FDA's Proposed Revocation of Soy Protein Health Claim Rests On Demanding Review of Selected Studies, Provides No Constitutional Analysis or Assessment of Enforcement Consequences

First Amendment Concerns

The preamble does not provide a constitutional analysis, despite the speech restrictions the final rule would impose and the questionable premise of the agency’s approach to evaluating the scientific data relating to the health claim. In particular, rather than evaluating all relevant sources as part of the “totality” of evidence that might be suggestive of a benefit, FDA completely disregards any source that FDA believes would not support a definitive conclusion. In principle, it could be possible for FDA to justify its approach to regulating health claims in food labeling in a restrictive manner, but the omission of an assessment of First Amendment limitations means affected manufacturers are unable to evaluate that justification for themselves.

Statutory Concerns

FDA appears to have departed from the statutory SSA standard, which directs the agency to adopt a “totality” approach to health claims substantiation and to consider the views of “experts qualified by scientific training and experience to evaluate such claims.”¹ The statutory language indicates that FDA should authorize a health claim according to a flexible standard that accounts for a wide range of evidence.² The proposed rule, however, interprets SSA more narrowly.

For animal and in vitro studies, for example, the preamble states, “we cannot draw any scientific conclusions” because “[s]uch studies cannot mimic the normal human physiology that may be involved in the risk reduction of CHD, nor can the studies mimic the human body’s response to the consumption of soy protein.”³ Like many other types of relevant scientific information, animal and in vitro studies are deemed categorically inadequate to substantiate a health claim and are disregarded entirely in the preamble’s substantiation analysis. The preamble asserts that “only well-designed, well-conducted human studies provide both the level of scientific rigor and generalizability to human populations needed to potentially support a health claim relationship.” FDA further limits the range of potentially supportive sources by reviewing only published sources⁴ and carefully defining the disease endpoints that the agency will use in identifying CHD risk reduction for the purposes of a health claim evaluation.⁵

SSA Or “Substantial Evidence”

In key respects, the SSA approach described in the preamble accompanying the proposed rule resembles the “substantial evidence” definition applied by FDA in the new drug context — both modes of analysis credit almost exclusively randomized, controlled trials (RCTs) and essentially ignore both non-RCT data sources and the perspectives of qualified experts. In the preamble, for example, FDA dismisses any source that the agency believes would not properly inform a scientific conclusion, without regard to whether, in fact, that source has informed such a conclusion.⁶ The preamble even indicates that government reports and other documents describing expert opinion must be completely eliminated from the agency’s analysis, as these sources do not typically include “detailed information on studies” and therefore FDA cannot “determine[e] whether the studies are flawed . . . .”⁷

Pursuant to FDA’s commitment to a “strong” interpretation of SSA,⁸ in reevaluating the health claim, FDA initially identified 709 publications from several sources, including the agency’s own literature review and submissions from stakeholders in response to an earlier notice.⁹ After eliminating any publication that FDA had found flawed, the agency focused its re-evaluation on 58 well-designed, well-executed intervention studies, which, according to FDA, revealed a high degree of inconsistency in their findings. Of the 58 studies, 19 showed a benefit in significantly reducing the risk of CHD, and four showed a benefit in lowering systolic or diastolic blood pressure with soy protein consumption; the remaining 35 studies did not show a benefit, either in significantly reducing the risk of CHD or in lowering systolic or diastolic blood pressure with soy protein consumption. FDA estimates that its conclusion, if finalized as proposed, would result in the elimination of the health claim from the labeling of 200-300 currently marketed food products,¹⁰ though the economic analysis in the preamble does not attempt to quantify the costs associated with lost sales.¹¹

If the proposal becomes effective, food manufacturers will be prohibited from using the health claim in labeling starting 30 days after the final rule’s date of publication in the Federal Register. Two FDA compliance policies are intended to ameliorate the effects of the new final rule.

First, FDA has said that manufacturers will have the option to use a qualified health claim about the relationship between soy protein and a reduced CHD risk if the proposed rule becomes final. According to the preamble, “while the totality of the publicly available scientific evidence does not support a finding of SSA, if, when we finalize this rule, we conclude there is no SSA but there is some credible evidence for the use of a qualified health claim about the relationship between soy protein and a reduced risk of CHD, we intend to issue a statement of enforcement discretion for the use of a qualified health claim.”¹² State-law claims are not necessarily preempted by federal law in the presence of an enforcement discretion letter authorizing a qualified health claim — meaning manufacturers using the qualified claim are subject to enforcement by state attorneys general and others relying on state consumer protection and other provisions.¹³ The proposed rule includes the required federalism discussion but does not address this dimension of the agency’s approach.

Second, FDA has said that manufacturers “will be allowed to keep the current authorized claim on their products” while the rulemaking proceeding is underway. This statement appears only in a public statement issued separately from the proposed rule itself.¹⁴ It is unclear whether such a non-binding FDA statement would operate as an effective litigation defense, particularly in view of the preamble analysis, which makes clear the agency’s view that the relevant scientific publications do not support the continued use of the claim.

¹21 U.S.C. § 343(r)(3).

² In its 2009 guidance, FDA stated that application of the SSA standard was intended to be “flexible, in recognizing the variability in the amount and type of data needed to support the validity of different substance/disease relationships.” Guidance § III.G.

³ 82 Fed. Reg. 50324, 50329 (Oct. 31, 2017).

⁴ Id. at 50327. FDA had previously said in guidance that, in evaluating whether scientific data reaches the level of SSA, it would not evaluate data and information that are not “available to the relevant scientific community.” Guidance § III.G.

⁵ 82 Fed. Reg. at 50329.

⁶ Id. at 50337 (describing SSA in terms of what “experts would find” rather than what experts have found).

⁷ Id. at 50329.

⁸ Id. at 50337.

⁹ Id. at 50329.

¹⁰ Id. at 50326.

¹¹ Id. at 50338. FDA does, however, recognize relabeling costs, as well as benefits associated with the enhanced enjoyment of foods by consumers who switch to products that they enjoy more or that have an authorized health claim in labeling. The economic analysis does not mention the benefits of CHD reduction from the consumption of foods containing soy protein.

¹² Id. at 50337. The FDA materials issued along with the proposed rule include conclusory statements describing the scientific data relating to the relationship between soy protein and CHD risk that are of the type that FDA might be willing to permit as qualified health claims, e.g., “While some evidence continues to suggest a relationship between soy protein and a reduced risk of heart disease — including evidence reviewed by the FDA when the claim was authorized — the totality of currently available scientific evidence calls into question the certainty of this relationship,” FDA, Statement from Susan Mayne, Ph.D. on Proposal to Revoke Health Claim That Soy Protein Reduces Risk of Heart Disease (Oct. 30, 2017), https://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm582744.htm.

¹³ See Reid v. Johnson & Johnson, 780 F.3d 952 (9th Cir. 2015).

¹⁴ See FDA, Statement from Susan Mayne.