Nearly all studies on the benefits of breastfeeding are meaningless

meaningless, word written in letterpress type blocks

Nearly all studies on the purported benefits of breastfeeding are meaningless.

Why?

Because the choice of the comparison groups is beset by a crippling flaw.

Nearly all breastfeeding studies compare two groups of infants, those that have been exclusively breastfed and those that have not. The researchers inevitably fail to include an important group: those infants whose did not get adequate nutrition when breastfeeding exclusively.

To understand why that is a problem so large as to render breastfeeding studies meaningless, it helps to consider another example.

Suppose we are trying to determine the benefits of ICUs (intensive care units). What happens if we compare hospital survival rates of patients who spend time in the ICU and those who don’t? I bet it won’t surprise you to find that the survival rate of patients who spend time in the ICU is lower than that of patients who are never treated in the ICU.

Many babies DON’T thrive on exclusive breastfeeding and mothers switch to formula in order to save their lives, promote their growth and preserve their brain function.

Does this mean that ICUs are harmful and we should abolish them? Of course not. We recognize that those patients who are treated in the ICU differ in fundamental ways from those who are not. Specifically, those who are treated in ICUs are far more ill than those who never spend time in the ICU, so it’s hardly surprising that they are less likely to survive. The difference in level of illness between the two groups is known as a confounding variable. Any study that does not address confounding variables produces results that are invalid.

But there’s an additional factor beyond confounding variables that renders the comparison meaningless; a key group is missing. If we want to know if ICUs are beneficial, we must compare very sick patients cared for in the ICU to very sick patients who are not cared for in the ICU. That produces a radically different conclusion; very sick patients treated in the ICU have better survival rates than very sick patients who are not treated in the ICU. In order to reach a valid conclusion, we must make a valid comparison.

How does that apply to breastfeeding studies?

Most breastfeeding studies compare babies who are exclusively breastfed with babies who are not exclusively breastfed. Even when you correct for confounding variables (mother’s socio-economic status and education level), you are still left with the wrong comparison. If we want to know the benefits of breastfeeding we need to add a third group: babies who are exclusively breastfed but are not thriving.

In order for an infant to thrive, the food it receives must meet three criteria:

1. It should contains all the nutrients and other factors that an infant needs.

2. It must be available in sufficient quantity to promote vigorous growth of the infant.

3. The infant must be able to access it easily.

Breastmilk contains the nutrients and other factors that an infant needs. But many women (up to 5%) do not produce enough breastmilk to fully support infant growth and development, and some babies are not capable of efficiently extracting milk from the breast. Some babies just can’t do it for anatomical reasons, because of weak muscle tone, or because they simply never get the hang of it.

If we want to know the difference between babies who are exclusively breastfed and those who are formula fed, we MUST include the babies who are being exclusively breastfed and failing to thrive.

Technically speaking, babies who are exclusively breastfeeding at 3 months or 6 months or any other arbitrary length of time represent only a subset of breastfed babies. That subset is babies who are thriving on exclusive breastfeeding. Comparing babies who are thriving on exclusive breastfeeding with those who are formula fed won’t tell us anything about the benefits of breastfeeding as compared to formula; we must include babies who are failing on exclusive breastfeeding. Of course, that would be a very difficult study to do because we live in a society with easy access to infant formula. We’d have to do that study in a society that lacks access to infant formula.

We’d need to create 3 groups:

  1. infants who are exclusively breastfed and are thriving
  2. infants who are exclusively breastfed, are failing and lack access to formula
  3. infants who are exclusively breastfed, are failing and are given formula supplements

To my knowledge, that study hasn’t been done for ethical reasons. But if we could conduct such a study, I suspect that we would find that infant formula has dramatic benefits as compared to exclusive breastfeeding. I’d be willing to bet that the death rate for exclusively breastfed infants would exceed by a wide margin the death rate of formula fed babies.

This point is critically important and the reason that it is routinely ignored reflects the lactivist bias that is pervasive in breastfeeding science. The bias is the lactivist lies that all women make enough breastmilk and all babies can easily access that breastmilk. The assumption, completely unjustified and not supported by scientific evidence, is that the only difference between formula fed infants and breastfed infants is that the mothers of formula fed infants couldn’t be bothered to breastfeed. The truth, which would smack any breastfeeding researcher in the face if he or she bothered to look, is that many babies DON’T thrive on exclusive breastfeeding and mothers switch to formula in order to save their lives, promote their growth and preserve their brain function.

