The FDA’s current guidance on evidence-based review for the science needed to support food and dietary supplement health claims was published in 2009. Although this guidance has not changed in the past decade, as demonstrated by two recent agency decisions regarding health claims, the agency’s interpretation of the guidance appears to be changing. The article reviews these two cases—revocation of the soy protein and coronary heart disease health claim and the conclusion that there is no credible evidence to support the vitamin D and decreased risk of multiple sclerosis health claim petition. The authors provide insight into how and which scientific evidence for health claims is being assessed today.
The US Food and Drug Administration (FDA) regulates the type of claims able to be made for dietary supplement products and with respect to communicating health benefits, permits structure function claims and health claims on these products. Structure function claims describe the role of a nutrient or dietary ingredient found in the supplement on the normal structure or function of the human body and do not require preauthorization by FDA. Health claims, first authorized through the Nutrition Labeling and Education Act of 1990 (NLEA) require preauthorization by FDA, describe a causal relationship between a substance and a reduction in the risk of a disease or health-related condition (e.g., cardiovascular disease, hypertension) for the general US population or a designated subpopulation. Specifically, FDA regulations define a health claim as “any claim made on the label or in labeling of a food, including a dietary supplement, that expressly or by implication, including ‘third party’ references, written statements, or vignettes, characterizes the relationship of any substance to a disease or health-related condition.”1 FDA intended health claims to help consumers make healthy food choices by providing consistent, science-based messaging on the relationship between a substance (food or component of food) and a disease or health-related condition. To support this goal under the NLEA, health claims must be based on Significant Scientific Agreement (SSA). This standard requires FDA, before issuing a health claim, to determine “based on the totality of publicly available scientific evidence (including evidence from well-designed studies conducted in a manner which is consistent with generally recognized scientific procedures and principles), that there is significant scientific agreement, among experts qualified by scientific training and experience to evaluate such claims, that the claim is supported by such evidence.”2
FDA recently published negative opinions on two health claims relevant to the nutrition and supplement industries. In one case, FDA was deliberating on a new petition (vitamin D and multiple sclerosis), and in the other, the agency took the unusual step of acting on an existing health claim (soy protein and heart health). Overall, the definition of a health claim has not changed, nor has the requirement for SSA, but as these two recent health claim denials will show, the level of detail needed to be considered in health claim petitions today are far greater than in the past. As will be discussed in this article, these cases demonstrate how the landscape for scientific evidence assessment has dramatically shifted.
The Case of Soy Protein and Reduced Risk of Coronary Heart Disease
FDA published the final rule authorizing a health claim for soy protein and reduced risk of heart disease in 1999.3 In this rule, FDA indicated that, after a totality of the evidence review, according to experts, Significant Scientific Agreement (SSA) for a link between soy protein consumption and decreased risk of heart disease existed. The final rule allowed companies to make the claim that products containing a certain level of soy protein among other nutritional factors that the product must meet “may reduce the risk of heart disease.” In October 2017, however, FDA published a proposed rule to revoke this health claim, indicating that, based on current guidelines, the totality of the evidence no longer demonstrated that the SSA standard had been met.4 Further, the SSA standard is intended only to be met when the validity of the relationship between a substance and disease is not likely to be reversed by new and evolving science. So, between 1999 and 2017, what changed?
The strongest evidence to support a health claim comes from human intervention studies. In 1999, FDA based the soy protein authorized health claim on review of 13 published intervention studies representing 14 clinical trials on soy protein and LDL-cholesterol and/or Total-Cholesterol (TC), which were considered biomarkers for heart disease risk. In its 2017 updated review, the agency included 58 human intervention studies, which included 46 intervention studies that measured blood TC or LDL-cholesterol and 12 that measured Systolic Blood Pressure (SBP) or Diastolic Blood Pressure (DBP). A much larger body of evidence was reviewed in 2017 and FDA then indicated that the additional evidence included a number of new studies that were inconclusive or did not support a relationship between soy protein and heart disease risk.
However, the approach to assessing evidence also has changed dramatically in the scientific community in general over the past decade, and the new studies were not the only difference in this re-evaluation of the evidence.
