicant, toxicity is limited to adaptive responses in the upper respiratory tract. As an inhaled tobacco product, IQOS may elicit an inflammatory response in the respiratory tract but this study provides no definitive information about carcinogenicity risk for humans.

The experimental approach taken in the organotypic studies included methods that are considered exploratory and have not been independently validated; hence, the usefulness of the data is limited. The 8- month switching/cessation study suggested that switching to Heatsticks after a short period of cigarette smoke exposure led to histopathological changes similar to smoking cessation; however, there were some design limitations that reduce reliability of these data.

After consideration of all the toxicological data presented, the demonstrated reductions in measured HPHC exposures and reduced histopathological changes indicate a possible relative benefit compared to CC for smokers who switch completely to IQOS. The toxicological profiles of the Marlboro, Smooth Menthol, and Fresh Menthol Heatsticks are identical except for the difference in the quantity of menthol added to the mentholated products. Although Marlboro, Smooth Menthol, and Fresh Menthol Heatstick aerosols contain chemicals which are different from those found in CC, and some of which may be toxic, the currently available information (discussed below) indicates the reduced exposures to the large number of HPHCs found in CC will likely result in reduced health risks for CC smokers who switch completely to IQOS. Reduced HPHC exposure also is beneficial for those who would be secondarily exposed to the aerosol as compared to environmental tobacco smoke.

To support the clinical evaluation of IQOS, the applicant provided four PK/PD studies, four reduced exposure studies, a summary of adverse events with an updated summary report submitted May 2018, review of published literature and post-marketing reports, and an actual use study which evaluated misuse of the products as well as overall use patterns in a “real world” environment.

The four single-use, randomized, 2-period, 4-sequence cross-over PK/PD studies assessed and compared the rate and extent of nicotine uptake in participants using THS 2.2 compared to smoking own-brand CC and nicotine replacement therapy products. Systemic nicotine exposure was similar after single and multiple

92

uses of IQOS and CC (both Marlboro and mentholated Heatsticks). THS 2.2 provides sufficient nicotine to produce user satisfaction. Self-report questionnaires found that THS 2.2 produced reinforcing effects close to those of CC. Results from the PK studies and the population model submitted account for the variability in nicotine PK across multiple factors including weight, CYP2A6 activity, sex, and race. Based on the study results, nicotine PK in smokers who switch to IQOS is similar to those who continued to smoke CC. The data indicate THS 2.2 has addictive potential and abuse liability similar to CC which means that while IQOS can provide an adequate nicotine source for dependent populations, there is a risk of developing addiction for non-tobacco users who begin using IQOS.

Four randomized, controlled, open-label, 3-arm parallel group studies (reduced exposure or REX studies) were conducted with the primary aim to investigate systemic exposure to BOE in smokers who switched to THS 2.2, continued to smoke CC, or abstained from smoking (SA) over a 5-day confinement period. Two of these studies (ZRHM-REXA-07-JP and ZRHM-REXA-08-US) had an 85-day ambulatory phase extension after the 5-day confinement period for a total study duration of 90 days. The BOE selected correspond with 14 HPHCs and two additional moieties found in cigarette smoke or filler. At the end of the 5-day confinement period, systemic exposure to 15 of the 16 selected chemicals described above decreased by 47-96%. (Nicotine - also an HPHC - was also measured and levels were not decreased.) The reductions of systemic exposure to 15 measured chemicals seen after switching from CC to THS 2.2 in all REX studies were statistically significant. The reductions were statistically significant over 5 days and the decreases persisted through the 90-day period. These BOE reductions in those that completely switched to IQOS, indicate reduced HPHC exposures, and, although not demonstrated by the studies in the application, these reductions in exposure are likely to result in reduced risk of tobacco-related disease.

All REX studies included measurements of several BOPH (referred to by the applicant as clinical risk endpoints or CREs) as secondary or exploratory study endpoints to determine if THS 2.2 use resulted in biological changes that may indicate a change in long-term disease risk. The applicant selected these biomarkers based on changes shown in previous smoking cessation studies, as well as peer-reviewed literature on the association with health risks. After independent review of the literature, FDA concludes that while each of the six markers have data suggesting a relationship with one or more tobacco-related diseases, none were strong predictors of future health risks. Many of these endpoints are more appropriate for longer-term studies, as changes in these measures are expected to take months to years. Some BOPH had desirable change trends in THS 2.2 users compared to the CC arm, but only white blood cell (WBC) count and sICAM-1 demonstrated differences in the two 90-day studies for THS 2.2, CC, and SA arms. The BOPH measures were not significantly improved over the relatively short duration of these studies; however, the trends may be informative for understanding potential effects on biological processes such as inflammation and oxidative stress.

The applicant provided a cumulative safety summary with information from the eight completed clinical studies, two on-going clinical studies, premarket safety surveillance covering six market research studies, and one perception/behavior study, as well as post-market surveillance studies outside the U.S. Although the applicant determined that most of the reported AEs were unrelated to product use, THS 2.2 exposure cannot be ruled out as contributing to or exacerbating those AEs typically associated with tobacco exposure (e.g., cough, headache, syncope). Bioresearch Monitoring inspections of two clinical investigators were conducted and no major issues or clinically significant deviations were found that would compromise data validity and integrity.

Post-marketing AE reports about IQOS have been sparse, despite increasingly widespread international marketing since its commercial introduction in Japan and Italy in 2014. A Safety Update Report published in

93

April 2016 reported two serious AEs (nervous system disorders/syncope). An updated Safety Report was submitted in May 2018 for the period covering 1/1/2017 thru 12/31/2017. The report identified previously unrecognized short-term health risks associated with THS including hypersensitivity reactions, an accidental child exposure, and a reported weather-related (heat and humidity) “burning sensation.” The applicant reports that(b) (4) (b) (4) These improvements/modifications are expected to decrease the occurrence rate of AEs and are therefore consistent with the conclusion that short-term risks of IQOS use are no greater than those associated with CC.

A review of published clinical literature provided by the applicant found one case report of acute eosinophilic pneumonia in a young adult Japanese male after increasing his consumption of Heatsticks. This disease has a known association with tobacco products and is not unique to THS 2.2. FDA conducted an independent clinical literature review and found no additional clinical reports. No apparent signals of adverse experience or other concerns related to product design have been identified related to IQOS in the countries where it is currently marketed. The data available in the clinical studies and other submitted information do not identify specific health-related issues for IQOS use beyond the concerns of CC use.

The Actual Use study assessed self-reported misuse of THS 2.2. Of 985 participants, 47 (4.8%) reported using Heatsticks without the IQOS device; the majority (97.9%) lit the Heatstick like a CC, and one participant chewed the Heatstick on one occasion. The applicant evaluated the potential for consumers to attempt to re-use Heatsticks. When re-use is attempted, the Heatsticks deliver small amounts of aerosol with 17% of nicotine and 12% of TPM as a new Heatstick. The applicant did not provide additional data on consumer misuse of the Holder by attempting to use a combusted product (e.g., cigar, CC); however, heating a tobacco product will only generate an aerosol if there is enough of an “aerosol forming agent,” such as glycerin. In addition, the tobacco in any conventional product inserted into the IQOS Holder would be heated only to a maximum of 350 °C - the maximum temperature of the heating blade. This temperature is much lower than the combustion threshold of tobacco (>400 °C). Furthermore, only products with a circumference of 22.9 mm or less would fit inside an IQOS Holder, which excludes most conventional US cigarettes.

On June 11, 2018, FDA received amendment MR0000117 to the MRTPAs. This amendment included the final study report for a randomized, controlled, open-label, two-arm, parallel group, multi-center clinical study of six months of ad libitum use of the non-menthol THS 2.2 compared to continued CC users in an ambulatory setting in the U.S. The primary study objective was to demonstrate favorable changes after six months across eight co-primary clinical risk endpoints (referred to by FDA as BOPH) for those switching from CC to THS as compared to continued CC use. Secondary objectives included self-reported product use, nicotine exposure levels, and evaluation of exposure reduction to selected HPHCs by measuring BOE in the THS group and the dual-use group.

Five of the eight BOPH endpoints showed a statistically significant change in smokers who switched (defined as > 70% THS use) from CC smoking to THS use. All BOPH shifted in the same direction as when smokers quit, as described in literature. For the BOE, users who switched to THS (i.e., >70% THS use) had reduced levels for most measures. Users who met the applicants’ criteria for dual use (1%-70% THS use) also had reduced BOE for most measures but the changes were smaller and not statistically significant. None of the BOE measures increased with THS use – even in those who were “dual-users” in the study.

There are limitations to the applicant’s statistical approach that affect the reliability of the statistical conclusions of the study. However, the study does provide evidence of reduced exposures associated with switching completely from CC to THS. Additionally, there was a trend for BOE reduction in subjects who

94

dual-used CC and THS. The rate of self-reported dual use during this longer study was lower (~34%) than the considerably higher dual use rates in previous studies (~58% in the actual use study). The short-term and long-term effects of dual use remain unclear, but these data provide minimal evidence that short-term dual use of IQOS and cigarettes does not appear to increase exposures to the selected HPHCs.

Overall, the clinical studies show exclusive use of IQOS has potential for reduced adverse effects on individual health compared to CC smoking. Regular, Smooth Menthol, and Fresh Menthol Heatsticks provide nicotine at levels similar to CC which relieves nicotine cravings and withdrawal symptoms. The short (5-day) studies demonstrate improvement in BOE for complete switchers, which indicates reduced HPHC exposures. These trends in improvement persisted in the 90-day studies even though some participants had reduced compliance and were probably using other tobacco products in addition to IQOS. Although not demonstrated in the studies in the application, the reduced exposure to HPHCs is likely associated with reduced tobacco-related disease risk. There were some small (non-significant) improvements in BOPH that were in a positive direction. Although dual use was common in the U.S. studies, the clinical study submitted June 2018 showed less dual use over time (six months) and a trend (although not statistically significant) for improved measures of exposure to HPHCs. Product misuse is uncommon, and the product design makes misuse unsatisfying. The clinical studies and the literature searches did not identify specific short-term health-related issues uniquely associated with use of these products. The clinical studies did not demonstrate any difference in PK, PD, or adverse effects between the Marlboro, Smooth Menthol, and Fresh Menthol Heatsticks. The currently available evidence indicates CC smokers who switch completely to IQOS will have reduced toxic exposures and, although not demonstrated by the studies in the applications, consequently, are likely to have less risk of tobacco-related diseases. CC smokers who use IQOS while continuing to smoke (dual use) do not appear to experience increased HPHC exposures and the limited available information indicate they may also have reduced HPHC exposures.

The likelihood of IQOS use by current CC smokers was assessed in the perception study, the actual use study, and the WOT. The perception study, conducted in the U.S., indicated ~2/3 of current smokers expressed some interest in trying IQOS. The interest level increased to ~80-90% if IQOS was offered by a friend. Smokers expressed “intent to use IQOS regularly if they tried and liked it” and rates of ~55-70%. There was no significant difference in any of these scores for smokers intending to quit vs. those not intending to quit. Results from the actual use study, conducted in the U.S., and the WOT, conducted in five other countries where IQOS is currently marketed, were variable. Although the actual use and WOT studies are not generalizable to U.S. cigarette smokers, all study participants were CC smokers and the information gained from these studies provides useful trends for consideration in review of these applications for marketing in the U.S. In the U.S., 34% of cigarette smokers in the study initiated IQOS use, defined as using at least 100 Heatsticks. In the WOT, the prevalence of initiating IQOS use ranged from 36% in Italy to 76% in South Korea. Participants in both the actual use study and WOT had to express interest in using IQOS prior to study enrollment; however, the findings suggest that some smokers will find IQOS appealing and acceptable enough to initiate product use. In the U.S. study, daily cigarette consumption decreased between baseline and the observational period for all IQOS use groups, with the largest decrease occurring in participants who were predominant Heatstick users at Week 6 (average decrease of 7.6 CPD).

Dual use of IQOS and CC was common in all countries in the pre- and post-market studies. Among current smokers in the actual use study, a majority (57.6%) used the IQOS in addition to conventional cigarettes when dual use is defined as between 5% to 95% Heatsticks. The patterns of use overall are similar when considering the type of Heatstick ordered by the participant (Menthol, Regular, both); switching, dual use, and exclusive cigarette use did not differ by the type of Heatstick respondents requested at baseline. When ƵƐŝŶŐ ƚŚĞ ĂƉƉůŝĐĂŶƚ͛Ɛ ĚĞĨŝŶŝƚŝŽŶ ĨŽƌ ƐǁŝƚĐŚŝŶŐ ;ŝ͘Ğ͕͘ шϵϱй ,ĞĂƚƐƚŝĐŬ ƵƐĞͿ͕ ůĞƐƐ ƚŚĂŶ ϴй ŽĨ ƉĂƌƚŝĐŝƉĂŶƚƐ ŝŶ ƚŚĞ

95

actual use study met the criteria for switching from cigarettes to IQOS. Participants who became exclusive IQOS users, however, seemed less likely to return to using mostly CCs, indicated by the steady prevalence of exclusive IQOS use throughout the 6-week observational period. Although it is possible that with additional follow-up time more participants would become exclusive IQOS users, data from the actual use and the WOT studies show that most smokers become dual users during the initial period of IQOS use. This is a concern since there is limited evidence about the effects that dual use of IQOS and CC (compared to complete switching) will have on long-term reduction of HPHC exposures and the health risks for tobacco- related diseases. While results from the PBA-07 study showed that IQOS use was associated with reduction in cigarette consumption, the health benefits of reducing cigarette consumption instead of quitting completely are unclear. However, based on the currently available evidence, dual use is unlikely to pose increased health risks compared to continued exclusive CC use.

The applicant provided data from two Japanese on-line post-marketing surveys. In a 2016 Japanese online cross-sectional survey of 2000 adult smokers and nonsmokers, 3.7% of respondents reported using “heat- not-burn” (heated) tobacco products. The prevalence of heated tobacco product use was higher among those aged 20-39 (~4 %) than those aged > 40 (~1 - 1.5%) and most (96.3%) were using “Marlboro Heatsticks with IQOS device.” Among respondents currently using heated tobacco products, 84.9% also used CC, most of them daily. In the second Japanese marketing survey, data on self-reported use of IQOS and cigarettes were also collected from 14,999 adult IQOS purchasers who registered their device in an online market research database. (Since purchasing and registering IQOS were criteria of inclusion, this may not be a representative sample of all users.) The proportion of IQOS purchasers who were “exclusively” using IQOS ;шϵϱйͿ ŝŶĐƌĞĂƐĞĚ ĨƌŽŵ ϱϮй ŝŶ :ĂŶƵĂƌLJ ϮϬϭϲ ƚŽ ϲϱй ŝŶ :ƵůLJ ϮϬϭϲ͘

The U.S. perception study (PBS-05-NOC) assessed perceptions and intention to use IQOS among a sub- population of former smokers and never smokers, including a subgroup of young adult (aged 18-25 years) never smokers. In this study the applicant developed label/labeling/advertising (LLA) materials, including an IQOS brochure, Heatsticks pack, and direct mail communication. The LLA materials provided information intended by the applicant to distinguish IQOS from e-cigarettes, including statements about “real tobacco” and the similarity in appearance of IQOS Heatsticks and CC. Never smokers in this study, including young adults of legal age to 25 years, were only exposed to the Regular Heatsticks pack. This is a study limitation since menthol cigarette smokers comprise one-third of the U.S. market, and the study did not assess the response to menthol LLA materials in never smokers. Among never smokers and young adult never smokers, < 1% who viewed the LLA materials indicated they would Definitely or Very Likely use IQOS. The results for former smokers were slightly higher; of those who viewed LLA materials with no additional IQOS information, 5-6% indicated they would Definitely or Very Likely use IQOS though the positive intent to try IQOS was higher if ‘offered by a friend.’ For comparison, the applicant also asked former smokers about their intention to use e-cigarettes regularly and asked never smokers about their intention to try e- cigarettes. Former smokers’ intention to use and never smokers’ intention to try IQOS appeared to be similar to or somewhat lower than their intention to use or try e-cigarettes, although the applicant provided no statistical analysis of these differences. Hypothetical scenario studies with no actual consequences associated with the decisions are difficult to interpret. While these types of studies provide an indication of intent to try or use the product, they cannot be considered absolute signals of behavior when/if IQOS is a marketed product.

As noted above, the applicant provided results from internet market surveys conducted in Japan. During the first one to two years after IQOS went on the Japanese market in 2014, use by adult former and never smokers was low (1.5% among former smokers and 1.2% among never smokers). Among survey responders in 2017, there were 3.6% current IQOS users and 2.0% of those aged 15-19 years reported current use of

96

IQOS in 2017. Of the 2017 survey responders 1.3% were never smokers, 2.1% were former smokers, 18.8% were current smokers with intention to quit, and 10.3% were current smokers with no intention to quit. These results may not equate to the anticipated U.S. experience as e-cigarettes containing nicotine require a prescription in Japan and use patterns may differ in the U.S.

