A study published in Pediatrics is claiming that bottle feeding might be a risk factor for pyloric stenosis (a condition in which the pyloric muscle malfunctions and impedes food (liquid) from going down into the small intestine, causing severe projectile vomiting). Actually, if you read the study, it seems that the authors are suggesting that formula feeding may cause pyloric stenosis; they argue that the rate of the condition went down when breastfeeding rates went up in Denmark, and that babies who were formerly breastfed and then switched to formula developed pyloric stenosis after making this dietary change.
The study itself is somewhat convincing. The authors looked at a Danish cohort of 70,148 babies born between 1996 and 2002. Of these babies, 65 eventually had surgery for pyloric stenosis (PS); 29 of these were bottle-fed.
Looking at these numbers, one might be confused – if 29 out of 65 babies were bottle fed, that means the other 36 were breastfed, right? And obviously, 36 is bigger than 29. (I dropped out of math my senior year in high school, but even I know that). However, the art of statistics makes numbers all irrational and annoyingly misleading, and because the number of breastfed babies in the cohort was so much larger than the number of formula fed babies (Denmark has impressive breastfeeding rates), statistically, the formula fed babies had a much greater chance of having PS. A 4.6-fold higher risk, to be exact. Scary stuff.
No need to panic just yet, though. First of all, while the study claims to have controlled for confounding factors, this controlling was based on telephone interviews – some of which were performed retrospectively. This type of data has a bad reputation in the research world, because, in the immortal words of Dr. House, everybody lies. We call it innocuous names like “recall bias”, but it amounts to the same thing. (Think about it: when you fill out those health forms at the doctor, and they ask about alcohol use, do you check “rarely” or “never”? What do those categories even mean? What you might consider “rarely” might be the equivalent of binge drinking to someone else. And don’t even get me started on the questions about smoking. There’s no box that describes “social smoking only when wine was involved in your early 20′s, with the occasional guilty cigarette at times of extreme stress, like once every 3 years”. So I – er, I mean, someone- might lie and say “never”. ) This study was also vulnerable to the same frustrating issues that plague all infant feeding research – it’s impossible to control adequately for everything.
Even so, there is something suspicious about the fact that when some of the PS babies who were formerly breastfed started formula, that was when they suddenly developed PS. And it’s odd that the babies who were never breastfed had the same risk as those who were combo fed or weaned early. If this weren’t the case, I might argue that it was likely that early symptoms of the PS may have provoked a change to bottle feeding; sometimes parents will end up formula feeding if a baby has trouble eating, or is spitting up a great deal, in the hopes that a gentle formula might do the trick.
What strikes me as problematic about this study, though, is that while the cohort includes a substantial amount of babies, the sample of babies who actually had PS is quite small. I also think there is something to be said for looking at a cohort in a country with high breastfeeding rates – because there must have been a reason that these other moms weren’t breastfeeding. I wonder, for example, if they asked about antidepressant use, or fertility drugs, or other medications that might have been taken during pregnancy and that would also lead a woman to either opt not to, or be unable to, breastfeed? Also, 91% of the babies with PS were male; there must be something either structural, behavioral or genetic in male human babies that predisposes them to this condition. The authors discuss this a bit in the report, and muse that it could be because boys overeat more than girls do (not sure if there is substantial scientific evidence for this claim or not); this would support the hypothesis that it is something in the act of bottle feeding, rather than the milk itself, that is conferring a protective effect. Lastly, I wish the researchers had thought to ask for reasons why the parents stopped breastfeeding; this might have led to other associations between the PS and feeding issues.
That said, I actually believe that this is one of the better studies I’ve seen regarding infant feeding. The discussion section takes the time to discuss the possible reasons for a protective effect from breastfeeding (or a negative effect from bottle feeding), elaborating on both the milk as a substance (they suggest that babies fed with breastmilk in bottles may be at just as high a risk, and admit that they did not receive adequate info to make a distinction between expressed breastmilk, donor milk, or formula) and on the act of bottle feeding itself. They also state that “formulas have improved over time and now approach the composition of breastmilk” and that this could be “a contributing reason for the decrease in PS incidence.” (And that leads my overtired brain to form more questions… what about the type of formula being given? Were these babies on soy? Formula with or without DHA? Partially hydrolyzed formulas?)
The ultimate consensus of these researchers is that this study “adds to the evidence supporting the advantage of exclusive breastfeeding in the first months after birth”. This is true, but I’m not quite sure it’s helpful, or even that scientifically prudent. If the aim of infant health studies is to improve infant health, wouldn’t it be more substantive to end with something like….
These findings suggest that there may be a protective effect against pyloric stenosis from exclusive breastfeeding, and should encourage further studies to decipher why this may be. Whether the relationship is merely associative, or causal, remains to be seen; further examination of the biological or mechanical reasons for the reduction in risk might help determine this, and lead to better outcomes no matter what the mother’s choice of feeding may be.
And pigs may fly wearing my daughter’s favorite leopard-print tutu.
In sum, this is an interesting study, and certainly leads to some big questions. But considering the total risk for PS was still only 0.1% in the cohort studied, I don’t think it’s something to really worry about as a formula feeding parent. If you have a family history of PS, and your baby is male, you might also want to choose a hydrolyzed protein formula and be extra cautious about over-feeding, as these factors relate to some of the plausible explanations for reduced risk offered by the researchers.
Hope that offers a bit of perspective on a study which associates yet another poor health outcome with formula feeding. Remember, as always, that associations are not causes – and that breastmilk can end up projectile-vomited all over the wall, too.