Sure, successful exclusive breastfeeding might be beneficial compared to formula feeding, but that’s something entirely different than claiming that breastfeeding is better than formula feeding. The truth may be the opposite: that when compared to exclusive breastfeeding (successful AND unsuccessful) formula feeding is not merely beneficial, it is life saving. And, therefore, many mothers who are formula feeding are doing so because of the extraordinary benefits of infant formula.

Those formula feeding mothers deserve our support and our praise, not the sneering derision of lactivists.

  • Eater of Worlds

    Breastmilk is so fantastic we have glycophosphates in it! Breastmilk truly does have everything.

    http://en.europeonline-magazine.eu/german-politicians-sound-alarm-after-weedkiller-found-in-breast-milk_399094.html

    • Daleth

      Seriously. I use an organic (actually “biodynamic”) sourced from cows that live on biodynamic farms in the Alps. I was like, those cows don’t spend their lives in polluted cities and occasionally snarf fast food for lunch–their milk is almost certainly cleaner than mine!

  • dimux

    Your first choice skepticalob Find Here

  • guest

    They always ignore the baby factor as they think every baby will naturally root for milk. My 36 week baby never rooted or tried to latch and never sucked when I forced a latch (which is a weak latch as forced on flat nipple and non responsive baby). She just stares at me or falls asleep. My milk came in after 6 days after a mastitis. Se got how to suck incontinueously though after a month!. but never made enough milk on pumping and all teas and meds. demand and supply didn’t work.

  • Jodi Hilla McCormack

    I’ve considered for years why nursing failed with my oldest, why she absolutely REFUSED to nurse. I think it comes down to two things:
    1. She suffered from reflux for the first month of her life. I’m also VERY short-waisted and she always ended up in a mostly horizontal position that was painful.
    2. She has an ADHD brain and we can look back now and see the signs SO SO early. She was fascinated to look at the world and I think nursing was far too dull (I used to take her out to the store on the weekends when she got fussy and she absolutely could not focus on a bottle in public). She also ignored her own signals so much that by the time she made any signs of being hungry, she was absolutely starving and screaming as loud as possible. This often occurred in the car as we left somewhere, taking away all the external stimuli and allowing her to examine the signals her body was sending. And this was all before 6 months!

    Do not let any study tell you bottle feeding causes ADHD – as the parent of two children with ADHD, I firmly believe ADHD causes formula feeding.

    • yentavegan

      You may believe that ADHD causes formula feeding but that does not make it true. ADHD is a whole can of worms unto itself. There is no agreed upon definition for ADHD and medical professionals as well as quasi medical wannabes are all to eager to treat it.

      • fiftyfifty1

        “There is no agreed upon definition for ADHD ”

        Sure there is. You can find the diagnostic criteria in the DSM5.

        • yentavegan

          I highly doubt the school social worker who diagnosed my high spirited child followed the criteria.

          • fiftyfifty1

            A school social worker, by law, cannot make a diagnosis of ADHD (unless by chance they happen to also be a licensed clinical psychologist, in which case they do have to use the criteria).

  • MichelleJo

    Ooooh, I would have loved this post when I was being told that breastfeeding works by supply and demand, so the baby must be getting enough nutrition. Yes. but. that. is. when. the. system. works. They don’t seem to get it that the rules of breastfeeding don’t apply if the mother doesn’t have sufficient supply!

    • yentavegan

      Were you told to count wet and dirty diapers? Did your infant not regain her/his birth weight by 14 days?

  • Cobalt

    Here’s an idea for a feeding study that might help better answer the “benefits of breast milk” question and would go a long way towards defining if donor milk is worthwhile for typical, healthy, term babies:

    Use only those who do not intend to breastfeed at all, this will eliminate an awful lot confounders. Give half of them formula, and half of them safe, screened, donor milk. Track outcomes.

    Sourcing that much donor milk would be difficult, but I think this idea has a lot of merit.

    • Who?

      Is that sound members of the lactivist community having simultaneous strokes?

    • anh

      yeah, but the real zealots insist the magic goes away unless the breast milk is drunk from the tap.

      • Cobalt

        Absolutely. But it’s still valuable to show if donor milk is worthwhile (the practice of informal donations/sales NEEDS this information), and is our best shot at equalizing variables. Studies of exclusive pumping and exclusively from the breast could be added to fill out the picture.

    • guest

      The other difficulty would be getting the moms to agree to donor milk – so many people think it’s disgusting.

      • Cobalt

        True. And it would have to be a large enough group for the outcomes to be statistically significant. I think the numbers could be found though.

      • Amy M

        Well, if the donor milk was screened and pastuerized, I think you’d have more people willing to use it.

        • Cobalt

          Especially if it’s at the same cost as formula. You could try making all the food free, but opportunists would sell the donor milk on Craigslist, use the money to buy formula, and pocket the difference.