In 2017, a face-to-face survey ǁĂƐ ĐŽŶĚƵĐƚĞĚ ŝŶ /ƚĂůLJ ĂŵŽŶŐ ϯϬϬϬн ƉĂƌƚŝĐŝƉĂŶƚƐ ĂŐĞĚ шϭϱ LJĞĂƌƐ ƐĞůĞĐƚĞĚ from the general Italian population. Based on cigarette smoking status, 1.0% of never smokers, 0.8% of former smokers, and 3.1% of current smokers reported having ever tried IQOS. Among participants who reported that they had never tried IQOS but were intending to try it, 1.7% were never smokers, 0.5% were former smokers, and 5.0% were current smokers. In the Italian study, 0.9% of those aged 15-24 years reported having ever tried IQOS.

These Japanese and Italian studies suggest that the prevalence of IQOS use is lower in never and former smokers compared to current smokers and that fewer youth than adults currently use IQOS in Japan or Italy. These two published survey studies are the only data currently available on the prevalence of IQOS use in youth. The U.S. perception study suggests a low prevalence of intention to use IQOS among never and former smokers. In the U.S., most cigarette smokers begin trial and progression to regular use before age 18. Overall, the available information suggests the prevalence of IQOS use is lower in never smokers compared to current smokers and that fewer youth than adults currently use IQOS in Japan and Italy. Given that IQOS is still a relatively new product, the extent to which youth will initiate and use IQOS is unknown though the trend from other countries indicates that this is uncommon. The current evidence indicates IQOS uptake by youth and nonsmokers will be low. Furthermore, the limited flavor choices may reduce IQOS’ appeal to youth. The social science reviewers have concerns that data regarding IQOS use in youth are limited; however, it could be difficult and impracticable to obtain data that would satisfy the reviewers’ concerns in a pre-marketing environment.

Both the U.S. actual use and ex-U.S. WOT studies evaluated the likelihood of current cigarette smokers switching to IQOS. During the six-week observational period of the actual use study, 33.8% of current smokers initiated use of HeatƐƚŝĐŬƐ ;ĚĞĨŝŶĞĚ ĂƐ ĐŽŶƐƵŵŝŶŐ шϭϬϬ ,ĞĂƚƐƚŝĐŬƐͿ͘ ŵŽŶŐ ƚŚŽƐĞ ǁŚŽ ƐƚĂƌƚĞĚ ƵƐŝŶŐ ,ĞĂƚƐƚŝĐŬƐ͕ ϭϲ͘ϯй ǁĞƌĞ ĞdžĐůƵƐŝǀĞůLJ ƵƐŝŶŐ ,ĞĂƚƐƚŝĐŬƐ ;шϵϱй ,ĞĂƚƐƚŝĐŬ ƵƐĞͿ ĚƵƌŝŶŐ tĞĞŬ ϲ͘ /Ŷ ƚŚĞ tKd͕ exclusive Heatstick use among those who had used at least 100 Heatsticks ranged from 7.8% in Switzerland to 21.5% in Japan. Exclusive and predominant IQOS use was most common in Asia. More IQOS users may quit smoking over time, but data from the actual use and WOT studies suggest that most smokers become dual users or at least go through a “dual use” phase before quitting. Although less than 10% of cigarette smokers in the U.S. actual use study switched to exclusive IQOS use during the study, the proportion of exclusive IQOS users remained steady during the 6-week observational period. This suggests that individuals who initiate IQOS and use Heatsticks for at least 95% of their tobacco intake are able to maintain exclusive IQOS use over time and potentially replace their use of CC with Heatsticks long-term. The toxicological and clinical studies did not demonstrate an increase in HPHCs for users consuming IQOS and CC and, although not statistically significant, some HPHC exposures appear to decrease.

The applicant presented results from a Population Health Impact Model to assess the possible effects of the proposed new products on population health in the U.S. This is a computational and simulation model that tracks tobacco prevalence and deaths from four specific smoking-related diseases: lung cancer (LC), ischemic heart disease (IHD), stroke, and COPD on a hypothetical population exposed to two tobacco products - CC and THS 2.2. The applicant concludes that “introducing THS into the US population appears to lead to a sizeable public health benefit in terms of reduced cigarette smoking and tobacco-related mortality.” FDA reviewers found no concerns with the statistical and computational aspects. However, there are limitations

97

to the modeling assumptions, e.g., the model only considers two products (cigarettes and IQOS), the population size does not change over time, there is no justification for the assumption that nonsmokers will not use IQOS, and the 20-year projection is relatively short for evaluating long-term health effects. Although the model is statistically valid, the overall analysis of the population model does not provide evidence to support the application.

The applicant provided sample labeling materials for the Heatstick packs, cartons, and the IQOS device package and printed film as well as the IQOS Tobacco Heating System User Guide and the IQOS Quick Start Guide. Apart from the warning information (discussed in more detail in section F.1), none of these materials raised concerns. The applicant conducted a study to evaluate the ability of prospective consumers to correctly understand and comply with THS Instructions for Use. Generally, most study participants were able to follow the instructions though there were some challenges with the cleaning instructions. The applicant notes these performances were scored based on participants’ first attempt at the use tasks; thus, results may not be indicative of performance after users are familiarized with the product through repeated attempts. The applicant has stated (b) (4) The additional support along with the instructions that are included with the IQOS device should resolve most consumer issues related to product use. Overall, the results demonstrate sufficient consumer understanding of the products and their use.

At request of FDA, the applicant provided a summary of their marketing plan for the PMTAs in the U.S. The applicant plans to (b) (4) The applicant states (b) (4)

. The applicant states they(b) (4) .

A. Recommendation for Marketing As discussed in Sections III C, III D, and IV F of this review, I recommend the PMTAs be authorized subject to the following changes to the proposed product labeling and advertising for IQOS: x Inclusion of the warning: “WARNING: This product contains nicotine. Nicotine is an addictive chemical.” on the package labels of all Heatsticks packs and of all kits containing Heatsticks packs as well as in all advertisements for such products and kits. Data shows that consumers do not accurately perceive the addiction risks of IQOS. Permitting IQOS to be marketed without this warning would not be appropriate for protection of public health. x Removal of the warning: “SURGEON GENERAL’S WARNING: Cigarette Smoke Contains Carbon Monoxide.” from the required warnings to be displayed on the product package labels and advertisements under FCLAA. Based on a fair evaluation of all material facts, the warning is misleading with respect to these products which, although categorized as cigarettes, do not produce carbon monoxide above environmental levels and do not increase CO-related health risks.

None of the grounds specified in Section 910(c)(2) of the FD&C Act apply. Specifically, I find the following: 1. Permitting the marketing of the products is appropriate for the protection of the public health, as described in Section 910(c)(4) of the FD&C Act (subject to the labeling and advertising changes described above);

98

2. The methods used in, and the facilities or controls used for, the manufacture, processing, and packing of these products do not fail to conform to the requirements in 906(e);93 3. Based on a fair evaluation of all material facts, the labeling (when subject to the changes described above) is not false or misleading in any particular; and 4. The products do not fail to conform to a tobacco product standard in effect under Section 907 of the FD&C Act.

I recommend FDA grant marketing authorization for the products described in the STNs, subject to the changes to the products’ package labels and advertisements, as described above: 1. PM0000424: Marlboro Heatsticks 2. PM0000425: Marlboro Smooth Menthol Heatsticks 3. PM0000426: Marlboro Fresh Menthol Heatsticks 4. PM0000479: IQOS Holder and Charger

B. Postmarketing Recommendations The applicant submitted information on the stability monitoring protocol that it intends to use post- approval of the new products. The applicant proposes to test the Heatsticks at(b) (4) (b) (4) (b) (4) ) over a period of (b) (4) of product shelf life (b) (4) (b) The applicant states that this storage condition was selected because of its much higher (4)geographical relevance for the U.S. market and because the product is reasonably expected to be exposed to this condition. CTP recommends that the applicant adopt this post-approval stability protocol for Heatsticks.

C. Postmarketing Recordkeeping, Retention, Reporting and Marketing Requirements The following language will be included in the marketing authorization:

The Food and Drug Administration (FDA) completed the review of your Premarket Tobacco Product Applications (PMTAs) submitted under section 910(b) of the Federal Food, Drug, and Cosmetic Act (FD&C Act), specified in Appendix A.

Based on our review of your PMTAs, we find that the marketing of the new tobacco products specified in Appendix A is appropriate for the protection of public health, and that you have met the other requirements of section 910(c) of the FD&C Act. This marketing order is subject to marketing requirements under section 910(c)(1)(B) of the FD&C Act and record retention and reporting requirements under section 910(f) of the FD&C Act, as outlined in Appendix B. Additionally, this order is conditioned upon the products conforming with any applicable current or future tobacco product standards, unless specifically exempted under this order or the product standard(s). Under the provisions of section 910, you may introduce or deliver for introduction into interstate commerce the new tobacco products, in accordance with the order requirements outlined in Appendix B.

The requirements in this order are intended to help ensure that the marketing of your products will continue to be appropriate for the protection of the public health, taking into account initiation among non- users, particularly youth. However, compliance with these requirements alone is not a guarantee that the marketing of the products will remain appropriate for the protection of the public health, particularly if, despite these measures, there is a significant uptake in youth initiation, for example. FDA will continue to

93 FDA has not yet promulgated any regulations under Section 906(e) of the FD&C Act.

99

monitor the marketing of your products.

This order does not constitute a finding that any of the products outside the scope of this authorization are in compliance with the FD&C Act and its implementing regulations. FDA has not evaluated other components or parts, or accessories that you may choose to market with the iQOS system, such as A/C power adapters, USB cables, charging docks, cleaners, disposal units, and pouches. To the extent that any premarket authorization requirements of section 910 of the FD&C Act apply, FDA does not intend to enforce them with respect to such products. However, it is your responsibility to ensure that these products comply with all other applicable laws and regulations. For example, if you choose to include the brand name “IQOS” on items other than the products authorized in these orders, you need to evaluate whether that would comply with 21 CFR 1140.34(a). In addition, we recommend you evaluate whether any of the branded accessories you plan to market would constitute advertising that requires the applicable warnings.

We note that, in your September 5, 2018 and March 25, 2019 amendments to your PMTAs, you include representations about your marketing plan for your products in the United States and indicate that you intend to focus marketing on adult cigarette smokers while limiting reach to unintended audiences. FDA encourages you to consider measures to limit youth-exposure to any of the products’ labeling, advertising, marketing, and/or promotion appearing in print media publications. Limiting youth exposure and initiation and use of the products as you have indicated in your PMTAs (i.e., complete switching to IQOS by adult cigarette smokers) are important components of consideration for the marketing of these products to continue to be appropriate for protection of the public health.

Also, in accordance with 40 CFR 1506.6, we will make your environmental assessments publicly available.

This order authorizing the marketing of these new tobacco products does not mean FDA “approved” the new tobacco products specified in Appendix A; therefore, you may not make any express or implied statement or representation directed to consumers that conveys, or misleads or would mislead consumers into believing, among other things, that the new tobacco products specified in Appendix A are “approved” by FDA.94 The products subject to this marketing order are subject to withdrawal or temporary suspension as described in section 910(d) of the FD&C Act.

We remind you that all regulated tobacco products, including the new tobacco products specified in Appendix A, are subject to the requirements of the FD&C Act and its implementing regulations. These requirements currently include, but are not limited to, annual registration, listing of products, listing of ingredients, reporting of harmful and potentially harmful constituents, and payment of user fees. There are also packaging, labeling, and advertising requirements with which you must comply. It is your responsibility to ensure the tobacco products specified in Appendix A comply with all applicable statutory and regulatory requirements. FDA will monitor your compliance with all applicable statutes and regulations.

1. Record Retention

Under section 910(f) of the FD&C Act, this order requires that you establish and maintain the records listed below. The records must be retained for a period of not less than four years from the date of distribution of the last batch of the new tobacco products listed in your marketing authorization. The records must be

94 See Section 301(tt) of the FD&C Act.

100

legible, written in English, and available for inspection and copying by officers or employees duly designated by the Secretary upon request:

x The PMTA submitted prior to product order x Periodic postmarket reports, as described below, and adverse experience reports, including all relevant documentation associated with the experience x Records of all nonclinical or clinical studies, including: o Source data; o Study protocols (including statistical analysis plan); o Amendments showing the dates and reasons for any protocol revisions; o Institutional Review Board (IRB) or Independent Ethics Committee (IEC) approvals or non- approvals; o Informed consent forms; o Correspondence with study monitors/investigators/contract research organizations/sponsors/IRB/IEC; o Investigator financial disclosure statements; o Progress reports; o Monitoring reports; o Adverse experience reports; o Case report forms/subject diaries/medical records/laboratory reports; o Subject data line listings/observation records; o Test article accountability records; o Study results/protocol summaries/study reports; and o Certifications and amendments to certifications x Records pertaining to the manufacture, in process and release testing, production process (including any changes to the process, facility, or controls), packaging, storage, and stability monitoring and testing (including protocol and results) of the products x Records pertaining to the sale, distribution, or other disposition of the products, specifically: o A list of distributors and retailers of the products, including brick-and-mortar and digital95; o Any available information (not to include personally identifiable information) about product purchases, such as purchasers’ demographics (e.g., age, gender, race/ethnicity, geographic region) and previous or current use of other tobacco products (i.e., dual use); o Policies and procedures regarding verification of the age and identity of purchasers of the products; and o Policies and procedures regarding restrictions on youth access to the products x Records pertaining to the products’ labeling, advertising, marketing, and/or promotion – whether conducted by you, on your behalf, or at your direction – including: o Specimens of all labeling, labels, inserts/onserts, instructions, and other accompanying information; o Copies of all advertising, marketing, and/or promotional materials published, disseminated to consumers, or for use in engaging or communicating with consumers; o Copies of any formative research studies conducted among any audiences in the formation of the labeling, advertising, marketing, and/or promotional materials, including qualitative and quantitative research studies used to determine message effectiveness, consumer knowledge, attitudes, beliefs, intentions, and behaviors toward using the products, and including copies of the stimuli used in testing;

95 For the purposes of this order, here and throughout the document, “digital” includes internet/online and mobile.

101

o Copies of any consumer evaluation research studies conducted among any audiences to determine the effectiveness of labeling, advertising, marketing, and/or promotional materials and any shifts in consumer knowledge, attitudes, beliefs, intentions, and behaviors toward using the products, and including copies of the stimuli used in testing; o Copies of any contractual agreements regarding the creation and/or dissemination of the products’ labeling, advertising, marketing, and/or promotional materials; o Copies of all advertising and marketing plans, including strategic creative briefs and paid media plans, by channel and by product, and the dollar amount(s) and flighting of such plans, by channel and by product, including any: ƒ Use of competent and reliable data sources, methodologies, and technologies to establish, maintain, and monitor highly targeted advertising and marketing plans and media buys; ƒ Targeting of specific adult audiences by age-range(s), including young adult audiences, ages 18-24, and other demographic and/or psychographic characteristics that reflect your intended target audience; ƒ Actions taken to restrict youth-access and limit youth-exposure to the products’ labeling, advertising, marketing, and/or promotion; ƒ Use of owned, earned, shared, and/or paid social media to create labeling for, advertise, market, and/or promote the products; ƒ Use of partners, influencers, bloggers, and/or brand ambassadors to create labeling for, advertise, market, and/or promote the products; ƒ Consumer engagements – whether conducted by you, on your behalf, or at your direction – including events at which the products were demonstrated; and/or ƒ Use of earned media and/or public-relations outreach to create labeling for, advertise, market, and/or promote the products o Copies of all records pertaining to media tracking and optimization, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic region), and all post- launch delivery-verification reports submitted to you from an accredited source, by channel, by product, and by audience demographics; and o Policies and procedures for real-time digital media monitoring to identify, correct, and prevent any delivery of advertising impressions to youth, ages 17 years and under, including documentation of such monitoring activities and implementation of corrective and preventive measures x Health hazard analyses, if performed voluntarily or directed by FDA x Records pertaining to any and all complaints associated with any of the products that you receive or of which you are aware

2. Serious and Unexpected Adverse Experiences Reporting

Under section 910(f) of the FD&C Act, this order requires that you report to the FDA all adverse experiences that are both serious and unexpected and your analysis of the association between the adverse experience and the tobacco product(s) within 15 calendar days after the report is received by you. These experiences may become known to you through any source including a customer complaint, request, or suggestion made as a result of an adverse experience, tobacco product defect, or failure, reported to you, or identified in the literature or media. Your information should be submitted with a cover letter that includes the following text in the subject line: SERIOUS UNEXPECTED ADVERSE EXPERIENCE REPORT FOR STN(s) XXX.

