What evidence for what health claims?

When DHA was first added to infant formula, the pitch to investors was that the product “may help close gaps researchers have found between the development of breast-fed and bottle-fed infants [14].” However, most of the research on DHA and ARA additives focuses on comparing infant formulas with different additives. Little of the research looks into how infant formulas with and without DHA compare with breastfeeding in terms of health effects. It might be that infant formula with DHA produces somewhat better results than infant formula without that additive, but if both produce far worse results than breastfeeding, that would be important to know. Parents need to make well informed choices not only in deciding which infant formula to use but also in deciding whether to use formula rather than breastfeed.

Using newer versions of the additive might have better health effects than older versions, and thus bring the product somewhat closer to the performance of breastfeeding. However, claiming that any infant formula with the fatty acid is “closest to human milk” or “as close as possible to human milk” is not the same as saying it is close to human milk. The claim may be misleading if the difference between “closer to” and “close to” escapes people’s notice.

When the manufacturers make claims about the health effects of their DHA and ARA that are added to infant formula, they pick out studies that support their views, and ignore studies that do not. They do not clearly identify the specific studies that supposedly support their specific claims. There is no clear consensus about what indicators should be used for the major claims.

To illustrate, one manufacturer’s website said:

Certain studies have also shown that infants fed formula supplemented with DHA and ARA exhibited:

Preterm infants who were fed DHA and ARA supplemented formula exhibited normal growth in terms of weight, length and head circumference and improved visual and mental development [15].

The manufacturer listed 28 studies said to be “Research Supporting the Importance of DHA & ARA in Infant Development,” with an additional seven listed as “Research Supporting the Importance of DHA for Pre-term and Near-term Infants.” It is not clear which of the studies is supposed to support which of the specific claims. Many were published prior to 2000, and thus have little relevance to current versions of manufactured DHA and ARA. Many of the studies listed have no relevance to any of the claims.

The list did not mention the many studies that found no significant health benefit from fatty acids added to infant formula. It did not mention that several major meta-analyses found no significant effect [16–18].

(The article by Qawasmi and colleagues published online by Pediatrics in June 2012 [16], concluded, “LCPUFA supplementation of infant formulas failed to show any significant effect on improving early infant cognition”. It was followed by another online article with a similar title in December 2012, this time concluding, “Current evidence suggests that LCPUFA supplementation of infant formulas improves infants’ visual acuity up to 12 months of age.” The journal published my letter inquiring about the reasons for this change, but there has been no reply. Their second article and my letter are available at http://pediatrics.aappublications.org/content/early/2012/12/12/peds.2012-0517.abstract/reply#pediatrics_el_54993. A third and final version of the article appeared in print in Pediatrics in January 2013, described at http://www.ncbi.nlm.nih.gov/pubmed/23248232. Its conclusion was the same as that in the December 2012 version. Apparently, seven studies were added to the set of twelve originally reviewed, leading to the change in the conclusion.)

The manufacturer said only that “certain studies” showed the effects that were claimed, but they did not specify which ones. There are also studies that fail to show those effects. The manufacturer has not demonstrated that the dominant trend in the relevant studies supports their claims.

Another page on the same website discussed “The Benefits of DHA and ARA During Infancy” under various categories: cognition, vision, immune response, etc. For each category, the document said there were studies that supported their claims regarding these benefits. At the end there was a link to an obscure list of references, with no indication of which were supposed to support which claims [19].

The manufacturer said, “Clinical studies have demonstrated numerous benefits for infants receiving DHA and ARA supplemented formula, including improved mental and visual development [20–22],” but it did not say which studies.

The manufacturer did not say which, if any, of the studies it listed were specifically about its currently featured product, trademarked as life’s DHA. For advertising purposes the manufacturer apparently wants life’s DHA to be viewed as very different from earlier versions of the manufactured fatty acid, but at the same time it suggests that earlier research based on other versions of DHA and the regulations on them should apply just as well to this latest version [23].

There is no consensus as to what indicators would be appropriate for assessing each of the different types of health claims made for fatty acids added to infant formula. Questions have been raised about the Bayley Scales of Infant Development (BSID) used in the meta-analysis by Qawasmi and colleagues, and in many other studies as well [16, 24]. The critics argue that these scales, used primarily for children up to 18 months of age, are not appropriate for assessing how these ingredients affect physical and mental development in the long run.

The FDA has expressed its doubts about manufacturers’ health claims:

16. What is the evidence that addition of DHA and ARA to infant formulas is beneficial?

The scientific evidence is mixed. Some studies in infants suggest that including these fatty acids in infant formulas may have positive effects on visual function and neural development over the short term. Other studies in infants do not confirm these benefits. There are no currently available published reports from clinical studies that address whether any long-term beneficial effects exist [25].