          That’s actually a major obstacle, and part of why WIC has to incentivise their breastfeeding packages. There are moms who will tell WIC they need formula even though they are breastfeeding so they can sell the formula for cash.

  • Busbus

    What a great point.

  • Sue

    Great article.

    This reflects what Bofa always says: Breast is best, ALL OTHER THINGS BEING EQUAL. But, when other things AREN’T equal (like supply), it may not be the overall best strategy.

    You are right – nobody would intentionally do a study where babies failing to thrive in exclusive BF are left to starve. But, in real life, people do that, falsely believing they are doing the right thing.

  • indigosky

    Thank you! This is EXACTLY what I have been trying to say for years, but as I lack eloquence and a science background, it comes out all garbled when I try to make this point.

  • Cobalt

    OT, but I enjoyed this and thought others here might as well.

    http://www.brainchildmag.com/2015/06/why-i-hate-dr-sears/#

    • Sue

      Great article.

      Ever noticed how many people “see the light” when they have a child with a disability or behavioural issues?

      Rational parents dont go looking for an external cause of blame or radically change their lifestyles – they continue to raise all their children rationally.

  • just me

    I disagree. Of course we don’t want people under feeding babies bc they’re afraid of formula. But I see nothing wrong with a study comparing ebf to non ebf that is comparing healthy ebf babies to healthy non ebf babies. That data is informative.

    Your premise is like saying ignore the studies that show health benefits from eating vegan, bc there’s this small subset of vegan people who have eating disorders and consume only 500 calories a day.

    And wouldn’t a study that does not break out a separate group of underfed babies skew the results such that the benefits of ebf would be reduced; the underfed babies would presumably be included in the ebf group.

    And I’m not anti ff by any means. I supplemented with my kids. But I think it’s disingenuous to pooh pooh all studies showing a benefit bc you don’t want women to feel shamed.

    There’s a difference between a study showing some type of benefit to bc and an article saying ff is detrimental. Not the same thing. I think some posters are conflating the two.

    • MegaMechaMeg

      I personally am actually very interested in the long term effects of inadequate nutrition on a baby. For a mother who is trying to make breastfeeding work, that information would be very helpful. If the typical infant can tolerate short periods of inadequate milk then the mother would be able to simply allow clusterfeeding or increased night feeding to signal her body to produce more and avoid the hassle of supplementation and pumping. If the typical infant is shown to not be resilliant it would be very good information to have to prevent harm and might help make the decision to supplement feel more cut and dry. It would be also be helpful from a pediatric standpoint to have an idea of where the point of diminishing returns comes when breastfeeding is not working. I am a data nerd and I like the idea of being able to make a fully informed decision backed by research.

      • Mattie

        “If the typical infant can tolerate short periods of inadequate milk then the mother would be able to simply allow clusterfeeding or increased night feeding to signal her body to produce more and avoid the hassle of supplementation and pumping.”

        I think, on the whole, babies can tolerate short periods of inadequate milk supply…physiologically, it is unlikely to hurt them. However, ‘short periods’ is sort of an ambiguous term, do we mean 1 feed, 1 day, 1 week…short to an adult can be an awful long time for a newborn, babies can do well losing a considerable amount of weight, there will likely be no lasting damage, but as a parent the difficulties you feel watching your baby lose weight and having issues breastfeeding might be harder than the supplementation and pumping.

        Saying that, cluster feeding is normal breastfeeding behaviour (although tough for the person feeding the baby) and night feeds, or pumping at night is essential at the start to build up supply (I believe because of higher levels of prolactin at night, but can’t entirely remember) so if supplementing try to breastfeed first at night, then ‘top up’ if needed.

        I agree that a study could be useful, but unethical because starving babies is generally seen to be wrong in most circles, even if you could find a group of women willing to ‘avoid formula at all costs even if the baby is starving’

        • MegaMechaMeg

          It would not personally be my choice to allow a baby in my care to be hungry for lack of breastmilk, but I do have friends who wanted to avoid formula and who went to great lengths to make that happen. Unfortunately if your milk supply is inadequate in the best case it will take a couple of days for your body to adjust and start producing enough. Without supplementation the child will be getting less than it wants, and I wouldn’t necessarily call that starving the baby, just a short term controlled hunger while the mother’s supply adjusts.
          Clusterfeeding and night feeds are the traditional way to recover supply, but formula topoffs and pumping sessions can also be used sucessfully. I don’t really know if it is known whether or not there is a risk in the formula free method, but I do know a solid handful of women for whom that is general practice and it seems like we could dig up enough in the EBF group for a population study, at least compared to formula top-offs.