102

For purposes of reporting under this order, serious adverse experience means an adverse experience that results in any of the following outcomes: x Death; x A life-threatening condition or illness; x Inpatient hospitalization or prolongation of existing hospitalization; x A persistent or significant incapacity or substantial disruption of the ability to conduct normal life functions; x A congenital anomaly/birth defect; or x Any other adverse experience that, based upon appropriate medical judgment, may jeopardize the health of a person and may require medical or surgical intervention to prevent one of the other outcomes listed in this definition.

For purposes of reporting under this order, unexpected adverse experience means an adverse experience occurring in one or more persons in which the nature, severity, or frequency of the experience is not consistent with: x The known or foreseeable risks associated with the use or exposure to the tobacco product as described in the PMTA (including the results of human subject investigations) and other relevant sources of information, such as product labeling and postmarket reports; x The expected natural progression of any underlying disease, disorder, or condition of the person(s) experiencing the adverse experience and the person’s predisposing risk factor profile for the adverse experience; or x The results of nonclinical laboratory studies.

3. Manufacturing Deviations

Under section 910(f) of the FD&C Act, this order requires that you establish and maintain records and reports of all manufacturing deviations, investigations, and corrective and preventive actions including, but not limited to, those deviations associated with processing, testing, packing, labeling, storage, holding, and distribution. For products that have been distributed, if a deviation occurs that you determine presents a reasonable probability that the tobacco product contains a manufacturing or other defect not ordinarily contained in tobacco products on the market that would cause serious, adverse health consequences or death you are required to report the deviation to FDA within 15 calendar days of identification.

4. Periodic Reporting

The information in the following postmarket periodic reports will help FDA determine whether continued marketing of your tobacco products is appropriate for the protection of public health and/or there are or may be other grounds for withdrawing or temporarily suspending the marketing authorization order.

Under section 910(f) of the FD&C Act, this order requires that you submit the following periodic reports to FDA on a quarterly basis, for a period of two years, beginning three months from the date of this order. For each three-month reporting period, these periodic reports must include: x A cover letter that includes the following text in your subject line: PERIODIC REPORT for PM0000424-PM0000426, PM0000479. The cover letter should include the STN(s) and corresponding tobacco product name(s), firm name, date of report, reporting period. x A summary of U.S. sales and distribution of the tobacco products, including total U.S. sales reported

103

in dollars, units, and volume, and broken down by U.S. census region, major retail markets, and channels where the products are sold (e.g., convenience stores, food and drug markets, big box retailers, digital platforms, tobacco specialty shops, company-owned stores). This summary must also be broken down by product (e.g., specific HeatStick flavor). x Data on product purchasers. Report any data collected about new purchasers, those who have switched tobacco products, and/or multiple product users. The results must be broken down by purchaser demographics (e.g., age, gender, race/ethnicity, geographic location) and must not include personally identifiable information.

Under section 910(f) of the FD&C Act, this order also requires that you also submit periodic reports to FDA on a quarterly basis, beginning three months from the date of this letter. For each three-month reporting period, these periodic reports must include: x A cover letter that includes the following text in your subject line: PERIODIC REPORT for PM0000424-PM0000426, PM0000479. The cover letter should include the STN(s) and corresponding tobacco product name(s), firm name, date of report, reporting period. x An analysis of the actual delivery of advertising impressions, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a breakout by age-group (i.e., adults, ages 25+; young adults, ages 18-24; and youth, ages 12-17 and ages 11 and under). This analysis must be verified against post-launch delivery-verification reports submitted to you from an accredited source. x A summary of media tracking and optimization, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a summary of real- time digital media monitoring to identify, correct, and prevent delivery of advertising impressions to youth, ages 17 and under, and including a summary of implementation of any corrective and preventive measures.

Under section 910(f) of the FD&C Act, this order also requires that you submit the following periodic reports to FDA on an annual basis, beginning twelve months from the date of this order. For each twelve-month reporting period, these periodic reports must include: x A cover letter that includes the following text in your subject line: ANNUAL REPORT for PM0000424- PM0000426, PM0000479. The cover letter should include the STN(s) and corresponding tobacco product name(s), firm name, date of report, reporting period. x A summary of how the marketing of the tobacco products continues to be appropriate for the protection of public health, which includes: o A status report of ongoing studies and a summary of completed studies about the tobacco products conducted by you, or on your behalf. o A summary of significant findings in publications not previously reported and full copies of the articles. This must include any new scientific data (published or otherwise) on the likelihood of product use by current users of tobacco products within the same tobacco product category, current users of tobacco products in other tobacco product categories, former users of any tobacco product, and youth and young adults. o A summary of reported adverse experiences for the tobacco products, which includes a listing of all adverse experiences, including the serious and unexpected adverse experiences previously reported. The summary must be accompanied by an analysis of the reports and a statement of any changes to risk information related to the products including nature, frequency, and potential aggravating factors. o A summary of U.S. sales and distribution of the tobacco products, not previously submitted,

104

including total U.S. sales reported in dollars, units, and volume, and broken down by U.S. census region, major retail markets, and channels where the products are sold (e.g., convenience stores, food and drug markets, big box retailers, digital platforms, tobacco specialty shops, company-owned stores). This summary must also be broken down by product (e.g., specific HeatStick flavor). o Data on product purchasers not previously submitted. Report any data collected about new purchasers, those who have switched tobacco products, and/or multiple product users. The results must be broken down by purchaser demographics (e.g., age, gender, race/ethnicity, geographic location) and must not include personally identifiable information. o A summary of the implementation and effectiveness of your policies and procedures regarding verification of the age and identity of purchasers of the products. o A summary of the implementation and effectiveness of your policies and procedures regarding restrictions on youth access to the products. o A description of each change made to the manufacturing process, facilities, or controls during the reporting period including: ƒ A comparison of each change to what was described in the PMTAs; ƒ The rationale for making each change; and ƒ A certification that the reported change did not result in any modification (including a change in design, any component, any part, or any constituent, including a smoke or aerosol constituent, or in the content, delivery, or form of nicotine, or any other additive or ingredient) of the tobacco products and the basis for concluding that each manufacturing change did not result in any modification to the products. o A summary of all manufacturing deviations, investigations, and corrective and preventive actions, including, but not limited to, those deviations associated with processing, testing, packing, labeling, storage, holding, and distribution and indicate any deviation(s) that may affect the characteristics of the products. o A summary of any stability monitoring and testing of the HeatSticks products, including monitoring and testing protocol (including batch/lot sampling) and results. o All final printed labeling (including all labeling variations, such as those reflecting different required warnings) not previously submitted, including the date the labeling was first disseminated and the date when the labeling was discontinued, and a description of all changes to the labeling. The labeling must include all the panels and be presented in the actual size and color with legible text. The labeling must include labels, inserts/onserts, instructions, and any other accompanying information or materials for the products. o All final full-color advertising, marketing, and/or promotional materials, published, disseminated to consumers, or for use in engaging or communicating with consumers not previously submitted, along with the original date such materials were first disseminated and the date they were discontinued, and a description of all changes to the materials. The materials must include all panels where applicable (e.g., print ads, point of sale signs) and reflect the actual size and colors used. For any materials that would not fit on an 8.5” x 11” piece of paper, you may resize and submit electronic versions of such materials in a format that FDA can review and with sufficient resolution to allow FDA to read lettering clearly. If resizing the advertisement does not allow for text to be read easily, the text may be provided separately and referenced. o A summary of all formative consumer research studies conducted – whether by you, on your behalf, or at your direction – among any audiences, in the formation of new labeling, advertising, marketing, and/or promotional materials, including qualitative and quantitative research studies used to determine message effectiveness, consumer knowledge, attitudes, beliefs, intentions and behaviors toward using the products, and including the findings of these

105

studies and copies of the stimuli used in testing. o A summary of all consumer evaluation research studies conducted – whether by you, on your behalf, or at your direction – among any audiences, to determine the effectiveness if labeling, advertising, marketing and/or promotional materials and any shifts in consumer knowledge, attitudes, beliefs, intentions, and behaviors toward using the products, and including the findings of these studies and copies of the stimuli used in testing. o A summary of the creation and dissemination of the products’ labeling, advertising, marketing, and/or promotional materials – whether conducted by you, on your behalf, or at your direction – including a list of all entities involved and a description of their involvement, including a description of contractual agreements with such entities. o A description of the implementation of all advertising and marketing plans, including strategic creative briefs and paid media plans – whether conducted by you, on your behalf, or at your direction – by channel and by product, and the dollar amount(s) and flighting of such plans, by channel and by product, including a description of any: ƒ Use of competent and reliable data sources, methodologies, and technologies to establish, maintain, and monitor highly targeted advertising and marketing plans and media buys; ƒ Targeting of specific adult audiences by age-range(s), including young adults, ages 18-24, and other demographic and/or psychographic characteristics that reflect the intended target audience, including a list of all data sources used to target advertising and marketing plans and media buys; ƒ Actions taken to restrict youth-access and limit youth-exposure to the products’ labeling, advertising, marketing, and/or promotion; ƒ Use of owned, earned, shared, and/or paid social media to create labeling for, advertise, market, and/or promote the products; ƒ Use of partners, influencers, bloggers, and/or brand ambassadors to create labeling for, advertise, market, and/or promote the products; ƒ Consumer engagements – whether conducted by you, on your behalf, or at your direction – including events at which the products were demonstrated; and/or ƒ Use of earned media and/or public-relations outreach to create labeling for, advertise, market, and/or promote the products; including the original date such plans were first used and the date they were discontinued, and a description of all changes to such plans since the last periodic report, by channel and by product. o An analysis of the actual delivery of advertising impressions, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a breakout by age-group (i.e., adults, ages 25+; young adults, ages 18-24; and youth, ages 12-17 and ages 11 and under), not previously submitted. This analysis must be verified against post- launch delivery-verification reports submitted to you from an accredited source. o A summary of media tracking and optimization, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a summary of real-time digital media monitoring to identify, correct, and prevent delivery of advertising impressions to youth, ages 17 and under, and including a summary of implementation of any corrective and preventive measures, not previously submitted.

Under sections 910(c)(1)(B) and 910(f) of the FD&C Act, this order also requires that you provide the following notifications to FDA. These notifications are not for pre-approval, but are required so that FDA can have timely access to your marketing plans and materials, and if needed, provide you advisory comments, including any concerns about their possible impact on youth appeal and tobacco use initiation and on the

106

finding that continued marketing of your products is appropriate for the protection of the public health. You may begin disseminating the materials 30 days after providing notification to FDA. x Provide FDA notification of all labeling, advertising, marketing, and/or promotional materials for which you have not previously provided notification, at least 30 days prior to the initial publication, dissemination to consumers, or use in engaging or communicating with consumers of such materials, and include in your notification: o Full-color copies of all such labeling, advertising, marketing, and/or promotional materials for the products. The materials must include all panels where applicable (e.g., print ads, point of sale signs) and reflect the actual size and colors used. For any materials that would not fit on an 8.5” x 11” piece of paper, you may resize and submit electronic versions of such materials in a format that FDA can review and with sufficient resolution to allow FDA to read lettering clearly. If resizing the advertisement does not allow for text to be read easily, the text may be provided separately and referenced. o All advertising and marketing plans, including strategic creative briefs and paid media plans, by channel and by product, and the dollar amount(s) and flighting of such plans, by channel and by product, including any plans to: ƒ Use competent and reliable data sources, methodologies, and technologies to establish, maintain, and monitor highly targeted advertising and marketing plans and media buys; ƒ Target specific adult audiences by age-range(s), including young adults, ages 18-24, and other demographic and/or psychographic characteristics that reflect your intended target audience, including a list of all data sources used to target advertising and marketing plans and media buys; ƒ Restrict youth-access and limit youth-exposure to the products’ labeling, advertising, marketing, and/or promotion; ƒ Use owned, earned, shared, and/or paid social media to create labeling for, advertise, market, and/or promote the products; ƒ Use partners, influencers, bloggers, and/or brand ambassadors to create labeling for, advertise, market, and/or promote the products; ƒ Conduct any consumer engagements – whether by you, on your behalf, or at your direction – including events at which the products will be demonstrated; and/or ƒ Use earned media and/or public-relations outreach to create labeling for, advertise, market, and/or promote the products.

5. Marketing Requirements

Under section 910(c)(1)(B) of the FD&C Act and in accordance with section 202(a) of the Family Smoking Prevention and Tobacco Control Act, this order requires:

x Inclusion of the warning statement: “WARNING: This product contains nicotine. Nicotine is an addictive chemical.” on the package labels of all HeatSticks packs and of all kits containing HeatSticks packs as well as in all advertisements for such products and kits.96 Specifically, the warning statement must appear directly on the package and must be clearly visible underneath any

96 This warning must appear on each package and each advertisement, in addition to the rotating Surgeon General warnings required under FCLAA (except the carbon monoxide warning, which is to be removed from the rotation of the Surgeon General warnings as described in this order). When FDA promulgates a final rule with respect to health warnings for cigarettes, FDA will reevaluate the conditions of marketing with respect to warnings for the products subject to this order.

107

cellophane or other clear wrapping as follows: o Be located in a conspicuous and prominent place on the two principal display panels of the package and the warning area must comprise at least 30 percent of each of the principal display panels; o Be printed in at least 12-point font size and the warning statement must occupy the greatest possible proportion of the warning area set aside for the required text; o Be printed in conspicuous and legible Helvetic bold or Arial bold type (or other sans serif fonts) and in black text on a white background or white text on a black background in a manner that contrasts by typography, layout, or color, with all other printed material on the package; o Be capitalized and punctuated as indicated in this order; and o Be centered in the warning area in which the text is required to be printed and positioned such that the text of the warning statement and the other information on the principal display panel have the same orientation. x For print advertisements and other advertisements with a visual component (including, for example, advertisements on signs, shelf-talkers, websites, mobile applications, and e-mail), the warning statement must appear in the upper portion of the area of the advertisement within the trim area as follows: o Occupy at least 20 percent of the area of the advertisement; o Appear in at least 12-point font size and the warning statement must occupy the greatest possible proportion of the warning area set aside for the required text; o Appear in conspicuous and legible Helvetica bold or Arial bold type (or other similar sans serif fonts) and in black text on a white background or white text on a black background in a manner that contrasts by typography, layout, or color, with all other material on the advertisement; o Be capitalized and punctuated as indicated in this order; o Be centered in the warning area in which the text is required to appear and positioned such that the text of the warning statement and the other textual information in the advertisement have the same orientation; and o Be surrounded by a rectangular border that is the same color as the text of the warning statement and that is not less than 3 millimeters (mm) or more than 4 mm. x Removal of the warning: “SURGEON GENERAL’S WARNING: Cigarette Smoke Contains Carbon Monoxide.” from the required warnings to be displayed on the product package labels and advertisements under the Federal Cigarette, Labeling and Advertising Act (FCLAA). x As a reminder, under section 4 of FCLAA (15 U.S.C. 1333), you must submit a warning plan to the United States Federal Trade Commission (FTC).

Under section 910(c)(1)(B) of the FD&C Act, this order requires you to:

x For any digital sales – whether conducted by you, on your behalf, or at your direction – establish, maintain, and monitor use of independent age- and identity-verification service(s) that compare customer information against independent, competent, and reliable data sources, such as public records, to prevent the sale of the products to individuals who are under the federal minimum legal age to purchase tobacco products. x For any of the products’ labeling, advertising, marketing, and/or promotion appearing in your owned digital properties (e.g., your company-owned, consumer-directed, product-branded website(s) and/or mobile applications) – whether conducted by you, on your behalf, or at your direction – establish, maintain, and monitor use of independent age- and identity-verification service(s) that compare consumer information against independent, competent, and reliable data sources, such as public records, at the first point of access to such properties, to restrict access to

108

such labeling, advertising, marketing, and/or promotion to only individuals who are at least of federal minimum legal age to purchase tobacco products. x For any of the products’ labeling, advertising, marketing, and/or promotion appearing in any shared digital properties (e.g., your product-branded social media accounts, pages and associated content; content promoting your products on your behalf disseminated through another entity’s social media accounts) – whether conducted by you, on your behalf, or at your direction – establish, maintain, and monitor use of the available site-, platform- and content- (e.g., post, video) specific age- restriction controls (e.g., age-restrict an entire product-branded account and all associated content disseminated through such account; ensure age-restriction of a specific video disseminated by an influencer promoting the products on your behalf through the influencer’s account), at the first point of access to such properties, to restrict access to such labeling, advertising, marketing, and/or promotion to only individuals who are at least of federal minimum legal age to purchase tobacco products. x For any of the products’ labeling, advertising, marketing, and/or promotion appearing in paid digital media (e.g., paid digital banner advertisements for the product(s) running on another company’s website; paid advertising for the product(s) running in social media; paid distribution of influencer content) – whether conducted by you, on your behalf, or at your direction: o Establish, maintain, and monitor use of competent and reliable data sources, methodologies, and technologies to precisely target delivery of such labeling, advertising, marketing, and/or promotion to only individuals who are at least of federal minimum legal age to purchase tobacco products. Such targeting must use only first- and/or second-party age-verified data, where: ƒ “First-party” age-verified data is data owned by you (e.g., your customer registration data collected via site traffic to your company-owned website; data you use in direct marketing to your adult smoking customers) that you have age-verified through independent, competent, and reliable data sources; and ƒ “Second-party” age-verified data is first-party data owned and age-verified by another competent and reliable entity (e.g., another company’s first-party user registration data) to which you have access. Such data must be age-verified by the second party. ƒ “First-party” and “second-party” data does not include data obtained from data aggregators who categorize consumers based on trackable activities and inferred interests (e.g., internet search terms, video interactions, browsing history, purchasing behaviors) to create demographic and psychographic profiles marketers may use to enhance audience targeting. Such data is not considered age-verified and can only be used in combination with first- and/or second-party age-verified data. x Establish, maintain, and monitor use of competent and reliable data sources, methodologies, and technologies (e.g., using an embedded tracking pixel in all digital advertising) – whether conducted by you, on your behalf, or at your direction – to track and measure actual delivery of all advertising impressions, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a breakout by age-group (i.e., adults, ages 25+; young adults, ages 18-24; and youth, ages 12-17 and ages 11 and under). Such monitoring requires real- time digital media tracking, and identifying, correcting, and preventing delivery of advertising impressions to youth, ages 17 and under. Such monitoring also requires post-launch delivery verification reports be submitted to you from an accredited source. x For any use of partners, influencers, bloggers, and/or brand ambassadors to create labeling for, advertise, market, and/or promote the products – whether conducted by you, on your behalf, or at your direction – disclose to consumers or viewers, via the use of statements such as “sponsored by [firm name]” in such labeling, advertising, marketing, and/or promotional materials, any

109 relationships between you and entities that create labeling for, advertise, market, and/or promote the products, on your behalf, or at your direction.