Reflective of the scientific community’s updated approach to assessing evidence, FDA also had updated its scientific evaluation standards between 1999 and 2009. In 1999, FDA published the first guidance for industry on the review of science against the SSA standard, providing steps to identify the relevant science, determine the quality of the studies and then assess the strength of the evidence as a whole. FDA described the process as a consistent and objective approach for scientific reviews and included consideration of study type (e.g., intervention, prospective cohort, case-control, cross-sectional), methodological quality, quantity of evidence (number of various types of studies and sample sizes), and whether the evidence could be generalized to the US population or a target subgroup. Assessment of the totality of the evidence also was described in this guidance, which included replication and the overall consistency of the total body of evidence. The field of evidence assessment was relatively new at the time and prior approaches in nutrition science had relied on expert panels rather than objective reviews. Therefore, scientific agreement on the standards and methods for objectively quantifying the amount of evidence, particularly for rating quality, had not yet been developed.
Since 1999, the fields of evidence-based medicine and evidence-based nutrition evolved and in 2009, FDA replaced its 1999 guidance with a new guidance on the assessment of evidence.5 Of the 13 publications included in FDA’s 1999 review of the soy protein health claim, only three met the new quality standards. Eight were excluded due to inappropriate controls and two were excluded because the statistics were not conducted between the control group and the soy protein intervention group. Thus, although the majority of studies included in FDA’s 1999 review appeared to meet the evidence standards at a macro level, the 2017 re-evaluation revealed that specific aspects of the studies’ design and implementation (e.g., control, statistics, etc.) did not meet the newer standards. As such, the past, mainly positive, studies were not considered in the updated assessment. Based on their re-evaluation, FDA determined that SSA does not exist among experts for the consumption of soy protein and decreased risk of heart disease.
Qualified Health Claims as an Option
Qualified health claims requiring a lower scientific standard of evidence than an authorized health claim are another option open to companies looking to make claims characterizing the relationship between a nutrient and a disease risk or health-related condition. Qualified health claims developed in response to a 1999 court decision regarding the limits the First Amendment places on FDA’s ability to prohibit health claims. Specifically, in Pearson v. Shalala, the court determined that, under the First Amendment, FDA is not permitted to prohibit the use of health claims unless the agency reasonably determines that there is no disclosure that can explain the limits on scientific evidence used to develop the claim and eliminate potential consumer deception as to the strength of this evidence.6 Known as “qualified health claims,” FDA allows these claims where emerging scientific evidence demonstrates a relationship between a substance and a disease or health-related condition, but such evidence does not rise to the level of the SSA standard needed for unqualified health claims. FDA authorizes these types of claims by issuing a letter of enforcement discretion indicating that they will not take action against an entity for violations of the Food, Drug and Cosmetic Act for making the claim, provided the claim is qualified to indicate the weight of the evidence and any scientific uncertainty. FDA includes language in these letters specifying the language a company should use to qualify a claim.
FDA guidance regarding qualified health claims indicates that claim language should reflect the level of scientific evidence with specificity and accuracy, such as noting the limited number or nature of studies that support a claim, and whether a substance plays an independent role in the relationship to the disease and/or requires the use or removal of another substance to provide benefits.7 FDA has published 18 qualified health claims subject to enforcement discretion, with seven of these on a food or nutrient decreasing the risk of cardiovascular disease.8 However, a challenge with qualified health claims is that the language of the claim can be cumbersome and not necessarily seen as positive by consumers. For example, for a recent qualified health claim discussing the link between cardiovascular disease and corn oil, FDA specified the following language should be used:
“Very limited and preliminary scientific evidence suggests that eating about 1 tablespoon (16 grams) of corn oil daily may reduce the risk of heart disease due to the unsaturated fat content in corn oil. FDA concludes that there is little scientific evidence supporting this claim. To achieve this possible benefit, corn oil is to replace a similar amount of saturated fat and not increase the total number of calories you eat in a day. One serving of this product contains [x] grams of corn oil.”9
Currently, FDA is considering the evidence for consumption of soy protein and reduced risk of cardiovascular disease for a potential qualified health claim. The comment period for this claim petition (FDA-2017-N-0763) closed in March 2018, and more than 1,200 comments were received by FDA. Due to high interest, it is anticipated FDA will within the next year provide some response on a qualified health claim for soy protein consumption and reduced risk of cardiovascular disease.
The Case of Vitamin D and Decreased Risk of Multiple Sclerosis
A recent petition requesting that FDA develop a health claim for vitamin D and decreased risk of multiple sclerosis provides an example of situations in which FDA may find that no language is acceptable, even for a qualified health claim. The petitioner requested that FDA approve a claim that “consumption of vitamin D may reduce the risk of developing multiple sclerosis (MS), a chronic autoimmune disorder that damages the body’s central nervous system” and provided, what appeared to be on its face, a large body of evidence to support this claim. On 12 January 2017, however, FDA denied the claim.10 A careful review of FDA’s determination in this instance demonstrates how new approaches to assessing scientific evidence on a more granular level has changed the evidence likely needed to support a health claim.