      • just me

        That’s cool. I’m not saying such data would not be important. I’m saying the lack of such a study doesn’t invalidate these other studies.

        • fiftyfifty1

          ” I’m saying the lack of such a study doesn’t invalidate these other studies.”

          Actually it does. The fact that not a single breastfeeding study besides the PROBIT study uses randomization or an intention to treat analysis does indeed make those studies invalid. As I mention above, it’s just like the hormone replacement studies. Women who took, and stayed on, HRT were an elite group. They were largely rich white, health conscious women who were “healthy adherents”. The non-HRT groups included women who could not remember to take a daily pill, could not afford the medicine, or who had had bad outcomes when they tried the hormone pills and quit. No wonder the HRT outcomes looked “better”. But when a well designed study, the WHI study (randomized, intention to treat analysis) came along it did indeed invalidate, literally overnight, all the hundreds of poorly designed studies that had come before.

          • Megan

            The WHI study on HRT is a great example.

    • Roadstergal

      “Your premise is like saying ignore the studies that show health benefits from eating vegan, bc there’s this small subset of vegan people who have eating disorders and consume only 500 calories a day.”

      I would say it’s more like – mandating people eat vegan because vegans have a better cholesterol profile than the population at large, ignoring the fact that increasing plant intake and reducing the intake of fats, including lowering animal product intake, can have the same health benefit while being more achievable for the majority of the population, and is more effective than shaming people for letting even the smallest bit of animal product cross their lips.

      It’s looking at the best outcome of the most extreme behavior and saying it’s the only option.

      • just me

        I disagree on the vegan thing but I’m not going to argue about it bc I just was using it as an example.

        Again, you’re conflating lactivists’ “breast milk only” with peer reviewed studies saying “hey there’s a benefit here”. Those studies aren’t saying this is the ONLY way.

        • Cobalt

          The peer reviewed studies exclude all the negative outcomes of breastfeeding that were avoided by switching to formula when breastfeeding was inappropriate.

          Imagine a drug safety study where everyone who had a severe reaction was excluded from the results because they had to stop taking the drug.

          • Amazed

            I could have been excluded from such a study. My allergic reaction to a widely spread anti-allergy drug was so severe that in less than a week, we realized what was going on (my arm was three times its normal size and purple). Should I be excluded from a study about the safety of this drug because I stopped almost immediately or because I was “unhealthy” (v. the drug MADE me unhealthy)?

          • guest

            Drug studies are not the same. If you do a drug trial, you aren’t accepted if you have a significant health problem unrelated to the condition the drug is supposed to treat. For example, if there’s a trial for a new asthma medication and you are HIV positive, you won’t be in the trial. They want to know what side effects are CAUSED by the drug, so you must not have any prior to start it (other than whatever the drug is supposed to treat.

            Formula is not a drug – it is a food. So a controlled study would want to study its effects on a baby capable of breastfeeding – not one who from the start can’t succeed at breastfeeding (whether on the part of the baby or the mom). If a baby capable of thriving on breastmilk could be shown not to thrive on formula, then we’d have the data we want.

          • guest

            To clarify a bit, we’re in a place right now where we think breastfeeding is best, but we also know some children need formula. What we aren’t completely sure of is how much formula is to blame for the health problems of formula fed babies. If formula is the second choice food after breastfeeding fails, then you can’t just study all EBF babies against all EFF babies, because EFF babies are largely self-selected to not do well on BM. But as I said above, if you could isolate babies that *would* do well breastfeeding and give them formula instead – then we’d have a better picture of whether formula *causes* problems (clean water and money to buy formula aside). But as long as all the breast fed babies studied are the ones who succeed at it and a large number of the formula fed babies studied are the ones who failed at breastfeeding, we don’t really know much of anything.

          • Amazed

            And if I do not have any health problem unrelated to the condition the drug is supposed to treat? I didn’t have any. The drug simply worsened the condition it was supposed to treat. It was a sting of some fly near my shoulder; next thing I know, off the ring went.

            Just like the way breastfeeding studies are done today. They take into account what mothers did – and most of them switched to formula when severe problems appeared. Just like I did with the drug. I needed a lot of time to get the arm down to normal and fully functional.

            I get what you say about the way a valuable study should be conducted but I can’s see how it can be done ethically. At the same time, I don’t think cases like mine should be excluded from results and effects of drugs simply because it turned out it was most emphatically not beneficial to me and was even harmful.

          • guest

            I think we could compare mothers who chose to EFF from day 1 for convenience/preference (vs. inability) to mothers who chose to EBF and succeeded. That would be a start, and while it might be hard to find the EFF mothers my choice in sufficient numbers, I think it would be ethical.