110

Appendix: The Public Health Rationale for Recommended Restrictions on New Tobacco Product Labeling, Advertising, Marketing, and Promotion

I. Background

Most tobacco use is established in adolescence and age of initiation plays a significant role in the progression from tobacco experimentation to regular use (HHS 2012). It is well established that industry practices, such as tobacco product labeling, advertising, marketing and promotion, substantially impact youth trial and uptake of tobacco product use. Part of FDA’s premarket review under the PMTA pathway is aimed at determining if marketing a new tobacco product would increase or decrease the likelihood that those who do not currently use tobacco products, will start using them. Firms seeking a marketing order for a new tobacco product not yet on the market may not have robust data on how U.S. consumers will perceive the specific product, including its risks, or the degree to which its labeling, advertising, marketing, and promotion may influence youth perception or appeal to youth. This memo describes FDA’s authorities under the Family Smoking and Tobacco Control Act (Tobacco Control Act) to monitor and restrict tobacco product marketing and related activities in the context of premarket tobacco product application review and authorization. Given FDA’s statutory mandate to protect young people from the dangers of tobacco use and ensure that the marketing of new tobacco products is appropriate for the protection of the public health, the agency can request and review labeling, advertising, marketing, and promotional materials and plans for new tobacco products that have received premarket authorization to ensure that there are no grounds for withdrawing authorization and restrict the marketing of such products as appropriate for the protection of public health. This will help FDA evaluate the potential impact of such materials on the likelihood of initiation and use of the new tobacco products by youth or others and provide the firm and/or the agency an opportunity to prevent or mitigate any related potential harms to the public health.

II. The Food, Drug, and Cosmetic Act, as Amended by the Tobacco Control Act: Congressional Findings and FDA Authorities Related to Tobacco Product Labeling, Advertising, Marketing, and Promotion

The Tobacco Control Act makes clear the harmful influence of tobacco product labeling, advertising, marketing and promotion on youth tobacco use, and the intent of Congress to give FDA the authority to restrict these activities. In the Tobacco Control Act, Congress finds that, “[t]obacco advertising and marketing contribute significantly to the use of nicotine-containing tobacco products by adolescents,” and “[b]ecause past efforts to restrict advertising and marketing of tobacco products have failed adequately to curb tobacco use by adolescents, comprehensive restrictions on the sale, promotion, and distribution of such products are needed.” TCA §2(5) and (6). Thus, Congress concludes, “[c]omprehensive advertising restrictions will have a positive effect on the smoking rates of young people,” and “[r]estrictions on advertising are necessary to prevent unrestricted tobacco advertising from undermining legislation prohibiting access to young people and providing for education about tobacco use.” TCA §2(25) and (26). These findings are underscored by section 906(d) of the FD&C Act, which grants FDA the authority to “require restrictions on the sale and distribution of a tobacco product, including restrictions on the access to, and the advertising and promotion of, the tobacco product, if [...] such regulation would be appropriate for the protection of public health,” and section 910(a)(2) of the FD&C Act, which grants FDA the authority

111

to require premarket review and authorization of a new tobacco product before such product may be legally marketed in the United States. Further, as part of premarket application review, FDA may require “information relevant to the subject matter of the application” to assist the agency in determining “whether the marketing of a tobacco product […] is appropriate for the protection of public health” (section 910(b)(1)(G) and 910(c)(4) of the FD&C Act). In an order authorizing the marketing of a new tobacco product, FDA may also restrict the sale and distribution of the tobacco product to the extent that the sale and distribution of a tobacco product may be restricted under section 906(d) of the FD&C Act. FD&C Act §910(c)(1)(B).

III. Effects of Youth-Exposure to Tobacco Product Labeling, Advertising, Marketing, and Promotion on Youth-Appeal, -Perception, and -Use of Tobacco Products A. Influence of Tobacco Product Marketing on Youth Tobacco Use, in General As noted in the FD&C Act, as amended by the Tobacco Control Act, a key consideration in determining whether the marketing of a tobacco product is appropriate for the protection of public health is whether the marketing of the product would increase or decrease the likelihood that those who do not use tobacco products, especially youth, will start using them. In addition to Congress’ findings in the Tobacco Control Act, there is a large body of scientific evidence that documents the potential harm of tobacco product labeling, advertising, marketing and promotion on youth tobacco use. In one of the first comprehensive reviews on the subject—the National Cancer Institute’s (NCI) 19th monograph, The Role of the Media in Promoting and Reducing Tobacco Use—authors conclude that “the total weight of evidence—from multiple types of studies, conducted by investigators from different disciplines, and using data from many countries—demonstrates a causal relationship between tobacco advertising and promotion and increased tobacco use” (NCI 1998). As such, the direct role of tobacco product marketing and related activities in increasing tobacco use in the United States, especially among youth, and the high rates of youth-exposure to tobacco marketing due to its ubiquity, are two key rationales cited by NCI for restricting tobacco product marketing and related activities. The 2012 Surgeon General’s report, Preventing Tobacco Use Among Youth and Young Adults, synthesizes more than 30 years of research on the topic. This report outlines similar findings—tobacco product labeling, advertising, marketing, and promotion influence a wide range of established risk factors for youth tobacco use by shaping attitudes, beliefs, and risk perceptions, and promoting pro-tobacco social and cultural norms. The report states, “there is strong empirical evidence, along with the tobacco industry’s own internal documents and trial testimony, as well as widely accepted principles of advertising and marketing that support the conclusion that tobacco manufacturers’ advertising, marketing, and promotions recruit new users as youth and continue to reinforce use among young adults” (HHS 2012). This evidence is sufficient to conclude that “marketing efforts and promotion by tobacco companies show a consistent dose-response relationship in the initiation and progression of tobacco use among young people” (HHS 2012). To illustrate these points, the report cites findings of studies that demonstrate “advertising and promotion by the tobacco industry are effective in raising awareness of smoking, increasing brand recognition, and creating favorable beliefs regarding tobacco use. There is strong and consistent evidence that marketing influences adolescent smoking behavior, including selection of brands, initiation of smoking, and overall consumption of cigarettes” (HHS 2012). Further, “research conducted by the tobacco industry consistently demonstrates that the brand imagery portrayed on packages is particularly influential during youth and young adulthood—the period in which smoking behavior and brand preferences develop,” and “displays of packages in retail outlets, commonly referred to as ‘powerwalls,’ have high visibility among youth and help to establish brand imagery and social norms at an early age” (HHS 2012). “Young people who are more

112

familiar with tobacco advertising can identify specific advertisements, have a favorite tobacco advertisement, or possess cigarette promotional items are more likely to begin smoking than their peers who do not have these characteristics,” and “adolescents who both owned cigarette promotional items and had a favorite cigarette advertisement” were more likely to progress from initiation of smoking to established smoking (HHS 2012). Research has found that a key tactic of tobacco companies seeking to attract and recruit youth users is to use advertising with aspirational imagery and themes known to resonate with younger audiences, such as independence, popularity, rebelliousness, attractiveness, and being “cool” (HHS 2012). Even tobacco advertising that purportedly targets adult users can have a profound influence on adolescent tobacco use behaviors if it creates positive feelings in youth toward the product; pleasant feelings motivate actions that consumers anticipate will reproduce those feelings (Slovic & Peters 2006). As such, youth are more likely to mimic behavior portrayed as favorable in advertising, such as tobacco use. Furthermore, youth often misjudge the risks and benefits of advertised products based on how they feel about them (Slovic & Peters 2006). If youth feel positively toward a product, they are more likely to perceive it as having lower risks and higher benefits. In addition, adolescents are “uniquely susceptible to social and environmental influences to use tobacco” given their developmental stage and are heavily influenced by peers, family members, prominent members of their community, celebrities, and other cultural icons and adult role models—especially those they perceive to be popular, attractive, and “cool” (HHS 2012). As such, images of tobacco use in various types of media are “a potentially powerful socializing force among adolescents, in part because they are communicated by people who are identified by youth as media stars,” and “adolescents actively rely on external information as they seek to shape their own identities, often looking to media stars as models of what to wear and what to do” (HHS 2012). These marketing campaigns may be misleading in that they imply positive, pervasive and/or pro-tobacco social norms that are inaccurate or overstated. The misleading impression can be enhanced by failing to disclose a sponsor’s relationship with a company or failing to reveal that the content was not organically generated independently of the sponsoring company. Because youth have a heightened sensitivity to normative influences, sponsored tobacco marketing content may encourage youth uptake of tobacco use (HHS 2012). B. Influence of Tobacco Product Marketing on Youth Tobacco Use in the Context of Novel Tobacco Products Much of the research spanning the past few decades has focused on the influence of tobacco product marketing on cigarette smoking in particular; however, companies that sell other types of tobacco products engage in the same labeling, advertising, marketing, and promotional practices used by cigarette companies. “[T]he traditional division of products, brand identities, and marketing between cigarette and smokeless tobacco companies has all but become nonexistent in recent years as major U.S. cigarette companies, including RJR and Altria, have acquired smokeless tobacco companies and have developed new smokeless tobacco products” (HHS 2012). Some of these products are even marketed with popular cigarette brand names (e.g., Camel Snus).

Beyond cigarette-specific marketing, research has found that youth exposed to in-store marketing of e- cigarettes, hookah, cigars, smokeless tobacco, and pipe tobacco were two to three times more likely to use those products as well as to initiate cigarette use (Cruz et al. 2018). Further, research exploring the influence of tobacco marketing on youth use of novel tobacco products, such as e-cigarettes, confirms that exposure and receptivity to tobacco advertising is significantly associated with tobacco initiation among adolescents. The 2016 Surgeon General’s report, E-cigarette Use Among Youth and Young Adults, concluded “e-cigarette products are marketed in a wide variety of channels that have broad reach among youth and young adults,”

113

and themes in e-cigarette marketing are “parallel to the themes and techniques that have been found to be appealing to youth and young adults in conventional cigarette advertising and promotion” (HHS 2016). The report also summarizes the results of several studies looking at the relationship between e-cigarette marketing and youth tobacco use. For example, an analysis of the 2011 National Youth Tobacco Survey (NYTS) found that “adolescents who reported frequent exposure to protobacco advertising at the point of sale and on the Internet (e.g., seeing ads most of the time or always) had significantly higher odds of ever using e-cigarettes, and there was a dose-response association between the number of marketing channels to which they were exposed and ever use” (HHS 2016; Agaku & Ayo-Yusuf 2014). Two analyses of 2014 NYTS data assessing exposure to e-cigarette advertising in different channels (i.e., internet, print, television and movies, retail stores) found that “exposure to each type of e-cigarette marketing was significantly associated with increased likelihood of ever having used and current use of e-cigarettes among middle and high school students. Exposure was also associated with susceptibility to use e-cigarettes among current nonusers. In multivariate models, as the number of channels of e-cigarette marketing exposure increased, the likelihood of use and susceptibility also increased” (HHS 2016; CDC 2016; Mantey et al. 2016). These findings are particularly relevant in the context of more recent NYTS data showing a substantial increase in youth use of e-cigarettes from 2017 to 2018 (Gentzke et al. 2019). This uptick in youth e-cigarette use also contributed significantly to the first increase in overall youth tobacco use in recent years (Gentzke et al. 2019). Recent studies have also assessed the influence of e-cigarette marketing on youth use of conventional cigarettes. For example, an analysis of data collected between 2013-2015 via the Population Assessment of Tobacco and Health study found youth receptivity was highest for e-cigarette advertising (compared to conventional cigarette, cigar, and smokeless tobacco product advertising), and receptivity to e-cigarette advertising was also associated with initiation of conventional cigarette smoking (Pierce et al. 2018). Another study had similar findings concluding that exposure to any e-cigarette advertising may play a role in teens’ decision to initiate e-cigarette and conventional cigarette use (Padon, Lochbuehler, et al. 2017). These findings further underscore the powerful influence of tobacco product labeling, advertising, marketing, and promotion within and between product types, and the need for marketing restrictions for novel tobacco products. C. Influence of Digital Tobacco Marketing on Youth Tobacco Use While all tobacco product labeling, advertising, marketing, and promotion has the potential to significantly influence youth tobacco use, digital97 labeling, advertising, marketing, and promotion is particularly concerning given that it is transforming traditional marketing practices and is highly targeted to young people. The Pew Research Center reports that a vast majority of teens have access to a home computer or smartphone and nearly half of teens report using the internet “almost constantly” (2018), which means that many youth are constantly being exposed to marketing of a variety of different products, including tobacco products. While there is overwhelming evidence that children, teens, and young adults are exposed to and influenced by marketing of unhealthy products in traditional media, the internet provides marketers with new, relatively inexpensive channels and tools for disseminating their messages (Dunlap et al. 2016). Research examining online engagement with tobacco marketing among youth found a sizable increase of engagement over time (Soneji, Yang, Knutzen, et al. 2018) and that the number of engagements is associated with tobacco use initiation, frequency of use, and progression to poly-product use (Soneji, Yang, Moran, et al. 2018). According to the 2012 Surgeon General’s report, “the techniques of digital marketing are part of sophisticated behavioral targeting in which the marketer collects data on the users’ every move (e.g., every click of the mouse, sign-up for a contest, forwarding to a friend) to enable ever more precisely

97 For the purposes of this appendix, here and throughout the document, “digital” includes internet/online and mobile.

114

targeted marketing” (HHS 2012). This precision marketing also represents an opportunity to limit youth- exposure to the digital marketing of tobacco products. Via social media applications, marketers gain access to detailed profiles of users and their friends. Social media has fundamentally altered the marketing landscape by moving young audiences from passive recipients of advertising to active participants in the co-creation and dissemination of marketing messages (Dunlap et al. 2016). Corporate brands leverage the use of social media by adolescents and young adults to target and engage with young audiences (Dunlap et al. 2016). Unlike traditional forms of advertising that target potential customers with ads, companies that join in the “complex network of relations” of social media “befriend” their customers, which is a particularly appealing approach for companies wanting their consumers to express their personality through brand association (Dunlap et al. 2016). “Marketers seek to create ‘brand ambassadors,’ [i.e., social-media influencers] who promote the product in the context of their online communications, whether or not such promotions are recognized by the users or receivers as marketing. The effect is to blur the distinction between marketing communications and market research” (HHS 2012). For example, a study examining message content on Twitter concluded that Twitter serves as an important platform for e-cigarette marketing (Chu et al. 2015). Whenever a message posted by an e-cigarette brand is “retweeted” by another user, the message has reached a new network of users. Additional retweets can provide a cascading spread within and outside an original poster’s network and cause the message to go viral. This exposure through a retweeting network allows rapid diffusion of messages across groups (Chu et al. 2015). However, Twitter content often reaches unintended audiences, including youth and other vulnerable populations, due to the platform’s exponential reach and relatively limited control over what types of people are exposed to specific messages (Chu et al. 2015). With more than 30% of today’s youth reporting they use Twitter, marketing and promotion of tobacco products through Twitter can influence youth (Pew Research Center 2018). In addition, a recent study found that sales growth of was accompanied by a variety of innovative, engaging, and wide-reaching campaigns on social media platforms popular among youth, such as Twitter, Instagram, and YouTube (Huang et al. 2018).