Which DHA?

When a manufacturer listed “Research Supporting the Importance of DHA & ARA in Infant Development,” it did not say all these studies would show the importance of manufactured DHA and ARA. The manufacturer implicitly assumed that evidence relating to human DHA and ARA as found in breast-milk would apply just as well to their manufactured versions. However, the manufacturer’s claims were specifically about “formula supplemented with DHA and ARA,” meaning the manufactured versions.

DHA in breast-milk tested in a chemistry laboratory might appear to be similar to a manufactured DHA. However, the real test of equivalence is whether they function in the same way. Do they lead to the same results for infants on relevant health indicators? The health effect of the two should not be claimed to be the same unless their performance is demonstrated to be the same [26].

It is also important to distinguish among various types of manufactured DHA. U.S. Pharmacopeia’s Food Chemicals Codex distinguish three types of algae-based DHA, corresponding to three different algal oils: schizochytrium, crypthecodinium, and ulkenia.

There are other DHAs that are not in the Food Chemicals Codex. For example, there is an egg-based DHA:

Baby’s Only Essentials® DHA & ARA Fatty Acid is naturally derived from the goodness of Egg Phospholipids using an aqueous (water) process. This differs from C. cohnii oil (algae) & M. alpina oil (fungus) used in all other organic and conventional infant formulas, which are treated with hexane solvent, acid, and bleach [27].

It is not clear how the different types of manufactured DHA compare with regard to their safety or their effectiveness. If the newer ones are said to be improvements, that implicitly acknowledges that the earlier ones must have been deficient.

Regulations regarding health claims

U.S. rules regarding health claims have been summarized as follows:

. . . . the Nutrition Labeling and Education Act . . . . set a high threshold for the scientific standard under which the U.S. Food and Drug Administration (FDA) may authorize health claims, this standard is known as the significant scientific agreement (SSA) standard. Subsequent legislation known as the Food and Drug Administration Modernization Act (FDAMA) provided an alternative to FDA review of the health claim where an U.S. government scientific body other than FDA concluded that there is SSA for a substance/disease relationship ([28], also see [29]).

Comparable regulations of health claims came into force for the European Union in January 2007 [30].

Following the practice in a number of other countries, in 2001 the FDA accepted the addition of manufactured DHA to infant formulas. However, the agency has not affirmed that this addition is beneficial, and it has not explicitly authorized the manufacturers to make claims regarding their benefits.

The FDA explains its role as follows:

4. Does FDA approve infant formulas before they are marketed?

No, FDA does not approve infant formulas before they can be marketed. However, all formulas marketed in the United States must meet federal nutrient requirements and infant formula manufacturers must notify the FDA prior to marketing a new formula. If an infant formula manufacturer does not provide the elements and assurances required in the notification for a new or reformulated infant formula, the formula is defined as adulterated under Section 412(a)(1) of the FFDCA and FDA has the authority to take compliance action if the new infant formula is marketed ([25], also see [31]).

From time to time FDA responds to petitions regarding health claims. There is a good illustration of how such petitions are handled in FDA’s response to a petition regarding a health claim for whey-protein partially hydrolyzed infant formula [32].

The FDA has not endorsed the manufacturers’ claims regarding the health benefits of adding DHA to infant formula. It may authorize health claims, under the standard known as significant scientific agreement (SSA). The law does not say the FDA must authorize health claims. There is no published SSA for DHA in infant formula [33].

The Infant Formula Council, a lobbying group for the manufacturers, said, “The use of LCPUFAs [long chain polyunsaturated fatty acids] in infant formulas has been reviewed and supported by the U.S. Food and Drug Administration [34].” Given FDA’s explanation that it does not approve infant formula and its statement that the evidence regarding the benefits of DHA and ARA in infant formula is mixed, it is difficult to guess what the Council might have meant by “reviewed and supported.”

There is little effective regulation with regard to health claims or the general nutritional adequacy of infant formula and its additives. The regulators regularly accept the manufacturers’ claims that the currently proposed versions are superior, but do not explore how the previous ones might have been deficient.

Current regulations for infant formula focus on ensuring that particular components are supplied in specified quantities, based on an ingredients list from the 1980s. They imply that any formula that includes the specified ingredients in the required amounts is safe and nutritionally adequate. However, breast-milk is a complex, changing, living thing, and not simply a static collection of inert ingredients. Many ingredients in breast-milk are absent in infant formula, or are provided in infant formula in forms that are different from those in breast-milk. There is little recognition of the fact that ingredients can act differently in different contexts.