            You weren’t in a study. If you were in such a study and you had your reaction, that would be the data we wanted. And indeed, some babies are allergic to formula – but we already know that. Some are even allergic to breast milk, though fewer. But that isn’t the data that is wanted here. We want to know about infections, obesity, IQ, etc. If you were drafted into a study and you *already had* a low IQ and were then formula fed, that proves nothing. You already had a low IQ, so formula didn’t cause it. When you compare successful breastfed babies to a mixed bag of unsuccessfully breastfed babies, partially successfully breastfed babies, and finally babies who could have been breastfed successfully but just weren’t, it’s an unfair comparison because some of the negative outcomes might have caused the problems with breastfeeding in the first place. Or, they might have been caused by formula. The problem is that you can’t tell with a study that does that.

            So it’s not the same as your allergy example at all.

          • Sarah

            I’d be glad to participate in such a study, though my daughter was in special care for the first 48 hours so despite being the picture of health now, she might not be suitable. The big problem is, in the UK at least, mothers who choose to formula feed are almost invariably from the poorest social groups. Anecdotally, I’m the only well educated, professional woman I know who’s EFF a child from birth because I had no interest in breastfeeding. And I don’t share Dr Amy’s confidence that social class can really be controlled for, either. It’s too complex and pervasive, and the effects of social inequality on health are only just beginning to be understood.

          • fiftyfifty1

            “I don’t share Dr Amy’s confidence that social class can really be controlled for, either.”

            Oh I doubt that Dr. Tuteur has any confidence that social class can really be controlled for. I believe her point is that at least you can TRY to control for it, unlike the problem of inadequate milk supply.

            The breastfeeding literature is just like the postmenopausal hormone replacement therapy literature before the Women’s Health Initiative randomized study. There were literally hundreds of observational studies all that purported to control for social class, and all that showed better outcomes for HRT. But then along came a well designed randomized study, and proved the exact opposite.

          • Sarah

            Well, she said in the post ‘even when you correct for confounding variables (mother’s socio-economic status and education level)’ without any caveats as to the possibiity of doing this. So I take her at her word until and unless she says otherwise.

          • guest

            It is complicated, but I’m not really satisfied with saying “poor babies are dumb and fat” and leaving it at that.

          • Sarah

            No. But equally, I’m not satisfied with accepting that social class is something that can be controlled for. That’s very, very far from certain.

          • fiftyfifty1

            “while it might be hard to find the EFF mothers {b}y choice in sufficient numbers,”

            Ha! I can find you find you a limitless supply of women who choose to EFF right off the bat. Does it matter to you that they are all inner city teens who haven’t finished high school?

            The truth is that the demographics of women who want and plan to exclusively formula feed are very different in many ways from women who want and plan to exclusively breastfeed.

          • guest

            That’s why you control for SES, and why I said it would be hard to find the EFF mothers you’d need.

          • moto_librarian

            Why are you unable to understand the socioeconomic issues at play here? There is bias off the bat with your proposed study because EBF babies tend to be born in privilege. Their moms can afford to stay home and breastfeed. Those who choose formula off the bat often do so because they have no paid family leave and are in tenuous financial positions. We know that socioeconomic factors play a huge role in IQ, obesity, etc. You would just be generating more meaningless data.

          • guest

            Um…I’m not? A good study would need to control for those factors. Believe it or not, there ARE white, middle class women who also don’t want to breast feed who could be included in the study, and poor black mothers who breastfeed who could be included on the other side – they’re just harder to find. And nowhere have I claimed that this is a perfect study, only that it would be better than what we’ve been getting so far (other than the sibling study, which was another good approach). Nor do I think Dr. Amy’s point is off in any way. I’m not sure what you’re taking issue with other than me not specifically mentioning controlling for SES as I wrote my comments late last night. I forgot to mention it, okay? Chill.

          • fiftyfifty1

            “Imagine a drug safety study where everyone who had a severe reaction was excluded from the results because they had to stop taking the drug.”

            Exactly. This is why drug studies are done with an intention to treat analysis. If a person is placed into the group receiving the drug, their outcome is included in that group, even if they have to stop taking the drug due to a severe reaction. In contrast, the only breastfeeding study we have with an intention to treat analysis is the Belarus PROBIT study.

        • Roadstergal

          They’re poorly done studies, is the issue. They’re studies showing benefits that are more related to the reasons why EBF was possible (as fiftyfifty mentions above, reasons that are independently associated with good outcomes) than the EBF itself.

          The better-controlled the study (specifically, the PROBIT and sibling studies), the fewer and more trivial the benefits seen for EBF.

