I wrote to an author of a couple of articles on this topic. My email:
Your 2014 Am J Clin Nutr and 2015 JPGN articles are being used to promote infant formulas in Asia and Australia. The marketing message is that your results show that the formulas are now closer than ever to breastmilk, as the outcomes measured are almost the same. It’s therefore important, in assessing the outcomes, to know just how the groups were defined. As you would know:
“As described (6), 160 formula-fed infants were recruited from March 2008 to February 2012 and randomized to receive EF or SF from inclusion until 6 months of age. After 6 months of age, no study formula (Experimental or Standard Formula) was distributed to the families who continued the feeding with support from their local baby health clinics….A breast-fed reference (BFR) group consisted of 80 infants. Inclusion criteria were <2 months of age, gestational age at birth 37 to 42 weeks, birth weight 2500 to 4500 g, absence of chronic illness, and exclusively formula-fed or, for the BFR group, exclusively breast-fed at inclusion, and mother having the intention to breast-feed (no formula) until 6 months. All of the groups were given a recommendation of only providing small amounts (taste portions) of complementary foods between 4 and 6 months."
So the breastfed infants were exclusively (solely) breastfed at time of entry, which could be up to two months old, and mothers were advised to give tastes of other foods between 4-6 months. The formula fed infants were recruited at up to two months old, and exclusively (solely) formula fed when included – also at up to two months of age; you note that some were breastfed for varying lengths of time, in the critical period of initial gut microbial colonisation.
The solely-breastfed-at-age-of-entry infants were not necessarily exclusively breastfed from birth to the point of entry, and then not even solely breastfed for more than perhaps two months before other foods were offered. (A “mother having the intention to breast-feed (no formula) until 6 months” does not ensure that she does so, and giving other foods affects breastmilk supply.) Both formula and other food exposures result in gut microbiome changes.
Similarly, given the positive community attitudes to breastfeeding in Scandinavia, and the high rate of breastfeeding discontinuation by 6 weeks in many countries, an exclusively formula-fed group recruited up to 2 months postpartum will always include children who have already had some breastmilk. The results section in the 2015 JPGN article states that those exclusively formula-fed at study entry had in fact received from 5-41 days of breastfeeding, with the mean in both groups being 23 days. In small groups of 80 or so, all three groups undoubtedly contained a significant number of mixed-fed infants. Working from basic science, this increases the similarity of outcomes.
Without meticulously accurate records, it is entirely possible in Sweden that ALL those defined as formula fed (FF) were given some colostrum and breastfeeding; while the group defined as breastfed (BFR) were given infant formula neonatally or in the process of getting breastmilk supply established. What care was taken to document hospital or other exposures to either breastmilk or infant formula? How many in each of these groups of 64-70 or so were in fact fed both breastmilk and formula before study entry at up to two months? Was any attempt made to assess the effects of timing and duration of different exposures on outcomes? The fact that there is so little difference in head circumference and also infection rates between the groups makes me wonder about the power of programming effects of early exposures.
Is it possible for you to assess in far more detail at the feeding histories of all children? Given the central importance of neonatal gut colonisation, it would be useful to define as the reference group a truly exclusively-breastfed-from-birth to 6 months cohort, and compare them with a formula-fed-from-birth group (i.e., the babies who will be using this formula if the current marketing works.) It may be that the results differ far more strongly when truly EBF infants are compared with truly EFF infants, especially if exclusive breastfeeding continues as WHO mandates, till around six months rather than four months.
A separate issue would be the impact of timing, quantity and variety of foods introduced from 4-6 months. (As noted above, the EBF reference group might have been solely BF only from around 2 months to around 4 months. It would be helpful if WHO recommendations were tested.)[An interesting aside for those conscious of language: when researchers posit benefits for breastmilk, the language is almost always tentative. When benefits are posited for infant formula variants, there is little hesitation in claiming benefit in unequivocal language, such as the new formulation “eradicated the gap in cognitive performance” [in healthy term mixed-fed infants, compared with the small not exclusively BFR group]. A less positive developmental outcome briefly noted in the study –“because of enhanced performance in the receptive verbal subscale, the BFR group performed better in the verbal domain than did both formula-fed groups” – did not make it into either conclusions or abstract, and so will not be widely noticed or reported. If that better result “in the verbal domain” had been recorded for the experimental formula group, would it have been omitted from the conclusion? Or would it have been recorded as having “eradicated the difference in the verbal domain seen in older higher-protein formula lacking MFGM additives.”]
I am writing this to you because massive industry campaigning is dependent on the basis of your pioneering studies – along with what seems to me quite illogical extrapolations from breastmilk science. For it is surely illogical to assume (and indeed for marketers to tell mothers) that because HMFGM components- in the living bioactive tissue that is breastmilk- have positive effects in human infants, industrially processed bovine or synthesized MFGM will have equally positive effects on human (not bovine) offspring – in the many complex mixes and batches that are infant formula, and in the differently constructed microbiome and organs and epigenome of the formula-fed infant. (What bioactivity of bovine milk remains after industrial processing and then home manufacture with water at 70C?)