IV. The Public Health Rationale for Requiring Submission of New Tobacco Product Labeling, Advertising, Marketing, and Promotional Materials and Plans and for Placing Restrictions on the Marketing of New Tobacco Products to Limit Youth-Exposure to Such Marketing A. Purpose of Marketing Requirements and Restrictions for New Tobacco Products, in General As noted in the introduction, FDA has a statutory mandate to ensure that the marketing of new tobacco products is appropriate for the protection of the public health. FDA’s premarket review under the PMTA pathway is aimed, in part, at determining if marketing a new tobacco product would increase or decrease the likelihood that those who do not currently use tobacco products will start using them. Among non-users, youth are a significant population of concern as their current stage of brain development makes them especially susceptible to nicotine addiction (HHS 2012). Prior sections of this memo have illuminated the powerful impact of tobacco product labeling, advertising, marketing, and promotion on youth-perceptions of tobacco products, youth-appeal of tobacco products, the likelihood of youth initiation and use of tobacco products, even when said marketing is purportedly targeted or designed to appeal to adults. Thus, for FDA to help ensure that the continued marketing of a new tobacco product is appropriate for the protection of public health, it is critical for FDA to conduct ongoing review and evaluation of the product’s labeling, advertising, marketing, and promotional materials and plans to assess any possible effects on perceptions, appeal, intentions, and behaviors among intended and unintended audiences, and to place appropriate restrictions on the product’s marketing and related activities from the outset to limit youth-exposure to such marketing.

115

Additionally, requiring a firm that receives marketing authorization for its products to provide labeling, advertising, marketing, and promotional materials and plans in advance of their use on an ongoing basis is not for pre-approval, but will provide FDA timely access to such materials and plans and, if needed, allow FDA to provide advisory comments, including any concerns about their possible impact on youth appeal and tobacco use initiation and on the finding that continued marketing of the products is appropriate for the protection of the public health. B. Reducing Youth-Appeal of Tobacco Product Marketing Generally, firms receiving marketing authorization for a new tobacco product should seek to reduce the youth-appeal of the tobacco product’s labeling, advertising, marketing, and promotional materials, including avoiding the use of imagery and themes known to resonate with youth, such as aspirational content depicting tobacco use as “cool,” attractive, rebellious, and/or risky, or as a means to make one more popular, desirable, or independent (HHS 2012). Other potential strategies for limiting youth-appeal of labeling, advertising, marketing, and promotional materials include focusing marketing content on instructional demonstrations and product comparisons and avoiding bright, bold, cheerful designs and colors, which can influence youths’ product choices because these characteristics affect their perception of the products, draw attention to them, and influence purchase decisions (Padon, Mahoney, et al. 2017; Akcay 2012; Lempert & Glantz 2016). Instead, labeling, advertising, marketing, and promotional materials should be clearly tailored to appeal to adults by using personalization strategies that make the content relevant and meaningful to adult recipients and should depict individuals who are similar to the target audience in terms of attributes, beliefs, and interests, in relatable situations that make it easier for adult viewers to engage with and connect to the advertising (Hawkins et al. 2008; Nielsen 2014). For example, advertising tailored to adult tobacco users would likely use headline and body copy that is relevant only to adults who might be considering switching products; would use models that are obviously older adults (ages 35-54) who look like and/or explicitly state they are tobacco users; and would portray people in realistic situations for tobacco users without making them look highly appealing or aspirational to other non-targeted populations, such as youth. C. Limiting Youth-Exposure to Tobacco Product Marketing Given the association between tobacco product marketing and youth initiation of tobacco use detailed in Section III, to help ensure the marketing of the products receiving marketing authorization under the PMTA pathway remains appropriate for the protection of public health, it is critical to limit youth-exposure to the products’ marketing, advertising, labeling, and promotion. Placing certain marketing restrictions98 on the newly authorized tobacco products from the outset, such as the media channels through which the firm markets its products, are essential components of limiting youth-exposure, and are thus appropriate for the protection of public health. 1. Restrictions on Paid Digital Tobacco Product Marketing The rise of digital marketing has changed media consumption habits over the past decade and created an increasingly complex media landscape where it is not yet possible to completely eliminate youth-exposure to tobacco marketing. However, the data sources, methodologies, and technologies used to deliver and track digital media consumption have also evolved, enabling product marketers to create sophisticated, highly targeted digital marketing plans and paid media buys designed to reach their intended audiences based on specific demographics, psychographics, and media passion-points while also limiting reach or

98 For the purposes of this appendix, this section focuses on restrictions related to digital tobacco product marketing. Considerations for other types of marketing restrictions may be addressed in the future, and the contents of this appendix should not be viewed as an exhaustive list.

116

Using targeting through the use of first- and/or second-party age-verified data (see table) does not mean that a firm will not be able to advertise at all in certain digital platforms, for example on certain websites that do not have age-restriction measures in place. Rather, even if a website does not have its own first- party age-verified data, tobacco advertising could still show up on that site. For example, if an adult that a tobacco marketer has age- and identity-verified as meeting the federal minimum legal age to purchase tobacco products through independent, competent, and reliable data sources visits TeenVogue.com, that adult could be delivered a tobacco ad on the site using the marketer’s first-party age-verified targeting data (regardless of whether TeenVogue.com has its own first-party age-verified data to share with the tobacco marketer), but an age-verified teen on TeenVogue.com would not be delivered the same tobacco ad as a result of this targeting. Therefore, through the use of targeting data, different individuals can see different ads when visiting the same website at the same time. This allows for a highly targeted approach to tobacco advertising delivery, which can help ensure that youth exposure is minimized, while at the same time not restricting access to adults. 2. Restrictions on Tobacco Product Social Media Marketing and the Use of Influencers, Bloggers, Brand Ambassadors, etc.

Although paid digital advertising can be effectively targeted using first- and second-party age-verified data to reach adults, there are other types of digital marketing cannot be targeted using this approach. For example, product-branded social media accounts essentially operate as both mini websites and “free” advertising channels offering a range of effective means of directly reaching and engaging consumers. In fact, “the ability to influence a large number of individuals, the minimal effort required to make influence attempts, and the flexibility to deploy a variety of influence strategies through information technologies are a potent combination making influence in online social networks considerably more compelling and pervasive than in conventional interpersonal interactions,” highlighting the need for close scrutiny of these methods (Subramani & Rajagopalan 2003). Further, one of the most effective digital marketing practices today—especially among youth who are particularly susceptible to social influences—is the use of “organic” depictions of tobacco use and endorsements of tobacco products by cultural icons and other influencers through their own social media accounts (HHS 2012). Thus, as part of ensuring digital media plans and buys for tobacco products are highly targeted to adults while limiting spill to youth, it is critical to mitigate against the incredible reach and influence of social media, including “organic” influencer promotion. Currently, there are no universal age-restriction controls on social media platforms and some do not offer any age-restriction options; however, many social media platforms are beginning to offer branded-account owners the option to age-restrict some or all of their account pages, followers, and content, including even specific posts, photos, videos, events, etc. These options still face a few additional limitations; for example, most social media platforms allow users to establish their own account profile settings, including self-reported age, and users are not age- or identity- verified. However, users are increasingly prompted to “link” digital profiles and accounts (e.g., option to sign-up for a new account using an existing email account or social media account), increasing the likelihood of more accurate self-reporting. As part of these restrictions, firms must ensure that their own social media accounts as well as those of any influencers promoting a tobacco product on a firm’s behalf use the available age-restriction controls to restrict youth access to any product promotion disseminated through social media accounts. Firms must also ensure the disclosure to consumers or viewers, via the use of statements such as “sponsored by [firm name],” of any relationships between the firm and entities that creating labeling for, advertise, marketing

118

and/or promote the product on the firm’s behalf to help prevent misleading marketing, which is especially likely to influence youth.

V. Proposed Marketing Restrictions in PMTA Authorization Orders

In this context, FDA should consider including detailed marketing restrictions and requirements, in addition to other requirements, for any new tobacco product receiving market authorization under sections 910(c)(1)(B) and 910(f) of the FD&C Act. FDA should determine such marketing restrictions and requirements on a case-by-case basis when issuing an order that the marketing of a tobacco product is appropriate for the protection of public health. Information that should be considered in these determinations includes, but is not limited to, information submitted to FDA by a firm seeking pre-market tobacco authorization regarding the firm’s intended labeling, advertising, marketing, and promotion of the products; use of industry practices known to substantially impact youth trial and uptake of tobacco product use; new and emerging technologies, media, and marketing practices; and existing applicable laws and legal agreements affecting the sales, distribution, marketing, advertising, labeling, and/or promotion of certain tobacco products. Generally, firms seeking marketing authorization for new tobacco products should seek to limit youth- exposure to the products’ labeling, advertising, marketing, and promotion. Restrictions in a marketing order should be aimed at the following with respect to advertising and marketing plans, including strategic creative briefs and paid media plans, by channel and by product: o Use of competent and reliable data sources, methodologies, and technologies to establish, maintain, and monitor highly targeted advertising and marketing plans and media buys; o Targeting of specific adult audiences by age-range(s), including young adults, ages 18-24, and other demographic and/or psychographic characteristics that reflect the intended target audience; o Actions taken to restrict youth-access and limit youth-exposure to the products’ labeling, advertising, marketing, and/or promotion; o Use of owned, earned, shared, and/or paid social media to create labeling for, advertise, market, and/or promote the products; o Use of partners, influencers, bloggers, and/or brand ambassadors to create labeling for, advertise, market, and/or promote the products; o Consumer engagements, including events at which the products were demonstrated; and/or o Use of earned media and/or public-relations outreach to create labeling for, advertise, market, and/or promote the products.

Firms should establish, maintain, and monitor use of independent age- and identity-verification service(s) that compare customer information against independent, competent, and reliable data sources, such as public records, to prevent digital sales of the products to individuals who are under the federal minimum legal age to purchase tobacco products. Firms should establish, maintain, and monitor use of independent age- and identity-verification service(s) that compare consumer information against independent, competent, and reliable data sources, such as public records, at the first point of access to any owned digital properties (e.g., the firm’s company-owned, consumer-directed, product-branded website(s) and/or mobile applications), to restrict access to any of the products’ labeling, advertising, marketing, and/or promotion appearing in such properties to only individuals who are at least of federal minimum legal age to purchase tobacco products. Firms should establish, maintain, and monitor use of the available site-, platform- and content- (e.g., post, video) specific age-restriction controls (e.g., age-restrict an entire product-branded account and all

119

associated content disseminated through such account; ensure age-restriction of a specific video disseminated by an influencer promoting the products on the firm’s behalf through the influencer’s account), at the first point of access to any shared digital properties (e.g., the firm’s product-branded social media accounts, pages and associated content; content promoting the products on the firm’s behalf disseminated through another entity’s social media accounts), to restrict access to any of the products’ labeling, advertising, marketing, and/or promotion appearing in such properties to only individuals who are at least of federal minimum legal age to purchase tobacco products. Firms should establish, maintain, and monitor use of competent and reliable data sources, methodologies, and technologies to precisely target delivery of any of the products’ labeling, advertising, marketing, and/or promotion appearing in paid digital media (e.g., paid digital banner advertisements for the product(s) running on another company’s website; paid advertising for the product(s) running in social media; paid distribution of influencer content), to restrict access to such labeling, advertising, marketing, and/or promotion to only individuals who are at least of federal minimum legal age to purchase tobacco products. Such targeting must use only first- and/or second-party age-verified data (see table). Firms should restrict advertising practices that are not and/or cannot be targeted using such data (e.g., tactics like “Run-of-Site,” “homepage takeovers,” “splashy buys”). Firms should establish, maintain, and monitor use of competent and reliable data sources, methodologies, and technologies (e.g., using an embedded tracking pixel in all digital advertising) to track and measure actual delivery of all advertising impressions, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a breakout by age-group (i.e., adults, ages 25+; young adults, ages 18-24; and youth, ages 12-17 and ages 11 and under). Such monitoring should require real-time digital media tracking, and identifying, correcting, and preventing delivery of advertising impressions to youth, ages 17 and under. Such monitoring also should require post-launch delivery verification reports be submitted to the firm from an accredited source (e.g., Media Ratings Council). Firms should disclose to consumers or viewers any relationships between the firm and entities that create labeling for, advertise, market, and/or promote the products, on the firm’s behalf or at the firm’s direction, via the use of statements such as “sponsored by [firm name]” in any such labeling, advertising, marketing, and/or promotional materials. It is vital to the continued protection of public health that FDA take these and other marketing-related considerations seriously when evaluating marketing plans to ensure they are sufficiently targeted to limit youth-exposure to tobacco product labeling, advertising, marketing, and promotion. The evaluation of these marketing plans, including evaluation of their potential impact on youth tobacco use, will help FDA determine whether the marketing, and continued marketing, of the products is appropriate for the protection of the public health.

VI. Conclusion

Given the level of evidence indicating the direct and powerful impact of tobacco marketing on youth tobacco use, and FDA’s statutory mandate to protect young people from the dangers of tobacco use, it is both reasonable and critical for firms to submit planned labeling, advertising, marketing, and promotional materials and plans for new tobacco products that are seeking or have received premarket authorization, and for FDA to place restrictions on the marketing of such products. This important safeguard will help FDA ensure, on an ongoing basis, that the continued marketing of new tobacco products remains appropriate for the protection of public health.

120

APPENDIX REFERENCES

Agaku IT, Ayo-Yusuf OA. The effect of exposure to pro-tobacco advertising on experimentation with emerging tobacco products among U.S. adolescents. Health Education and Behavior. 2014;41(3):275–80.

Akcay, O. Marketing to teenagers: The influence of color, ethnicity and gender. International Journal of Business and Social Science. 2012;3(22).

Centers for Disease Control and Prevention (CDC). Tobacco use among middle and high school students— United States, 2011–2015. Morbidity and Mortality Weekly Report. 2016;65(14):361–7.

Chu KH, Unger JB, Cruz TB, Soto DW. Electronic cigarettes on twitter–spreading the appeal of flavors. Tobacco Regulatory Science. 2015;1(1):36-41.

Cruz TB, McConnell R, Low BW, Unger JB, Pentz MA, Urman R, Berhane K, Chou CP, Liu F, Barrington-Trimis, JL. Tobacco marketing and subsequent use of cigarettes, e-cigarettes and hookah in adolescents. Nicotine & Tobacco Research. 2018;1-7.

Gentzke AS, Creamer M, Cullen KA, et al. Vital Signs: Tobacco Product Use Among Middle and High School Students — United States, 2011–2018. MMWR Morb Mortal Wkly Rep. 2019;68:157–164.

Interactive Advertising Bureau (IAB). Defining the data stack: How the power of data is impacting the rise of the direct brand economy. IAB Data Center of Excellence, 2018. https://www.iab.com/insights/defining-the- data-stack-2018/. Accessed March 19, 2019.

Hawkins RP, Kreuter M, Resnicow K, Fishbein M, Dijkstra A. Understanding tailoring in communicating about health. Health Education Research. 2008;23(3):454-466.

Huang, J, Duan Z, Kwok J, et al. Vaping versus JUULing: how the extraordinary growth and marketing of JUUL transformed the US retail e-cigarette market. Tobacco Control. 2018; tobaccocontrol-2018.

Lempert LK, Glantz SA. Implications of tobacco industry research on packaging colors for designing health warning labels. Nicotine & Tobacco Research. 2016;18(9):1910-1914.

Mantey DS, Cooper MR, Clendennen SL, Pasch KE, Perry CL. E-Cigarette marketing exposure is associated with e-cigarette use among U.S. youth. Journal of Adolescent Health. 2016;58:686–90.

National Cancer Institute (NCI). Monograph 19. The role of the media in promoting and reducing tobacco use. National Institutes of Health, 1998. https://cancercontrol.cancer.gov/brp/tcrb/monographs/19/m19 complete.pdf. Accessed January 17, 2019.

Padon AA, Lochbuehler K, Maloney EK, Cappella JN. A randomized trial of the effect of youth appealing e- cigarette advertising on susceptibility to use e-cigarettes among youth. Nicotine & Tobacco Research. 2017.

Padon AA, Maloney EK, Cappella JN. Youth-targeted e-cigarette marketing in the US. Tobacco Regulatory Science. 2017;3(1):95-101.

121

Pew Research Center. Teen, social media & technology 2018. http://www.pewinternet.org/2018/05/31/teens-social-media-technology-2018/. Updated May 21, 2018. Accessed January 3, 2019.

Pierce JP, Sargent JD, Portnoy DB, et al. Association Between Receptivity to Tobacco Advertising and Progression to Tobacco Use in Youth and Young Adults in the PATH Study. JAMA Pediatrics. 2018;172(5):444-451.