        • fiftyfifty1

          “with peer reviewed studies saying “hey there’s a benefit here”. ”

          No, those studies don’t find “there’s a benefit here”. They find that exclusive breastfeeding is CORRELATED with good outcomes. Correlation is not the same as causation. Studies with robust design that eliminate or nearly eliminate confounders, do not show any benefits, or at best trivial health benefits to breastfeeding (in countries with clean water and reliable access to commercial formula).

    • fiftyfifty1

      No, you are not understanding. Dr. Tuteur is not saying the studies are weak out of a misguided desire to save mother’s feelings. She’s saying it because the studies ARE weak to the point of being totally useless. The only 2 studies we have that have any sort of robust design at all are the discordant sib studies and the randomized Belarus PROBIT study. Neither study shows anything other than trivial benefits of breastfeeding.

      All other observational studies of breastfeeding are severely marred by confounding. Women who are exclusively breastfeeding at 3 or 6 months in this country are women who 1. wanted to breastfeed, 2. make plenty of milk 3. have a baby that sucks normally.

      These are 3 positives right off the bat. It is like the studies that look at vegans. Sure they seem healthy (except for all the anorexics, but everybody always excludes them saying they shouldn’t count) but that’s because they are a bunch of health-conscious almost exclusively well off white people who can afford the time and money to cook produce, legumes etc.. But how does a vegan diet stack up if you were to actually *randomize* people to do it? Would it really be healthier than say, being randomized to a Mediterranean diet or a DASH type diet? Hard to know. Although at least you could ethically do the study, unlike with breastfeeding.

      • Roadstergal

        You said it better. I would just add 4: have sufficient SES to be a SAHM/work from home.

      • Cobalt

        Skittles are vegan (according to PETA). If I had to eat vegan, I would eat a LOT of Skittles. Not only would I be unlikely to lose weight, I would probably quadruple my diabetes risk.

        • just me

          You proved my point. Let’s ignore studies showing the benefits of eating vegan, because there’s vegan candy!

          • Cobalt

            If I was randomized to the vegan group in a study, my results would be dramatically different than what is currently seen in those who CHOOSE to be vegan. The benefits of veganism for me would, in fact, be mostly risks. Risks currently unaccounted for in studies of vegans.

            Just like breastfeeding. If you only study those who choose it and have the resources to maintain it, you miss all the negative outcomes avoided by those for whom breastfeeding is inappropriate. If you forced it on a random sample, those negative outcomes would be revealed.

          • Gatita

            Just like breastfeeding. If you only study those who choose it and have the resources to maintain it, you miss all the negative outcomes avoided by those for whom breastfeeding is inappropriate. If you forced it on a random sample, those negative outcomes would be revealed.

            This is it. What Amy was saying. I was having trouble with it too but this makes it crystal clear. Thank you!

          • guest

            I think you could do a study of moms who choose to ebf successfully and moms who choose to eff simply because they like it better (vs. can’t get the latch right, low milk production, FTT, etc.). That might tell us something. As far as I know, it’s never been done. The sibling study would be the closest, but I don’t think they cared why a mother did formula with one vs. the other.

          • just me

            This is kind of what I was getting at. I never said extrapolate to people with issues preventing them from ebf.

          • fiftyfifty1

            But comparing women who wanted to breastfeed (and did so without problem) with women who chose formula from the start would be a terrible comparison. Who chooses exclusive breastfeeding in our country? Mainly well off, healthy, college educated white women who either plan to stay home or have flexible jobs that allow pumping. Who chooses to formula feed off the bat in our country? This group is heavily weighted with young, poor African Americans, those with hourly wage jobs without maternity leave and women with severe health problems. Sound like a fair comparison to you?

    • Amy Tuteur, MD

      Actually comparing health EBF babies to healthy formula fed babies is NOT informative; it’s deceptive. That’s my point.

      It’s like comparing drunk drivers who didn’t have an accident to non-drunk drivers and then claiming that drunk driving is safe since the drunk drivers in the study had no accidents. If you exclude the bad outcomes, you can’t draw conclusions about the benefits.

      The choice for many women is not between easy exclusive breastfeeding vs. formula feeding. The choice is between poor exclusive breastfeeding vs. formula feeding. Studies that look only at successful breastfeeding can’t be extrapolated to women who are having problems.