And secondly, if, when part of a dehydrated and reconstituted powder mix, any new ingredient IS active in vivo, is it always acting in ways that promote growth and protection appropriate for human young? It may well be beneficial, of course, as some industry-funded research is indicating, including yours. If that is so, the earlier industry decision to stop using milk fat should be seen as a mistake, as some scientists have admitted. All international standards for formula should then insist on the inclusion of BMFGM components as soon as possible.
But anything powerful enough to be helpful is also powerful enough to be damaging… and that will not become obvious until more than 80 healthy term, partly breastfed, Swedish infants have been exposed – and then only if there is ongoing independent post-market surveillance. [Which the USFDA requests of manufacturers, but is not given, as freedom of information requests have revealed.] Longterm followup is crucial.
And I do have another concern. In most consumer testing of products, independent testing agencies with no financial links to a manufacturer either purchase standard product off the shelf, or carry out independent assessment any product supplied. It would be easy to skew results by slight differences or greater quality control of the new product to be compared with an older one. Such deception ought to be unlikely, and I hesitate to raise it. But the stakes in a $70billion industry are high. I know of one early comparative trial of an extensively hydrolysed product, in which severe reactions prompted the researchers to conduct a belated independent analysis that revealed large amounts of intact protein. The trial was stopped. Another batch supplied by the company caused fewer reactions, and those results were published, ensuring widespread acceptance of the product. The negative performance of the first batch was never publicly reported. But it was the off-the-shelf product that babies were consuming. As you would know, label quantities cannot be exact representations of what the can contains. Were both supplied formulas independently analysed and verified as differing only in the ways mentioned in the article? I hope researchers would not accept at face value compositional data supplied by vested interests.
I apologise in advance if I have been less than tactful in this email – I often am! However, I feel you should know the way your research is being used, as I feel it creates some moral responsibility. Do you consider that the addition of bovine milk fat constituents makes infant formula closer to breastmilk, as the marketing by Abbott, Mead Johnson and others is claiming? If not, will you dissociate yourself publicly from that claim, and reinforce the message that exclusive breastfeeding from birth to around 6 months is in fact important to health lifelong, and that even the best formula cannot match its outcomes? (Pluripotent stem cells are unlikely to be included in formula any time soon!)
I’m attaching an image of the marketing in Vietnam. The pearl in gold casing has strong emotive power in Asia. But the claims of better intellectual, emotional, growth, and behavioural outcomes are incredibly powerful in cultures where women believe their own breastmilk cannot be as good as something produced by clever scientists like yourself and your colleagues. This is how industry is using your work.
I am asking you to think more widely about the consequences of the research you are engaged in, and to be more conscious of how it is reported, and the damage this can do to breastfeeding promotion. I agree entirely that infant formula needs to be as good as it can be, and support the research to improve it, since we cannot do without it in societies as currently constructed. But societies can change, and would, if scientific papers, by omission, did not inadvertently support the illusion of equivalence of current formula and breastmilk. Human omnivores survive and grow on many varied and inappropriate diets. I would argue that the vertically communicated inflammatory disease epidemics we face are proof that both the absence of breastfeeding, and the presence of ersatz substitutes for breastmilk, is responsible for compounding intergenerational damage. To me, a milk hypothesis makes more sense than the hygiene hypothesis, which it subsumes. I’d be happy to supply you with a pdf of the book I’ve written which argues that case.
I will look forward to updates about your work at ISRHML in Sweden next year.
All best wishes,
Author: Milk Matters: Infant Feeding and Immune Disorder
I received a perfectly reasonable and polite reply. The most relevant part of the reply:
“we have tried to be as clear as possible that comparisons between the formula-fed and breast-fed infants must be made with caution. We have also stated that it´s too early to give any recommendations on MFGM supplementation to infant formulas, more studies are needed.
The reports from our study are written to the scientific community. Unfortunately, as scientists, we have no influence on how the information from the study is used by companies. We have always been very clear that breastfeeding is the optimal nutrition for infants and will continue to stress that and we hope that the present study in the future can contribute to a better nutrition for those infants that can not be breast-fed.”
This issue is discussed by Mills, Hall and Berry: What are interesterified fats and should we be worried about them in our diet? Nutr Bull. 2017 Jun; 42(2): 153–158.Published online 2017 May 8. doi:10.1111/nbu.12264(PMID:28729812 PMCID:PMC5497165) http://europepmc.org/articles/PMC5497165
They say, “the impact on cardiovascular health is unknown. The use of IE fats instead of animal fats is assumed to be beneficial as it should help meet the population target for saturated fatty acids intake of <11% daily energy intake (Department of Health 1991), but there has been no systematic attempt to establish whether this is the case.”