It should not be assumed that DHA additives that are manufactured by one process have the same health effects as DHA manufactured by another process. Similarly, it should not be assumed that the health effects of manufactured DHA in infant formula are the same as the health effects of natural DHA in breast-milk. These propositions need to be verified empirically.

GRAS determination

The FDA’s Guidelines Concerning Notification and Testing of Infant Formulas say:

A manufacturer must notify FDA 90 days before the first processing of any infant formula for commercial or charitable distribution for human consumption that differs fundamentally in processing or in composition from any previous formulation produced by the manufacturer [35].

These incremental changes are handled by determining whether the proposed additives are GRAS, Generally Regarded as Safe. It is not the FDA itself that assesses their safety. The manufacturer is supposed to assess their safety and notify the FDA of its findings. To illustrate, in 2001 the FDA responded to requests from manufacturers to have their versions of these fatty acids, ARASCO and DHASCO, characterized as GRAS. In its response, the FDA said:

The agency has not, however, made its own determination regarding the GRAS status of the subject use of ARASCO and DHASCO [36].

The FDA had no questions about the documents submitted to it about the companies’ requests for GRAS status for these products, and it acknowledged that the FDA had not made its own independent determination regarding their status. The FDA did not say whether it agreed with the requests.

The FDA’s Inventory of GRAS Notices shows that the agency records the details of the notifications regarding GRAS that it receives [37]. The FDA does not affirm that the product is safe.

The GRAS determination process is explained in FDA’s Frequently Asked Questions about GRAS[38]. The answer to question 10 is of particular interest:

Does FDA currently have a program to affirm that one or more uses of a food substance are GRAS?

In a proposed rule that FDA published in 1997 (62 Fed. Reg. 18938; April 17, 1997), FDA explained why the agency could no longer devote resources to the voluntary GRAS affirmation petition process that is described in 21 CFR 170.35(c) and proposed to abolish that process and replace it with a notification procedure. The agency has not yet issued a final rule however, and the petition procedure remains in the agency’s regulations. However, at this time FDA is not committing resources to the review of GRAS affirmation petitions.

In other words, the FDA does not evaluate the GRAS petitions it receives.

The question is not whether these optional ingredients are in fact safe to use. The GRAS question is about whether they can reasonably be assumed to be safe. The distinction is important. To illustrate, one could be unsure whether infant formula made with genetically modified soy is safe for infants, but at the same time be certain that it should not be assumed to be safe.

The GRAS concept was developed to avoid having to scientifically assess familiar products that were known to be safe on the basis of extensive experience in diverse conditions. In the case of novel products such as manufactured DHA and ARA intended for highly vulnerable infants, the designation seems unreasonable. Moreover, depending on the manufacturers of the products to make the judgment about whether the product should be considered GRAS seems unreasonable. Children should not be made to bear the risk entailed in making that assumption.

Postmarket surveillance

The FDA’s acceptance of the GRAS notification allows manufacturers to begin marketing the product. However, in its letters to the manufacturers, the FDA said:

FDA would expect any infant formula manufacturer who lawfully markets infant formula containing ARASCO and DHASCO to monitor, through scientific studies and rigorous post-market surveillance, infants who consume such a formula. We also would expect regular reports of such studies and post-market surveillance. Because the use of ARASCO and DHASCO in infant formula would be based on the GRAS provision of the FFDCA, we also would expect that these reports would not be considered to be confidential so that the broader scientific community can contribute to this continuing evaluation [36, 39].

Thus, there was an expectation of surveillance of the products’ safety and health effects after marketing was underway.

In response to an inquiry from the Cornucopia Institute in 2009, the FDA said it had received no postmarket surveillance reports on the safety and effectiveness of these products. Nevertheless, the industry proceeded with large-scale advertising campaigns promoting infant formula and many other food products based on its claims about the benefits of these additives [40]. There is no indication that the manufacturer has ever done such studies [10, 41].

Numerous “Adverse Reaction Reports” that appear to be related to DHA/ARA in infant formula have been submitted to the FDA. The Cornucopia Institute has made some of them available on its website [42]. The adverse reactions described by mothers included gas, diarrhea, vomiting, bowel obstruction, gastric reflux, colic, and constipation.

It is unfortunate that FDA relies on the industry for post-market surveillance of the effects of optional ingredients because the value of the industry’s reports would be highly questionable. It is even more problematic that the FDA now seems to be indifferent to the failure of the industry to conduct such studies.