    • Courtney84

      My son was supplemented on day 3 and 4 of life due to excessive weightloss and milk that had not yet come in. He received only breast milk from day 5 through 9 weeks. My fertility returned early and tanked my milk supply. We started combo feeding around 9 weeks. By 5.5 months he was fully weaned from the breast due to issues with supply and feeding behaviors that indicated he wasn’t able to eat until satisfied at the breast. At 6.5 months he was diagnosed with an uncommon allergic condition – FPIES. So my kid goes in the non-EBF group. Late milk, insuffiecient milk, and likely allergic reaction to proteins sometimes present in the milk caused feeding difficulties including a baby who’d been born in average percentiles to be down to 0.4 percentile for height at 4 months old. So if we were to put my kid in the non- EBF group it’d look like he was shrinking and developed an allergic condtion because Quel horreur! we fed him formula. But infact feeding him formula was likely much healthier for him because a) there was sufficient amounts and b) he was not allergic to it.

      This is the meaningless ness that is being pointed out. Nearly all parents who have a baby who is not doing well on exclusive breast milk (supply, allergies, physical difficulty with removing milk at the breast, feeding issues of unknown origin) is going to end up combo feeding or moving to exclusive formula such that infants with problems are skewed into the non ebf group.

      • Courtney84

        This is the same problem that most of the followers here rail against in studies of homebirth that look at actual place of birth instead of intended place of birth. A woman sees a midwife. She goes into spontaneous labor around 41 weeks. The midwife has not been monitoring the mother or fetus to the same standard an OB would. After many hours of laboring at home the babies heart rate is either dangerously low or can’t be found. They call the ambulance and the laboring woman arrives at the hospital. At the hospital an emergency c-section is performed the baby is born brain damage or daed, and the mother goes on to develop a post op infection which is more common in emergent c-sections. When this outcome goes into a hospital birth category it negatively skews hospital results. Laboring in the hospital would likely have caught the baby’s distress sooner – the baby would have been far more likely to be delivered alive and neurologically intact. It is also possible that with more warning she’d have been less likely to contract an infection.

        Comparing EBF to non EBF infants is highly similar, if not the same as, lumping the home birth transfers in with the planned hospital births. Many woman who PLAN to EBF end up in situations where they must switch to partial or full formula feeding. Putting these babies into the non-EBF group skews that data for infants who are combo fed or formula fed intentionally. I suspect, if we eliminated all of the children who’s parents intended to EBF but were “forced” into combo/formula feeding we’d find that the reported differences in breast vs. formula are much smaller if not unquantifiable.

        • SuperGDZ

          Ditto when comparing the outcomes of c-sections and vaginal births – where are the women who attempted a vaginal birth but ended up with a c-section counted? So many women go to extraordinary lengths to avoid a c-section because they believe a vaginal birth carries a lower risk of complications – but this is only true of a successful vaginal birth. It’s circular.

          • Roadstergal

            Yeah, there need to be more studies looking at outcomes of only planned, pre-labor C-sections. Because the one I’ve seen make it look pretty good. Compare that to the outcome of all attempts at vaginal birth (including emergency C-sections after a failed VB attempt), and that’s a properly useful comparison.

  • LizzieSt

    I am imagining what people will be saying about all this in the year 2055.

    “Grandma, I just read that in your day, there were people who thought it was better to let a baby starve than to feed her any formula. That can’t be true, can it?”

    “It was true. Many people believed that every mother was capable of producing enough breastmilk to keep a baby well-nourished. They became very angry if a mother used formula, even to supplement a low milk supply.”

    “But that’s crazy!”

    “Yes, it was.”

  • Liz Leyden

    OT- Slate has published another piece on infant feeding. It’s a pretty solid piece on the science behind infant formula, but the author insists on presenting formula as inferior to breastfeeding.

    http://www.slate.com/articles/double_x/doublex/2015/06/formula_fed_vs_breast_fed_babies_can_we_create_a_better_formula.html

    • Cobalt

      Well, it’s a start.

  • Angharad

    I feel like a lot of the effects of formula feeding that lactivists like to cite could actually be reasons for switching to formula. Allergies – give the baby something they’re certain not to be allergic to. Asthma – if the baby is having a hard time breathing they may do better with a bottle. IQ – if a baby has a developmental disability that affects both IQ and oral motor skills they may do better with a bottle. Etc, etc. I feel like this may be a classic case of correlation being mistaken for causation.

  • Mel

    Plus, this study has been done in cattle many times.

    In an extremely simplified system where dams are selected only for ability to breed and raise a calf, primary lactation failure is as high as 10%.

    Calves that aren’t thriving on milk will be stunted unless fed supplemental formula known as Milk replacer.

    Final carcass weight is unchanged; the cost of formula-raised is higher, but a dead calf gets no profit.

    • anotheramy

      You just triggered a fond memory for me: ” bottle calves” or ” pail calves”!! 🙂 I loved to feed the “pail calves” on my uncle’s farm when I was kid. Too bad there were no bovine lactavists to inform me that pail calves were actually inferior calves or to shame the cow into feeding the calf the way nature intended. /sarcasm

  • Amy M

    I think that the discordant sibling study was pretty good, in terms of discouraging confounders. Would that be considered a totally different kind of study, or could discordant siblings be worked into the groups Dr. Amy mentioned?