The petition presented 85 scientific articles. FDA reviewed these publications as well as those identified from their own search. Of the 85 studies submitted in the petition, the vast majority were reviews, safety, animal or in vitro studies, or studies not related to MS. FDA will consider these types of studies as background or supportive information only, but not as primary evidence for a health claim. The petition also provided seven intervention studies that evaluated the relationship between vitamin D intake and MS and 22 observational studies, which could be considered as primary evidence in support of the claim.
For the intervention studies, FDA found that the studies evaluated the effect of vitamin D in individuals previously diagnosed with MS, ranging from early stage to late stage. Health claims are intended to highlight the reduced risk of a disease in people who do not have the disease that is the subject of the claim; therefore, FDA concluded these studies, which evaluated vitamin D’s effect on individuals already diagnosed with MS, did not meet the criteria to support a claim about the reduced risk of developing MS.
FDA also conducted a thorough review of the observational studies and found that 15 of these studies included people previously diagnosed with MS which, similar to the intervention studies, FDA concluded could not be used to support a claim about reduced disease risk. Of the seven remaining studies, FDA reviewed the approach used in these studies to quantify vitamin D status, which included estimations from dietary intake assessments as well serum vitamin D levels, and found these methods included too many confounding variables to confidently quantify vitamin D levels.
Overall, the agency concluded:
“Through this rigorous assessment, the FDA determined that scientific conclusions could not be drawn about the relationship between vitamin D intake and MS risk. Based on the agency’s review of the totality of publicly available scientific evidence, the FDA has concluded that there is no credible evidence of a relationship between intake of vitamin D and a reduced risk of MS.”
This case represents one of the most difficult questions for developing evidence for a health claim for a food or dietary supplement—what is an appropriate biomarker for the endpoint? Health claims, whether unqualified or qualified, can only reflect the effect of the nutrient or dietary substance on the reduced risk of a disease, and not on the disease itself. The effect a substance may have in diagnosing, curing, mitigating, treating or preventing a disease is the purview of a drug, and such claims are not permitted to be made about foods and dietary supplements. Therefore, the endpoints in studies used as support for health claims cannot reflect treatment, prevention, mitigation or cure of a disease or condition. Thus, determining what reflects a decreased risk, compared to treatment or management of symptoms, can be difficult.
With regard to the vitamin D/MS health claim petition, FDA noted “no surrogate endpoints have been recognized for evaluating the risk of MS.” The petitioners suggested that Clinically Isolated Syndrome (CIS) could be used as a surrogate endpoint/biomarker, but FDA concluded this syndrome is not considered to be a surrogate endpoint/biomarker; rather, individuals who are diagnosed with CIS are considered to have MS. Therefore, the presence of CIS would be considered presence of MS. Any intervention to decrease or prevent CIS, based on this definition, would be considered mitigation of the disease and represent a drug effect, not a health claim effect regarding decreased risk. In evaluating health claim petitions for risk reduction, FDA considers evidence from studies in individuals already diagnosed with the disease only if it is scientifically appropriate to extrapolate to individuals who do not have the disease.
Further, in observational studies, the accuracy of biomarkers and assessment tools for dietary intake are currently in question. As noted in the vitamin D/MS petition, over the past decade several cases exist in which intervention studies have not found the same relationships as those identified in observational studies and inaccuracies in dietary intakes and nutrient status of the populations have been implicated as a major issue in observational studies. The long-accepted approach of dietary intake analysis for determining nutrient status is no longer considered a gold-standard and new scientifically validated approaches have not been forthcoming.
While FDA allows for qualified health claims or those that have credible science but do not meet the higher standard of SSA required for an authorized health claim, FDA indicated in this case no qualifying disclaimer would be sufficient to characterize the evidence. That is, FDA indicated there is no credible scientific evidence to support the claim
What can we learn from these two cases?
The approach to evaluating evidence in the nutrition and medical fields continues to evolve. The vitamin D/MS health claim petition contains significant insight into how FDA assesses observational studies. It also points out the need to assure the endpoints and biomarkers are the current accepted and validated approaches. This is a challenge in developing strategies for research intended to be used for health claim petitions. In their 2010 report on Biomarkers and Surrogate Markers for Chronic Disease, the Institute of Medicine of the National Academy of Sciences (NAS) accepted few identified biomarkers for indicating a decreased risk of diseases.11 However, the report did indicate the area of biomarker validation and application is dynamic, and promoted the use of an expert panel to assess not only the validation of the biomarker, but also the context of its use, in application to health claims. Therefore, one approach to navigating this area may be to develop an expert panel report on the appropriate biomarkers and endpoints before conducting the scientific assessment for a health claim.