Slovic P, Peters E. Risk perception and affect. Current Directions in Psychological Science. 2006; 15(6):322- 325.

Soneji S, Yang J, Knutzen KE, Moran MB, Tan ASL, Sargent J & Choi K. Online tobacco marketing and subsequent tobacco use. Pediatrics. 2018; 141(2):2017-2927.

Soneji S, Yang J, Moran MB, Tan ASL, Sargent J, Knutzen KE & Choi K. Engagement with online tobacco marketing among adolescents in the US: 2013-2014 to 2014-2015. Nicotine & Tobacco Research. 2018;1-8.

Subramani MR, Rajagopalan B. Knowledge-sharing and influence in online social networks via . Communications of the ACM. 2003; 46(12):300-307.

The Nielsen Company (Nielsen). Effective advertising—more than a creative black box. https://www.nielsen.com/us/en/insights/news/2014/effective-advertising-more-than-a-creative-black- box.html. Updated May 19, 2014. Accessed January 4, 2019.

U.S. Department of Health and Human Services (HHS). Preventing Tobacco Use Among Youth and Young Adults. A Report of the Surgeon General. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2012. https://www.surgeongeneral.gov/library/reports/preventing-youth-tobacco-use/full-report.pdf. Accessed January 18, 2019.

U.S. Department of Health and Human Services (HHS). E-cigarette Use Among Youth and Young Adults. A Report of the Surgeon General. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2016. https://www.cdc.gov/tobacco/data statistics/sgr/e- cigarettes/pdfs/2016 sgr entire report 508.pdf. Accessed February 4, 2019.

122

Attachment 2: Philip Morris Products S.A. Marketing Order, FDA Submission Tracking Numbers (STNs): PM0000424, PM0000425, PM0000426, and PM0000479 (April 30, 2019) (“FDA Marketing Order”)

U.S. Food & Drug Administration 10903 New Hampshire Avenue Silver Spring, MD 20993 www.fda.gov

April 30, 2019 MARKETING ORDER

Philip Morris Products S.A. ATTENTION: Dr. Jeffrey Walker, CEO Teton Regulatory Sciences 1125 E. Glacier Road Jackson, Wyoming 83001-4876

FDA Submission Tracking Numbers (STNs): PM0000424-PM0000426, PM0000479

Dear Dr. Walker:

The Food and Drug Administration (FDA) completed the review of your Premarket Tobacco Product Applications (PMTAs) submitted under section 910(b) of the Federal Food, Drug, and Cosmetic Act (FD&C Act), specified in Appendix A.

Based on our review of your PMTAs, we find that the marketing of the new tobacco products specified in Appendix A is appropriate for the protection of public health, and that you have met the other requirements of section 910(c) of the FD&C Act. This marketing order is subject to marketing requirements under section 910(c)(1)(B) of the FD&C Act and record retention and reporting requirements under section 910(f) of the FD&C Act, as outlined in Appendix B. Additionally, this order is conditioned upon the products conforming with any applicable current or future tobacco product standards, unless specifically exempted under this order or the product standard(s). Under the provisions of section 910, you may introduce or deliver for introduction into interstate commerce the new tobacco products, in accordance with the order requirements outlined in Appendix B.

The requirements in this order are intended to help ensure that the marketing of your products will continue to be appropriate for the protection of the public health, taking into account initiation among non-users, particularly youth. However, compliance with these requirements alone is not a guarantee that the marketing of the products will remain appropriate for the protection of the public health, particularly if, despite these measures, there is a significant uptake in youth initiation, for example. FDA will continue to monitor the marketing of your products.

This order does not constitute a finding that any of the products outside the scope of this authorization are in compliance with the FD&C Act and its implementing regulations. FDA has not evaluated other components or parts, or accessories that you may choose to market with the iQOS system, such as A/C power adapters, USB cables, charging docks, cleaners, disposal units, and pouches. To the extent that any premarket authorization requirements of section 910 of the FD&C Act apply, FDA does not intend to enforce them with respect to such products. However, it is your responsibility to ensure that these products comply with all other applicable laws and regulations. For example, if you choose to include the brand name “IQOS” on items other than the products authorized in these orders, you need to evaluate whether that would comply with 21 CFR 1140.34(a). In addition, we recommend you evaluate whether any of the branded accessories you plan to market would constitute advertising that requires the applicable warnings.

PM0000424-PM0000426, PM0000479 Page 2 of 15

We note that, in your September 5, 2018 and March 25, 2019 amendments to your PMTAs, you include representations about your marketing plan for your products in the United States and indicate that you intend to focus marketing on adult cigarette smokers while limiting reach to unintended audiences. FDA encourages you to consider measures to limit youth-exposure to any of the products’ labeling, advertising, marketing, and/or promotion appearing in print media publications. Limiting youth exposure and initiation and use of the products as you have indicated in your PMTAs (i.e., complete switching to IQOS by adult cigarette smokers) are important components of consideration for the marketing of these products to continue to be appropriate for protection of the public health.

Also, in accordance with 40 CFR 1506.6, we will make your environmental assessments publicly available.

This order authorizing the marketing of these new tobacco products does not mean FDA “approved” the new tobacco products specified in Appendix A; therefore, you may not make any express or implied statement or representation directed to consumers that conveys, or misleads or would mislead consumers into believing, among other things, that the new tobacco products specified in Appendix A are “approved” by FDA.1 The products subject to this marketing order are subject to withdrawal or temporary suspension as described in section 910(d) of the FD&C Act.

We remind you that all regulated tobacco products, including the new tobacco products specified in Appendix A, are subject to the requirements of the FD&C Act and its implementing regulations. These requirements currently include, but are not limited to, annual registration, listing of products, listing of ingredients, reporting of harmful and potentially harmful constituents, and payment of user fees. There are also packaging, labeling, and advertising requirements with which you must comply. It is your responsibility to ensure the tobacco products specified in Appendix A comply with all applicable statutory and regulatory requirements. FDA will monitor your compliance with all applicable statutes and regulations.

If you discontinue the manufacture, preparation, compounding or processing for commercial distribution of these tobacco products and later decide to reintroduce the products into the market, please contact the Office of Science prior to reintroduction.

For more information on your responsibilities under the FD&C Act, we encourage you to visit our website at http://www.fda.gov/TobaccoProducts. You may also obtain information by contacting us at 1-877-CTP-1373, [email protected], or [email protected].

We encourage you to submit all regulatory correspondence electronically via the CTP Portal (http://www.fda.gov/TobaccoProducts/GuidanceComplianceRegulatoryInformation/Manufacturing/uc

1 See Section 301(tt) of the FD&C Act.

PM0000424-PM0000426, PM0000479 Page 3 of 15

m515047.htm)2 using eSubmitter (http://www.fda.gov/ForIndustry/FDAeSubmitter). Alternatively, submissions may be mailed to:

Food and Drug Administration Center for Tobacco Products Document Control Center (DCC) Building 71, Room G335 10903 New Hampshire Avenue Silver Spring, MD 20993-0002

The CTP Portal and the FDA Electronic Submission Gateway (ESG) are generally available 24 hours a day, seven days a week; if the upload is successful, submissions are considered received by DCC on the day of upload. Submissions delivered to DCC by courier or physical mail will be considered timely if received during delivery hours on or before the due date3; if the due date falls on a weekend or holiday, the delivery must be received on or before the preceding business day. We are unable to accept regulatory submissions by e-mail.

If you have any questions, you may contact Donna Cheung, Regulatory Health Project Manager, at (240) 402-5340 or [email protected].

Sincerely,

Digitally signed by Matthew R. Holman -S Date: 2019.04.30 08:29:11 -04'00'

Matthew R. Holman, Ph.D. Director Office of Science Center for Tobacco Products

2 The FDA’s Electronic Submission Gateway (ESG) is still available as an alternative to the CTP Portal. 3 See http://www.fda.gov/tobaccoproducts/aboutctp/contactus/default.htm

Appendices A and B Page 6 of 15

Appendix B

Postmarket Recordkeeping, Retention, Reporting, and Marketing Requirements

I. Record Retention

Under section 910(f) of the FD&C Act, this order requires that you establish and maintain the records listed below. The records must be retained for a period of not less than four years from the date of distribution of the last batch of the new tobacco products listed in your marketing authorization. The records must be legible, written in English, and available for inspection and copying by officers or employees duly designated by the Secretary upon request:

x The PMTA submitted prior to product order x Periodic postmarket reports, as described below, and adverse experience reports, including all relevant documentation associated with the experience x Records of all nonclinical or clinical studies, including: o Source data; o Study protocols (including statistical analysis plan); o Amendments showing the dates and reasons for any protocol revisions; o Institutional Review Board (IRB) or Independent Ethics Committee (IEC) approvals or non- approvals; o Informed consent forms; o Correspondence with study monitors/investigators/contract research organizations/sponsors/IRB/IEC; o Investigator financial disclosure statements; o Progress reports; o Monitoring reports; o Adverse experience reports; o Case report forms/subject diaries/medical records/laboratory reports; o Subject data line listings/observation records; o Test article accountability records; o Study results/protocol summaries/study reports; and o Certifications and amendments to certifications x Records pertaining to the manufacture, in process and release testing, production process (including any changes to the process, facility, or controls), packaging, storage, and stability monitoring and testing (including protocol and results) of the products x Records pertaining to the sale, distribution, or other disposition of the products, specifically: o A list of distributors and retailers of the products, including brick-and-mortar and digital8; o Any available information (not to include personally identifiable information) about product purchases, such as purchasers’ demographics (e.g., age, gender, race/ethnicity, geographic region) and previous or current use of other tobacco products (i.e., dual use); o Policies and procedures regarding verification of the age and identity of purchasers of the products; and o Policies and procedures regarding restrictions on youth access to the products

8 For the purposes of this order, here and throughout the document, “digital” includes internet/online and mobile.

Appendices A and B Page 7 of 15

x Records pertaining to the products’ labeling, advertising, marketing, and/or promotion – whether conducted by you, on your behalf, or at your direction – including: o Specimens of all labeling, labels, inserts/onserts, instructions, and other accompanying information; o Copies of all advertising, marketing, and/or promotional materials published, disseminated to consumers, or for use in engaging or communicating with consumers; o Copies of any formative research studies conducted among any audiences in the formation of the labeling, advertising, marketing, and/or promotional materials, including qualitative and quantitative research studies used to determine message effectiveness, consumer knowledge, attitudes, beliefs, intentions, and behaviors toward using the products, and including copies of the stimuli used in testing; o Copies of any consumer evaluation research studies conducted among any audiences to determine the effectiveness of labeling, advertising, marketing, and/or promotional materials and any shifts in consumer knowledge, attitudes, beliefs, intentions, and behaviors toward using the products, and including copies of the stimuli used in testing; o Copies of any contractual agreements regarding the creation and/or dissemination of the products’ labeling, advertising, marketing, and/or promotional materials; o Copies of all advertising and marketing plans, including strategic creative briefs and paid media plans, by channel and by product, and the dollar amount(s) and flighting of such plans, by channel and by product, including any: ƒ Use of competent and reliable data sources, methodologies, and technologies to establish, maintain, and monitor highly targeted advertising and marketing plans and media buys; ƒ Targeting of specific adult audiences by age-range(s), including young adult audiences, ages 18-24, and other demographic and/or psychographic characteristics that reflect your intended target audience; ƒ Actions taken to restrict youth-access and limit youth-exposure to the products’ labeling, advertising, marketing, and/or promotion; ƒ Use of owned, earned, shared, and/or paid social media to create labeling for, advertise, market, and/or promote the products; ƒ Use of partners, influencers, bloggers, and/or brand ambassadors to create labeling for, advertise, market, and/or promote the products; ƒ Consumer engagements – whether conducted by you, on your behalf, or at your direction – including events at which the products were demonstrated; and/or ƒ Use of earned media and/or public-relations outreach to create labeling for, advertise, market, and/or promote the products o Copies of all records pertaining to media tracking and optimization, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic region), and all post-launch delivery-verification reports submitted to you from an accredited source, by channel, by product, and by audience demographics; and o Policies and procedures for real-time digital media monitoring to identify, correct, and prevent any delivery of advertising impressions to youth, ages 17 years and under, including documentation of such monitoring activities and implementation of corrective and preventive measures x Health hazard analyses, if performed voluntarily or directed by FDA x Records pertaining to any and all complaints associated with any of the products that you receive or of which you are aware

Appendices A and B Page 8 of 15

II. Serious and Unexpected Adverse Experiences Reporting

Under section 910(f) of the FD&C Act, this order requires that you report to the FDA all adverse experiences that are both serious and unexpected and your analysis of the association between the adverse experience and the tobacco product(s) within 15 calendar days after the report is received by you. These experiences may become known to you through any source including a customer complaint, request, or suggestion made as a result of an adverse experience, tobacco product defect, or failure, reported to you, or identified in the literature or media. Your information should be submitted with a cover letter that includes the following text in the subject line: SERIOUS UNEXPECTED ADVERSE EXPERIENCE REPORT FOR STN(s) XXX.

For purposes of reporting under this order, serious adverse experience means an adverse experience that results in any of the following outcomes: x Death; x A life-threatening condition or illness; x Inpatient hospitalization or prolongation of existing hospitalization; x A persistent or significant incapacity or substantial disruption of the ability to conduct normal life functions; x A congenital anomaly/birth defect; or x Any other adverse experience that, based upon appropriate medical judgment, may jeopardize the health of a person and may require medical or surgical intervention to prevent one of the other outcomes listed in this definition.

For purposes of reporting under this order, unexpected adverse experience means an adverse experience occurring in one or more persons in which the nature, severity, or frequency of the experience is not consistent with: x The known or foreseeable risks associated with the use or exposure to the tobacco product as described in the PMTA (including the results of human subject investigations) and other relevant sources of information, such as product labeling and postmarket reports; x The expected natural progression of any underlying disease, disorder, or condition of the person(s) experiencing the adverse experience and the person’s predisposing risk factor profile for the adverse experience; or x The results of nonclinical laboratory studies.

III. Manufacturing Deviations

Under section 910(f) of the FD&C Act, this order requires that you establish and maintain records and reports of all manufacturing deviations, investigations, and corrective and preventive actions including, but not limited to, those deviations associated with processing, testing, packing, labeling, storage, holding, and distribution. For products that have been distributed, if a deviation occurs that you determine presents a reasonable probability that the tobacco product contains a manufacturing or other defect not ordinarily contained in tobacco products on the market that would cause serious, adverse health consequences or death you are required to report the deviation to FDA within 15 calendar days of identification.

Appendices A and B Page 9 of 15

IV. Periodic Reporting

The information in the following postmarket periodic reports will help FDA determine whether continued marketing of your tobacco products is appropriate for the protection of public health and/or there are or may be other grounds for withdrawing or temporarily suspending the marketing authorization order.

Under section 910(f) of the FD&C Act, this order requires that you submit the following periodic reports to FDA on a quarterly basis, for a period of two years, beginning three months from the date of this order. For each three-month reporting period, these periodic reports must include: x A cover letter that includes the following text in your subject line: PERIODIC REPORT for PM0000424-PM0000426, PM0000479. The cover letter should include the STN(s) and corresponding tobacco product name(s), firm name, date of report, reporting period. x A summary of U.S. sales and distribution of the tobacco products, including total U.S. sales reported in dollars, units, and volume, and broken down by U.S. census region, major retail markets, and channels where the products are sold (e.g., convenience stores, food and drug markets, big box retailers, digital platforms, tobacco specialty shops, company-owned stores). This summary must also be broken down by product (e.g., specific HeatStick flavor). x Data on product purchasers. Report any data collected about new purchasers, those who have switched tobacco products, and/or multiple product users. The results must be broken down by purchaser demographics (e.g., age, gender, race/ethnicity, geographic location) and must not include personally identifiable information.

Under section 910(f) of the FD&C Act, this order also requires that you also submit periodic reports to FDA on a quarterly basis, beginning three months from the date of this letter. For each three-month reporting period, these periodic reports must include: x A cover letter that includes the following text in your subject line: PERIODIC REPORT for PM0000424-PM0000426, PM0000479. The cover letter should include the STN(s) and corresponding tobacco product name(s), firm name, date of report, reporting period. x An analysis of the actual delivery of advertising impressions, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a breakout by age-group (i.e., adults, ages 25+; young adults, ages 18-24; and youth, ages 12-17 and ages 11 and under). This analysis must be verified against post-launch delivery-verification reports submitted to you from an accredited source. x A summary of media tracking and optimization, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a summary of real-time digital media monitoring to identify, correct, and prevent delivery of advertising impressions to youth, ages 17 and under, and including a summary of implementation of any corrective and preventive measures.