  • DaisyGrrl

    What happens in studies now when women in the breastfeeding group start formula feeding? Are those results discarded or are they transferred to the formula feeding group? I’m only familiar enough with the literature to know that the benefits of breastfeeding are so minor that I don’t give a crap how a woman chooses to feed her baby as long as baby is fed.

    If babies who are switched to formula at some point are either dropped from the results or are transferred to the formula group, then that’s a glaring problem that is further poisoning the results of these studies.

    • MegaMechaMeg

      Like Dr. Amy said, I think the real statistical problem is combination feeding. If it were a matter of breastfeeding for x days and then stopping it would be easy to look for relationships, but there are so many permutations of combo feeding that it gets hard to know where to draw the line between breast and bottle. Combo feeding spans anywhere between one formula bottle a day to top offs after nursing to breastmilk when the mother is present and formula while she is at work. When you toss in things like whether on not the baby was recieving adequate nutrition at all times the whole mess gets so complex that you might as well be throwing divining rods. At the end of the day I think all we really can be sure of is that feeding your baby is better than not feeding them whatever path you take. I would take it as a personal favor if people stuck to the two approved classes of food intended for babies and didn’t screw around with camel milk and the internet, but you can’t live peoples lives for them.

    • Amy Tuteur, MD

      There are no randomized studies. Babies are put in the breastfeeding or formula feeding groups based on what their mothers have done, not what they promised to do.

      • fiftyfifty1

        Yes, there are no randomized studies except for the Belarus PROBIT study which is randomized on a community level instead of an individual level. And it is the only study we have that uses an intention to treat analysis.

  • Roadstergal

    “To my knowledge, that study hasn’t been done for ethical reasons.”

    It’s a bit complicated by confounders, but we already have that comparison in parts of the developing world where they don’t have access to clean water. But a few confounders never bothered the lactivists before…

    • Cobalt

      The reason we know it’s impossible (for ethical reasons) do Dr. A’s described study is because we already KNOW babies will die, based on simple observation of areas and times where formula is unavailable.

  • Daleth

    *Forehead smack* oh duh! I never thought of this before. So breastfeeding studies include the following kids in the “formula fed” group:

    – Babies who couldn’t BF well (which includes most babies with Down’s syndrome, doesn’t it? They tend to have muscle tone issues that make BF’ing difficult for them)…

    – Babies who were formula fed or combo-fed because they were premature and in the NICU (which includes babies who were extreme preemies, with a much higher risk of developmental and other health issues)…

    – Babies who were formula fed because their mothers had health issues that required medication incompatible with breastfeeding (including mental health issues)…

    …and the studies compare those babies’ health, development and academic performance with those of babies who don’t have Down’s, weren’t premature or at least not extremely premature, and had moms without serious health problems.

    Um!

    • Spamamander

      At the risk of harping on the subject, your #1 is correct. My daughter was 8lbs 2oz when born (she wasn’t even ON the chart for newborns with Down syndrome! The nurses in NICU felt awful they couldn’t feed her initially until tests were done). Even with her size there was lack of muscle tone in her mouth and tongue that kept her from really getting a latch. Even in adulthood with years of occupational therapy people with DS often have a hard time keeping their tongue behind their lips and enunciating certain sounds.

      • MaineJen

        My aunt had Down Syndrome; even at age 50 she spoke with a pronounced lisp and couldn’t say certain words. She definitely knew how to make herself understood, though. And she had enough dexterity in her fingers to do crochet and latchhook rugs, her favorite hobby. 🙂

      • Sarah1035

        My oldest son has 16p11.2 microdeletion, a newish genetic discovery with a highly variable presentation. But one relatively common feature is poor infant feeding. I beat my head against the wall battling poor production on my part for 8 weeks. When I switched to formula I found he couldn’t drink well from a bottle either. His latch looked fine and he never had “low facial tone” but he just lacked suction power. He grew very slowly until he started solid foods. I just had another son on June 10th and he’s already growing out of clothes. It’s so wierd to have a baby that can eat well.

        • Cobalt

          Congrats on the newbie! And yes, it is disorienting to have a “typical” baby after only ever parenting one that is particularly challenging in some way. My first had a list of food sensitivities, digestive issues, and subsequent complications in growth and health. The next was a fantastic and unfussy eater without health issues, and I didn’t know what to do half the time because I didn’t need to do all the extra stuff the oldest required. I worried over not having anything to worry about!