Most recently, the NAS Health and Medicine Division published a report on Guiding Principles for Developing Dietary Reference Intakes Based on Chronic Disease12 providing recommendations for assessment of the totality of evidence applicable to making public health recommendations. Although not directly about health claims, this report reflects the current scientific consensus for the methods that should be used to evaluate a body of evidence for a substance and health outcome relationship, and which should be considered to help understand how FDA or other agencies may assess a body of evidence. Finally, in assessing the evidence, it is not enough to simply look at the type of studies and how many are available. Evidence assessment has evolved to requiring evaluators to review in much more detail how the different steps in the studies were conducted.
- 21 CFR 101.14(a).
- 21 USC 343(r)(3)(B)(i). FDA adopted this same standard by regulation for dietary supplement health claims. See 21 CFR 101.14(c).
- 21 CFR 101.82.
- Food Labeling: Health Claims; Soy Protein and Coronary Health Disease. Proposed Rule. 31 October 2017. https://www.federalregister.gov/documents/2017/10/31/2017-23629/food-labeling-health-claims-soy-protein-and-coronary-heart-disease. Accessed 11 July 2018.
- FDA Guidance for Industry: Evidence-Based Review System for the Scientific Evaluation of Health Claims. January 2009. FDA website. https://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/ucm073332.htm. Accessed 11 July 2018.
- Pearson v. Shalala, 164 F.3d 650 (DC Cir. 1999).
- Op cit 5.
- Summary of Qualified Health Claims Subject to Enforcement Discretion. FDA website. https://www.fda.gov/Food/LabelingNutrition/ucm073992.htm. Accessed 11 July 2018.
- Qualified Health Claims: Letter of Enforcement Discretion—Corn Oil and Corn Oil-Containing Products and a Reduced Risk of Heart Disease (Docket No. 2006P-0243). http://wayback.archive-it.org/7993/20171114183735/https://www.fda.gov/Food/IngredientsPackagingLabeling/LabelingNutrition/ucm072956.htm. Accessed 11 July 2018.
- FDA Response, Qualified Health Claim Petition—Vitamin D and Reduction of the Risk of Multiple Sclerosis (Docket No. FDA-2016-Q-2646). FDA website. https://www.fda.gov/downloads/Food/LabelingNutrition/UCM592302.pdf. Accessed 11 July 2018.
- Biomarkers and Surrogate Endpoints in Chronic Disease (2010). https://www.nap.edu/catalog/12869/evaluation-of-biomarkers-and-surrogate-endpoints-in-chronic-disease. Accessed 11 July 2018.
- Developing Dietary Reference Intakes Based on Chronic Disease. 3 August 2017. http://nationalacademies.org/hmd/Reports/2017/guiding-principles-for-developing-dietary-reference-intakes-based-on-chronic-disease.aspx. Accessed 11 July 2018.
About the Authors
DeAnn Liska, PhD, is the senior director of nutrition science and biostatistics at Biofortis Research, a division of Mérieux NutriSciences, where she leads the team responsible for scientific consultation and design and interpretation of clinical trials. Liska has more than 20 years of experience in leadership roles in the food and nutrition industry. She may be contacted at email@example.com.
Megan Olsen, JD, is the assistant general counsel for the Council for Responsible Nutrition (CRN) in Washington, DC. At CRN, she provides legal counsel and advice to CRN’s staff and members in the areas of legislation, regulatory compliance and advocacy, and international policy development with respect to dietary supplement and nutrition issues. Prior to joining CRN, Olsen was in-house counsel for Walgreen Co., where she provided legal advice about FDA, Federal Trade Commission and other consumer protection regulatory requirements for a wide-variety of consumer products, including conventional food, dietary supplements, OTCs and cosmetics. Olsen began her career at Kelley, Drye and Warren, LLP working on a variety of consumer protection, regulatory and advertising law issues. She may be contacted at firstname.lastname@example.org.
Cite as: Liska D and Olsen M. “Evaluating Science for Health Claims: An Evolving Landscape.” Regulatory Focus. July 2018. Regulatory Affairs Professionals Society.