Under section 910(f) of the FD&C Act, this order also requires that you submit the following periodic reports to FDA on an annual basis, beginning twelve months from the date of this order. For each twelve-month reporting period, these periodic reports must include: x A cover letter that includes the following text in your subject line: ANNUAL REPORT for PM0000424-PM0000426, PM0000479. The cover letter should include the STN(s) and corresponding tobacco product name(s), firm name, date of report, reporting period. x A summary of how the marketing of the tobacco products continues to be appropriate for the protection of public health, which includes:

Appendices A and B Page 10 of 15

o A status report of ongoing studies and a summary of completed studies about the tobacco products conducted by you, or on your behalf. o A summary of significant findings in publications not previously reported and full copies of the articles. This must include any new scientific data (published or otherwise) on the likelihood of product use by current users of tobacco products within the same tobacco product category, current users of tobacco products in other tobacco product categories, former users of any tobacco product, and youth and young adults. o A summary of reported adverse experiences for the tobacco products, which includes a listing of all adverse experiences, including the serious and unexpected adverse experiences previously reported. The summary must be accompanied by an analysis of the reports and a statement of any changes to risk information related to the products including nature, frequency, and potential aggravating factors. o A summary of U.S. sales and distribution of the tobacco products, not previously submitted, including total U.S. sales reported in dollars, units, and volume, and broken down by U.S. census region, major retail markets, and channels where the products are sold (e.g., convenience stores, food and drug markets, big box retailers, digital platforms, tobacco specialty shops, company-owned stores). This summary must also be broken down by product (e.g., specific HeatStick flavor). o Data on product purchasers not previously submitted. Report any data collected about new purchasers, those who have switched tobacco products, and/or multiple product users. The results must be broken down by purchaser demographics (e.g., age, gender, race/ethnicity, geographic location) and must not include personally identifiable information. o A summary of the implementation and effectiveness of your policies and procedures regarding verification of the age and identity of purchasers of the products. o A summary of the implementation and effectiveness of your policies and procedures regarding restrictions on youth access to the products. o A description of each change made to the manufacturing process, facilities, or controls during the reporting period including: ƒ A comparison of each change to what was described in the PMTAs; ƒ The rationale for making each change; and ƒ A certification that the reported change did not result in any modification (including a change in design, any component, any part, or any constituent, including a smoke or aerosol constituent, or in the content, delivery, or form of nicotine, or any other additive or ingredient) of the tobacco products and the basis for concluding that each manufacturing change did not result in any modification to the products. o A summary of all manufacturing deviations, investigations, and corrective and preventive actions, including, but not limited to, those deviations associated with processing, testing, packing, labeling, storage, holding, and distribution and indicate any deviation(s) that may affect the characteristics of the products. o A summary of any stability monitoring and testing of the HeatSticks products, including monitoring and testing protocol (including batch/lot sampling) and results. o All final printed labeling (including all labeling variations, such as those reflecting different required warnings) not previously submitted, including the date the labeling was first disseminated and the date when the labeling was discontinued, and a description of all changes to the labeling. The labeling must include all the panels and be presented in the actual size and color with legible text. The labeling must include labels, inserts/onserts, instructions, and any other accompanying information or materials for the products.

Appendices A and B Page 11 of 15

o All final full-color advertising, marketing, and/or promotional materials, published, disseminated to consumers, or for use in engaging or communicating with consumers not previously submitted, along with the original date such materials were first disseminated and the date they were discontinued, and a description of all changes to the materials. The materials must include all panels where applicable (e.g., print ads, point of sale signs) and reflect the actual size and colors used. For any materials that would not fit on an 8.5” x 11” piece of paper, you may resize and submit electronic versions of such materials in a format that FDA can review and with sufficient resolution to allow FDA to read lettering clearly. If resizing the advertisement does not allow for text to be read easily, the text may be provided separately and referenced. o A summary of all formative consumer research studies conducted – whether by you, on your behalf, or at your direction – among any audiences, in the formation of new labeling, advertising, marketing, and/or promotional materials, including qualitative and quantitative research studies used to determine message effectiveness, consumer knowledge, attitudes, beliefs, intentions and behaviors toward using the products, and including the findings of these studies and copies of the stimuli used in testing. o A summary of all consumer evaluation research studies conducted – whether by you, on your behalf, or at your direction – among any audiences, to determine the effectiveness if labeling, advertising, marketing and/or promotional materials and any shifts in consumer knowledge, attitudes, beliefs, intentions, and behaviors toward using the products, and including the findings of these studies and copies of the stimuli used in testing. o A summary of the creation and dissemination of the products’ labeling, advertising, marketing, and/or promotional materials – whether conducted by you, on your behalf, or at your direction – including a list of all entities involved and a description of their involvement, including a description of contractual agreements with such entities. o A description of the implementation of all advertising and marketing plans, including strategic creative briefs and paid media plans – whether conducted by you, on your behalf, or at your direction – by channel and by product, and the dollar amount(s) and flighting of such plans, by channel and by product, including a description of any: ƒ Use of competent and reliable data sources, methodologies, and technologies to establish, maintain, and monitor highly targeted advertising and marketing plans and media buys; ƒ Targeting of specific adult audiences by age-range(s), including young adults, ages 18-24, and other demographic and/or psychographic characteristics that reflect the intended target audience, including a list of all data sources used to target advertising and marketing plans and media buys; ƒ Actions taken to restrict youth-access and limit youth-exposure to the products’ labeling, advertising, marketing, and/or promotion; ƒ Use of owned, earned, shared, and/or paid social media to create labeling for, advertise, market, and/or promote the products; ƒ Use of partners, influencers, bloggers, and/or brand ambassadors to create labeling for, advertise, market, and/or promote the products; ƒ Consumer engagements – whether conducted by you, on your behalf, or at your direction – including events at which the products were demonstrated; and/or ƒ Use of earned media and/or public-relations outreach to create labeling for, advertise, market, and/or promote the products;

Appendices A and B Page 12 of 15

including the original date such plans were first used and the date they were discontinued, and a description of all changes to such plans since the last periodic report, by channel and by product. o An analysis of the actual delivery of advertising impressions, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a breakout by age-group (i.e., adults, ages 25+; young adults, ages 18-24; and youth, ages 12-17 and ages 11 and under), not previously submitted. This analysis must be verified against post-launch delivery-verification reports submitted to you from an accredited source. o A summary of media tracking and optimization, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a summary of real-time digital media monitoring to identify, correct, and prevent delivery of advertising impressions to youth, ages 17 and under, and including a summary of implementation of any corrective and preventive measures, not previously submitted.

Under sections 910(c)(1)(B) and 910(f) of the FD&C Act, this order also requires that you provide the following notifications to FDA. These notifications are not for pre-approval, but are required so that FDA can have timely access to your marketing plans and materials, and if needed, provide you advisory comments, including any concerns about their possible impact on youth appeal and tobacco use initiation and on the finding that continued marketing of your products is appropriate for the protection of the public health. You may begin disseminating the materials 30 days after providing notification to FDA. x Provide FDA notification of all labeling, advertising, marketing, and/or promotional materials for which you have not previously provided notification, at least 30 days prior to the initial publication, dissemination to consumers, or use in engaging or communicating with consumers of such materials, and include in your notification: o Full-color copies of all such labeling, advertising, marketing, and/or promotional materials for the products. The materials must include all panels where applicable (e.g., print ads, point of sale signs) and reflect the actual size and colors used. For any materials that would not fit on an 8.5” x 11” piece of paper, you may resize and submit electronic versions of such materials in a format that FDA can review and with sufficient resolution to allow FDA to read lettering clearly. If resizing the advertisement does not allow for text to be read easily, the text may be provided separately and referenced. o All advertising and marketing plans, including strategic creative briefs and paid media plans, by channel and by product, and the dollar amount(s) and flighting of such plans, by channel and by product, including any plans to: ƒ Use competent and reliable data sources, methodologies, and technologies to establish, maintain, and monitor highly targeted advertising and marketing plans and media buys; ƒ Target specific adult audiences by age-range(s), including young adults, ages 18-24, and other demographic and/or psychographic characteristics that reflect your intended target audience, including a list of all data sources used to target advertising and marketing plans and media buys; ƒ Restrict youth-access and limit youth-exposure to the products’ labeling, advertising, marketing, and/or promotion; ƒ Use owned, earned, shared, and/or paid social media to create labeling for, advertise, market, and/or promote the products; ƒ Use partners, influencers, bloggers, and/or brand ambassadors to create labeling for, advertise, market, and/or promote the products;

Appendices A and B Page 13 of 15

ƒ Conduct any consumer engagements – whether by you, on your behalf, or at your direction – including events at which the products will be demonstrated; and/or ƒ Use earned media and/or public-relations outreach to create labeling for, advertise, market, and/or promote the products.

V. Marketing Requirements

Under section 910(c)(1)(B) of the FD&C Act and in accordance with section 202(a) of the Family Smoking Prevention and Tobacco Control Act, this order requires:

x Inclusion of the warning statement: “WARNING: This product contains nicotine. Nicotine is an addictive chemical.” on the package labels of all HeatSticks packs and of all kits containing HeatSticks packs as well as in all advertisements for such products and kits.9 Specifically, the warning statement must appear directly on the package and must be clearly visible underneath any cellophane or other clear wrapping as follows: o Be located in a conspicuous and prominent place on the two principal display panels of the package and the warning area must comprise at least 30 percent of each of the principal display panels; o Be printed in at least 12-point font size and the warning statement must occupy the greatest possible proportion of the warning area set aside for the required text; o Be printed in conspicuous and legible Helvetic bold or Arial bold type (or other sans serif fonts) and in black text on a white background or white text on a black background in a manner that contrasts by typography, layout, or color, with all other printed material on the package; o Be capitalized and punctuated as indicated in this order; and o Be centered in the warning area in which the text is required to be printed and positioned such that the text of the warning statement and the other information on the principal display panel have the same orientation. x For print advertisements and other advertisements with a visual component (including, for example, advertisements on signs, shelf-talkers, websites, mobile applications, and e-mail), the warning statement must appear in the upper portion of the area of the advertisement within the trim area as follows: o Occupy at least 20 percent of the area of the advertisement; o Appear in at least 12-point font size and the warning statement must occupy the greatest possible proportion of the warning area set aside for the required text; o Appear in conspicuous and legible Helvetica bold or Arial bold type (or other similar sans serif fonts) and in black text on a white background or white text on a black background in a manner that contrasts by typography, layout, or color, with all other material on the advertisement; o Be capitalized and punctuated as indicated in this order; o Be centered in the warning area in which the text is required to appear and positioned such that the text of the warning statement and the other textual information in the advertisement have the same orientation; and

9 This warning must appear on each package and each advertisement, in addition to the rotating Surgeon General warnings required under FCLAA (except the carbon monoxide warning, which is to be removed from the rotation of the Surgeon General warnings as described in this order). When FDA promulgates a final rule with respect to health warnings for cigarettes, FDA will reevaluate the conditions of marketing with respect to warnings for the products subject to this order.

Appendices A and B Page 14 of 15

o Be surrounded by a rectangular border that is the same color as the text of the warning statement and that is not less than 3 millimeters (mm) or more than 4 mm. x Removal of the warning: “SURGEON GENERAL’S WARNING: Cigarette Smoke Contains Carbon Monoxide.” from the required warnings to be displayed on the product package labels and advertisements under the Federal Cigarette, Labeling and Advertising Act (FCLAA). x As a reminder, under section 4 of FCLAA (15 U.S.C. 1333), you must submit a warning plan to the United States Federal Trade Commission (FTC).

Under section 910(c)(1)(B) of the FD&C Act, this order requires you to:

x For any digital sales – whether conducted by you, on your behalf, or at your direction – establish, maintain, and monitor use of independent age- and identity-verification service(s) that compare customer information against independent, competent, and reliable data sources, such as public records, to prevent the sale of the products to individuals who are under the federal minimum legal age to purchase tobacco products. x For any of the products’ labeling, advertising, marketing, and/or promotion appearing in your owned digital properties (e.g., your company-owned, consumer-directed, product-branded website(s) and/or mobile applications) – whether conducted by you, on your behalf, or at your direction – establish, maintain, and monitor use of independent age- and identity-verification service(s) that compare consumer information against independent, competent, and reliable data sources, such as public records, at the first point of access to such properties, to restrict access to such labeling, advertising, marketing, and/or promotion to only individuals who are at least of federal minimum legal age to purchase tobacco products. x For any of the products’ labeling, advertising, marketing, and/or promotion appearing in any shared digital properties (e.g., your product-branded social media accounts, pages and associated content; content promoting your products on your behalf disseminated through another entity’s social media accounts) – whether conducted by you, on your behalf, or at your direction – establish, maintain, and monitor use of the available site-, platform- and content- (e.g., post, video) specific age-restriction controls (e.g., age-restrict an entire product-branded account and all associated content disseminated through such account; ensure age-restriction of a specific video disseminated by an influencer promoting the products on your behalf through the influencer’s account), at the first point of access to such properties, to restrict access to such labeling, advertising, marketing, and/or promotion to only individuals who are at least of federal minimum legal age to purchase tobacco products. x For any of the products’ labeling, advertising, marketing, and/or promotion appearing in paid digital media (e.g., paid digital banner advertisements for the product(s) running on another company’s website; paid advertising for the product(s) running in social media; paid distribution of influencer content) – whether conducted by you, on your behalf, or at your direction: o Establish, maintain, and monitor use of competent and reliable data sources, methodologies, and technologies to precisely target delivery of such labeling, advertising, marketing, and/or promotion to only individuals who are at least of federal minimum legal age to purchase tobacco products. Such targeting must use only first- and/or second-party age-verified data, where: ƒ “First-party” age-verified data is data owned by you (e.g., your customer registration data collected via site traffic to your company-owned website; data you use in direct marketing to your adult smoking customers) that you have age-verified through independent, competent, and reliable data sources; and

Appendices A and B Page 15 of 15

ƒ “Second-party” age-verified data is first-party data owned and age-verified by another competent and reliable entity (e.g., another company’s first-party user registration data) to which you have access. Such data must be age-verified by the second party. ƒ “First-party” and “second-party” data does not include data obtained from data aggregators who categorize consumers based on trackable activities and inferred interests (e.g., internet search terms, video interactions, browsing history, purchasing behaviors) to create demographic and psychographic profiles marketers may use to enhance audience targeting. Such data is not considered age-verified and can only be used in combination with first- and/or second-party age-verified data. x Establish, maintain, and monitor use of competent and reliable data sources, methodologies, and technologies (e.g., using an embedded tracking pixel in all digital advertising) – whether conducted by you, on your behalf, or at your direction – to track and measure actual delivery of all advertising impressions, by channel, by product, and by audience demographics (e.g., age, gender, race/ethnicity, geographic location), including a breakout by age-group (i.e., adults, ages 25+; young adults, ages 18-24; and youth, ages 12-17 and ages 11 and under). Such monitoring requires real-time digital media tracking, and identifying, correcting, and preventing delivery of advertising impressions to youth, ages 17 and under. Such monitoring also requires post-launch delivery verification reports be submitted to you from an accredited source. x For any use of partners, influencers, bloggers, and/or brand ambassadors to create labeling for, advertise, market, and/or promote the products – whether conducted by you, on your behalf, or at your direction – disclose to consumers or viewers, via the use of statements such as “sponsored by [firm name]” in such labeling, advertising, marketing, and/or promotional materials, any relationships between you and entities that create labeling for, advertise, market, and/or promote the products, on your behalf, or at your direction.

Attachment 3: Declaration of Yvonne DeVerry

DECLARATION OF YVONNE DEVERRY

1. The matters set forth herein are truthful and accurate to the best of my knowledge. The information contained herein is based on my personal knowledge, except where otherwise stated, and in those instances the information is based on my information and belief. Background 2. I have been employed by Altria Client Services LLC ("ALCS") and Philip Morris USA Inc. ("PM USA") for more than 17 years and have 28 years' experience in the field of consumer product packaging development and design.l 3. From 1991-2002, I held various roles in the printing and packaging industry. In those roles, I gained experience on various aspects of the industry, including technical printing expertise and packaging development. 4. I joined PM USA in 2002 as an Associate Staff Engineer within the Packaging Services Department. In that role, I was charged with developing packaging materials for the domestic and businesses. In subsequent years, I held different positions within the Packaging Services organization, including Staff Engineer and Manager of Domestic Packaging Development, where my areas of responsibility included development and execution of the packaging for brand launches, redesigns, and promotions for Altria's tobacco operating companies, including PM USA. 5. In 2013, after a 3-year rotational assignment, I returned to the Packaging Design and Innovation department as a Senior Manager to lead the department in the design, development, and commercialization of revenue and promotional packaging to meet marketing, regulatory, and manufacturing requirements for Altria's tobacco operating companies, including PM USA.

1 ALCS is a separate, wholly-owned subsidiary of Altria Group, Inc., as is PM USA. ALCS provides a range of professional services and additional support to the parent company and each of the operating companies, such as packaging services, pursuant to specific contracts or agreements. Page 1 of 5 6. In my current role as Senior Director of Packaging Design and Innovation, I lead the long-term strategic direction for Packaging Design, Development and Execution for Altria's tobacco operating companies, including material and specification requirements to ensure regulatory compliance. 7. Throughout my 28-year career in the packaging field, I have gained knowledge and expertise with printing and converting techniques used in the development and execution of primary and tertiary packaging for consumer products. During the 17 years I have spent with ALCS and PM USA, I have become intimately familiar with the packaging production techniques used by our suppliers, as well as with the regulatory and labeling requirements with which our tobacco operating companies must comply. 8. PM USA is a wholly-owned subsidiary of Altria. It is the largest cigarette manufacturer in the United States. PM USA's brands comprise approximately 50 percent of cigarette sales in the United States. Altria acquired Nat Sherman in 2018. Nat Sherman also sells cigarettes. 9. I have evaluated FDA's graphic warnings proposal ("Proposed Rule") as it applies to cigarette packaging and have considered implementation issues raised by the Proposed Rule. 10. When the Final Rule goes into effect, I wi111ikely be responsible for implementing the requirements for cigarette packages. Random and Equal Distribution Issues 11. The Proposed Rule requires that manufacturers randomly display 13 warnings in as equal a number of times as is possible, and that the warnings be randomly distributed over a 12- month period in all areas of the United States in which the cigarette is marketed. 12. To the best of my knowledge, the requirement to print 13 graphics in equal numbers is unheard of in the printing and packaging industry. It is common to print, for example, 9 or 12 different graphics. 13. Thirteen is both an odd and prime number that is fundamentally incompatible with industry-wide printing practices and the random and equal distribution requirements of the Proposed Rule.

Page 2 of 5 14. When planning a print run, packaging designers arrange the desired package graphics in a grid layout. That layout depends on the package style and press used. Each rotation of the press cylinder prints the grid pattern onto packaging materials. 15. This grid pattern is based on the technical capabilities of the printing press, which are fixed. 16. Our hard pack cigarette packages are printed on 32-inch presses or 26-inch presses. The 32-inch press prints in an 8x3 layout. Our 26-inch press prints in a 6x3 layout. Cartons use 2layouts: 1x2 and 2x2. 17. Soft packs are printed in 2layouts: 9x5 and 8x5. We are exploring the possibility of retrofitting our soft pack presses to print in an 8x61ayout. 18. The 8x3 layout illustrates implementation issues common to all layout patterns. 19. Without drastically changing printing and packaging practices and adopting unproven methods, the requirement for 13 warnings could result in outcomes where one warning appears approximately 10 percent less frequently than another warning. Other presses and layouts could result in nearly 35 percent more of one warning than another. 20. We are exploring alternate means to attempt to comply the proposed requirement for 13 warnings. For example, we are exploring the feasibility of printing blanks in the layout and designing a process to cut out the blanks and discard them. We are also exploring the feasibility of adding a new stack of the 13th warning on a pallet containing the other warnings. 21. We do not know if any of these novel alternatives could prove to be commercially scalable, or would ultimately satisfy the applicable requirements of the Final Rule. Any of the alternatives, if they are even possible, are orders of magnitude more complex and significantly more expensive than randomly and equally distributing 12 or 9 warnings. 22. We have estimated that it could be 40 to 165 percent costlier to meet the requirement for 13 warnings than for 12 or 9 warnings. These estimates are from a materials perspective and assume the changes are possible, can be commercially scaled, and do not present any unforeseen obstacles. From a materials perspective, the cost to print 12 or 9 warnings is about the same. 23. If equal numbers of warnings are produced by the printer, randomization occurs though normal manufacturing and distributions processes.

Page 3 of 5 24. Manufacturers could achieve a random and equal distribution of warnings if the number of required warnings were 12 or 9. 25. Manufacturers could more easily achieve random and equal distribution for 12 or 9 warnings if FDA allows them to print two different warnings on the front and back of a single package. 26. To the best of my knowledge about industry printing and packaging practices, other tobacco companies and their suppliers would face similar issues if the proposed requirement for 13 warnings were to take effect. Additional Considerations 27. Reference 18 of the Proposed Rule provides each proposed warning in a single shape and single height-to-width ratio ("aspect ratio"). 28. PM USA and Nat Sherman create packages in various sizes. We provide a sampling of our packages in Exhibit A. 29. PM USA and Nat Sherman are unable to fill the required 50 percent warning area of all their packs and cartons with the proposed warning without adapting the warnings' aspect ratio. 30. PM USA and Nat Sherman can fill the required 50 percent warning area by changing the aspect ratio, but this will result in minor distortions of the warning. Some distortions will be slight and some more significant. We show the results of those adaptations in Exhibit B. 31. PM USA and Nat Sherman can maintain the aspect ratio of the proposed warnings, but this will result in empty space within the 50 percent required area. Exhibit C illustrates this point. 32. Content printed on the hinged lid of hard packs may shift up or down by up to about a millimeter at the point where the lid meets the pack front. This variability is likely to be most noticeable when text appears across the hinged lid and the pack front. 33. Normal variation in the manufacturing process may cause a small gap in the hinged lid, approximately 1.5 mm in size, where the lid meets the front of the package. This effect cannot be overcome through manufacturing practices.

Page 4 of 5 34. Packages with hinged lids may also result in individual words being split by the hinge. We show this in Exhibit D. To avoid this, FDA should provide manufacturers with font suitcases and allow manufacturers to flow text within a designated area. 35. The tops of cigarette soft packs are folded down after the cigarettes are placed in the packages and are then held together by an adhesive closure. 36. A 0.482-inch overlap of the closure onto the front of the pack would be sufficient to ensure that there is enough adhesive for the package to remain closed while also accounting for manufacturing-related variation. A 0.375-inch overlap for the closure would be insufficient. 37. FDA should provide industry with the "printer's proofs" for printing packaging. These proofs would help ensure that suppliers can accurately and consistently reproduce color graphics. It is standard industry practice to supply printing houses with these proofs. 3 8. To my knowledge and belief, manufacturers do not calibrate their equipment with one another and use different inlc application methods and substrates.

Executed on this the~ ay of October 2019, at Richmond, Virginia.

By: YVONNE DEVERRY

Page 5 of 5 EXHIBIT A (page 1 of 2) Representative Sample* of Cigarette Pack Dimensions Manufactured by PMUSA *PMUSA has 14 Different Pack Dimensions – These 5 are a Representative Sample Pink line represents the 50% boundary of the required warning

54.8mm x 74.3mm

54.8mm x 85.5mm

51.4mm x 123.2mm

47.5mm x 99.2mm 27.9mm x 100.6mm EXHIBIT A (page 2 of 2) Representative Sample* of Cigarette Carton Dimensions Manufactured by PM USA *PMUSA has 14 Different Carton Dimensions – These 5 are a Representative Sample Pink line represents the 50% boundary of the required warning

281.6mm x 75.7mm 270.7mm x 103.2mm

282.6mm x 87.2mm 265.7mm x 125.0mm

147.0mm x 101.5mm EXHIBIT B (page 1 of 2) Proposed Warning Adapted to Fit 50% of Representative Sample* of Cigarette Pack Dimensions *PMUSA has 14 Different Pack Dimensions – These 5 are a Representative Sample Pink line represents the 50% boundary of the required warning

54.8mm x 74.3mm

54.8mm x 85.5mm

51.4mm x 123.2mm

47.5mm x 99.2mm 27.9mm x 100.6mm EXHIBIT B (page 2 of 2) Proposed Warning adapted to fit 50% of Representative Sample* of Cigarette Carton Dimensions *PMUSA has 14 Different Carton Dimensions – These 5 are a Representative Sample Pink line represents the 50% boundary of the required warning

281.6mm x 75.7mm 270.7mm x 103.2mm

282.6mm x 87.2mm 265.7mm x 125.0mm

147.0mm x 101.5mm EXHIBIT C (page 1 of 2) Proposed Warning Placed in a Representative Sample* of Cigarette Pack Dimensions Without Adapting to 50% Area (Black Represents Remainder of Required 50% Warning Area) *PMUSA has 14 Different Pack Dimensions – These 5 are a Representative Sample Pink line represents the 50% boundary of the required warning

54.8mm x 74.3mm

54.8mm x 85.5mm

51.4mm x 123.2mm

47.5mm x 99.2mm 27.9mm x 100.6mm EXHIBIT C (page 2 of 2) Proposed Warning Placed in a Representative Sample* of Cigarette Carton Dimensions Without Adapting to 50% Area (Black Represents Remainder of Required 50% Warning Area) *PMUSA has 14 Different Carton Dimensions – These 5 are a Representative Sample Pink line represents the 50% boundary of the required warning

281.6mm x 75.7mm 270.7mm x 103.2mm

282.6mm x 87.2mm 265.7mm x 125.0mm

147.0mm x 101.5mm EXHIBIT D (page 1 of 1) Proposed Warning Adapted to Fit the 50% Required Area of a Representative Sample* of Cigarette Hinge Lid Packs Black Line Represents the Location of the Hinged Lid on Proposed Warning *PMUSA has 12 Different Hinged Lid Packs – These 5 are a Representative Sample Pink line represents the 50% boundary of the required warning

54.8mm x 74.3mm

54.8mm x 85.5mm

51.4mm x 123.2mm

47.5mm x 99.2mm 27.9mm x 100.6mm

Attachment 4: Comments from Altria Client Services LLC Docket No. FDA-2018-N-3552 November 26, 2018

Altria Altria Client Services

Jose Luis Murillo Senior Vice President Regulatory Affairs

November 26, 2018

Via Electronic Submission

Dockets Management Staff (HFA-305) Food and Drug Administration 5630 Fishers Lane, Rm. 1061 Rockville, MD 20852

Re: Docket No. FDA-2018-N-3552 (83 Fed. Reg. 48,625, September 26, 2018)— Comments on "Agency Information Collection Activities; Proposed Collection; Comment Request; Experimental Study of Cigarette Warnings"

Altria Client Services (ALCS"), on behalf of Philip Morris USA Inc. ("PM USA") and Sherman Group Holdings LLC and its subsidiaries ("Nat Sherman"),1 submits these comments to the Food and Drug Administration's ("FDA" or the "Agency") above-captioned docket entitled "Agency Information Collection Activities; Proposed Collection; Comment Request; Experimental Study of Cigarette Warnings" ("Notice).

Altria, on behalf of its tobacco operating companies, actively supported the passage of the Tobacco Control Act2 ("TCA" or the "Acr) because we believed that a national regulatory framework could contribute to resolving many of the complex issues that surround tobacco products. We also believe that regulation could create guidelines for accurate and scientifically grounded communications about tobacco products to adult tobacco consumers and could provide smokers with accurate risk information. We support a single, consistent public health message on the role of cigarette smoking in the development of disease in smokers, and on smoking and addiction.

1 PM USA and Nat Sherman are wholly-owned subsidiaries of Altria Group, Inc. ("Altrie). PM USA manufactures cigarettes and Nat Sherman manufactures cigarettes, cigars, and pipe tobacco. ALCS provides certain services, including regulatory affairs, to the Altria family of companies. "we and "our are used throughout to refer to PM USA and Nat Sherman. 2 Public Law 111-31, 123 Stat. 1776, June 22, 2009. Codified at 21 U.S.C. § 301.

Altria Client Services LLC 2325 Bells Road Richmond, Virginia 23234 (804) 335-2879 [email protected] The Notice invites comments on, among other things, "ways to enhance the quality, utility, and clarity of the information to be collectecr as part of FDA's research into graphic warnings.34 We believe there are ways FDA can more effectively meet the Notice's stated objective of collecting data regarding exposure to the graphic health warnings relative to the text-only Surgeon General's warnings.5 Our suggestions are as follows:

Study Design & Methodology

• A new graphic alone would be expected to increase noticeability, but may not impact understanding. If FDA is introducing both new graphics and new warning statements simultaneously, a different study design should be considered to assess whether exposure to the graphic warning "would promote greater public understanding."6 Graphic warnings should be compared relative to the new text-only warning statement, not the current, familiar text-only Surgeon General's warnings. Within this study there should be test conditions where the new information shown as a text only warning statement would serve as the internal controls for whether the pictorial aspects "provide new information, increase self-reported learning, or change beliefs about the negative health consequences of cigarette smoking."7

• Certain demographic (e.g., age, socioeconomic status) and other (e.g., nicotine dependence among smokers) factors may impact whether "the warning was perceived to be a fact or an opinioe8 and "if the warning made [participants] think about the health risks of cigarette smoking."9 Potential moderating factors should be thoroughly evaluated during the course of this study.

• When developing this study, FDA should consider previous research which has shown that use of graphical warnings can produce an opposite effect to the desired outcome, consistent with behavioral theory (i.e., exposure to aversive stimuli can produce an avoidance reaction).1°

3 The Act amended the Federal Cigarette Labeling and Advertising Act to require the Secretary of Health and Human Services to issue regulations for "Graphic Label Statements." See TCA § 201. A federal court found that the resulting regulations, issued the Food and Drug, violated the First Amendment. The research described in this Notice is intended to support a new rulemaking concerning Section 201's graphic warning requirements. See (76 Fed. Reg. 36,628, Jun. 22, 2011). See also R.J. Reynolds Tobacco Co. v. U.S. Food & Drug Admin., 823 F. Supp. 2d 36, 53 (D.D.C. 2011). See also R.J. Reynolds Tobacco Co. v. U.S. Food & Drug Admin., 845 F. Supp. 2d 266, 277 (D.D.C. 2012). 4 83 Fed. Reg. 48,626 (Sept. 26, 2018). 5 Id at 48,627. 6 Id at 48,626. 7 Id at 48,627. 8 Id 9 Id 10 Hall, M. G., Sheeran, P., Noar, S. M., Ribisl, K. M., Bach, L. E., & Brewer, N. T. (2016). Reactance to Health Warnings Scale: Development and Validation. Ann Behav Med, 50(5), 736-750. doi:10.1007/s12160-016-9799-3, Hardcastle, S. J., Chan, D. C., Caudwell, K. M., Sultan, S., Cranwell, J., Chatzisarantis, N. L., & Hagger, M. S. (2016). Larger and More Prominent Graphic Health Warnings on Plain-Packaged Tobacco Products and Avoidant Responses in Current Smokers: a Qualitative Study. Int J Behav Med, 23(1), 94-101. doi:10.1007/s12529-015-9487- x, Moodie, C., Bauld, L., Ford, A., & Mackintosh, A. M. (2014). Young women smokers response to using plain

2 • There is ample research indicating that the general public already believes that cigarette smoking is extremely harmful,11 which would suggest that exposure to graphical warnings will not meaningfully increase risk perceptions. An analysis conducted by ALCS on PATH Waves 1 (2013-2014), 2 (2014-2015), and 3 (2015-2016) datal2 shows that at least 90% of all adult respondents in each wave believe that cigarettes are at least "very harmful" to health. The study design should include pre/post measures of risk perceptions to evaluate whether the graphic warnings meaningfully increase likely pre- existing high levels of incoming risk perceptions.

Outcomes Measurement

• FDA should consider assessing comprehension of the new warnings objectively (i.e., evaluating recall of specific content, evaluating comprehension of disease risk) rather than participants indicating only that they learned (i.e., "self-reported learning from the warnine13). FDA should prioritize measuring the impact of the warning on behavior (e.g., quit intentions among cigarette smokers, initiation intentions among non-users) over concepts such as the evaluating whether the warning is "informative"14 and "grabbed...attention."15

• FDA has not shared the self-report scales used to measure the primary outcomes, making it impossible for ALCS to determine the validity, reliability, and sensitivity of the self- report scales.

We appreciate the opportunity to provide input on this Notice and would be happy to discuss our suggestions in more detail. If you have any questions, please contact me at (804)335-2879.

Sincerely, }-nmAnt A, r,C, A-A-V1V

cigarette packaging: qualitative findings from a naturalistic study. BMC Public Health, 14, 812. doi:10.1186/1471- 2458-14-812 II Foundation for a Smoke-Free World. (2018). The State of Smoking 2018: Global Survey Findings and Insights. Retrieved from https://www.smokefreeworld.org/sites/default/files/uploads/derek-yach-press-conference- presentation.pdf, Popova, L., Owusu, D., Weaver, S. R., Kemp, C. B., Mertz, C. K., Pechacek, T. F., & Slovic, P. (2018). Affect, risk perception, and the use of cigarettes and e-cigarettes: a population study of U.S. adults. BMC Public Health, 18(1), 395. doi:10.1186/s12889-018-5306-z, Song, A. V., Morrell, H. E., Cornell, J. L., Ramos, M. E., Biehl, M., Kropp, R. Y., & Halpern-Felsher, B. L. (2009). Perceptions of smoking-related risks and benefits as redictors of adolescent smoking initiation. Am J Public Health, 99(3), 487-492. doi:10.2105/AJPH.2008.137679 P2 ALCS internal analysis. 13 83 Fed. Reg. 48,627. 14 Id. 15 Id

3