Is Breastfeeding Always Best?

  • Cohort studies have persuaded new mothers to breastfeed over bottle feeding while many studies show confounding results
  • Children who are formula fed can have excellent health outcomes, often not any worse than breast fed children

These days it seems unquestionable that breastfeeding is superior to formula. The American Academy of Pediatrics (AAP) recommends exclusive breastfeeding for six months, when possible. Nursing has been reported to help babies grow up smarter (higher IQ’s) and confer on them a number of health benefits. Breastfeeding advocate group La Leche League states on their website that: “breastfed babies have a decreased likelihood for allergies and dental caries…appropriate jaw, teeth and speech development as well as overall facial development. This means that people who were artificially fed may experience more trips to doctors and dentists”. Breastfeeding has been reported to decrease the prevalence of asthma, allergies, illness with diarrhea and many more conditions.

However, when one looks closer at the evidence, it becomes significantly less clear that nursing offers any advantage over formula at all.

There are huge discrepancies in nursing rates across populations of different ethnic, racial and socioeconomic backgrounds. For example, the overall rate of American moms who have breastfed for any length of time is 75%. However, when that figure is broken down by socio-economic differences, some crucial discrepancies come to light. Women who were eligible to receive WIC services (lower income) were tallied at 67% breast feeders but those who were not eligible (higher income) were reported at 85%. There are also significant differences across people of different racial and cultural groups. As we’ll see, these discrepancies are problematic for attaining accurate information in the studies that have been done.

Most studies looking at breastfeeding and its potential health advantages compare two groups: babies who breast feed versus babies who formula feed.  The idea is to show that one method of feeding leads to better health outcomes over the other. Those outcomes include: number of respiratory infections, the child’s IQ, how likely they are to be obese and several others.

But there’s a problem with that approach: people who come from certain social, cultural and economic backgrounds are more likely to nurse their children. Wealthier parents choose nursing more frequently than their poorer counterparts. In addition, education seems to play a crucial role: the more a woman is educated, the more likely she is to nurse her children.


Photo Credit: Flickr/Nerissa’s ring  (cc license). 

As economist Emily Oster has written “in the U.S. (and most developed countries), white, wealthy women with a lot of education are much, much more likely to nurse their babies than the rest of the population. But these demographic characteristics are also linked to better outcomes for infants even independent of breastfeeding.”

The problem we run into is that when looking at those health outcomes, it’s hard to tease out whether it was the method of feeding that led to them, or something else in the child’s life and environment. It has been well documented that children of better educated, wealthier parents, generally experience more positive health outcomes. Conversely, children whose parents have attained lower degrees of education and/or tend to live in relative poverty, have worse health outcomes.

In general, breastfeeding is a difficult topic to study not only because of all the confounding results one encounters. Randomized trials (studies in which the people being studied are randomly assigned to a breast or formula feeding group) give the highest quality data but are nearly impossible and likely unethical to conduct. To study breastfeeding benefits with a randomized trial, researchers would have to assign parents to breast or bottle feeding using a lottery. It is hard to imagine parents agreeing to “OK Mrs. Smith, the computer picked formula for you, so we’re going to ask you to formula feed your baby for the next year or so.”

As a result, there are essentially no randomized studies on this topic, and that really hurts the quality of the information existing breastfeeding studies provide.

The AAP admits in its own breastfeeding recommendations “Major methodological issues have been raised as to the quality of some of these studies, especially as to the size of the study populations, quality of the data set, inadequate adjustment for confounders, absence of distinguishing between “any” or “exclusive” breastfeeding, and lack of a defined causal relationship between breastfeeding and the specific outcome. In addition, there are inherent practical and ethical issues that have precluded prospective randomized interventional trials of different feeding regimens. As such, the majority of published reports are observational cohort studies and systematic reviews/meta-analyses.” In other words, they are admitting that the science they have used to make their recommendations is particularly flawed.

In reality, the studies that have been done on breast feeding so far are essentially all non-randomized (the subjects choose for themselves formula versus breastfeeding). That creates many problems for the quality of the results. It is very hard to account for all the factors that contribute to a woman’s decision to nurse versus formula feed. For example, if we’re thinking about how infants feed and their IQs as they get older, we have to be aware of what other forces in their lives might be influencing their IQ, such as the mother’s educational level.





Photo Credit: Flickr/myllissa (cc license

In addition to the problem of confounders, the studies generally used to make recommendations are small in size which makes them less powerful and useful.

One more problem is that almost all the research ever done on the topic of breast versus formula feeding and how it affects children’s health has been retrospective. Retrospective (from the Latin “to look back”) studies collect their data from information that’s already out there before the research began. It’s generally accepted that these types of studies make it much more likely for the results to be misinterpreted due to problems with researcher bias. When looking at data in retrospect, researchers who desire a particular outcome (such as “breastfed babies have less infections”), are more likely to focus on the data that supports their wishes.

In situations where confounding variables are difficult to control for, scientists look for creative ways to produce cleaner results. Three such studies are worth looking at. One is the Scandinavian PROBIT trial. While the women in this study chose initially whether to breast or bottle feed, they were later assigned  to one of two groups. In one group, the researchers followed women who were breastfeeding initially and were given support and encouragement to continue to do so. The other group was not given this type of support. In a way, this study did what we’ve been looking for: randomly assigning women to two groups and following them over a period of time.

As one would expect, in the group that received more support and encouragement more women continued breastfeeding longer than in the non-supported group. Then, the researchers looked at health outcomes for the babies. The results showed something very  different from what was previously believed in the past. While two conditions seemed to slightly benefit from breastfeeding (diarrhea and eczema), several others did not. For ear infections, croup, colds and wheezing, both breast and formula fed babies had the same number of illnesses.

pexels-photo (6)

Photo Credit: Pexels/Roman Pohorecki

The most compelling evidence that breastfeeding may not offer as robust of an advantage over formula as previously touted came in 2014 with the publication of a very well designed and powerful study. In “Is Breast Truly Best?”, the researchers set out to remove the effects of confounders on the data. They did this with two approaches. First, as their data source, they used a the NLSY79 study which followed about 12,000 children from all walks of life from birth until age 14. This study was particularly helpful because it was large, followed very diverse groups of families and, most importantly, was prospective.

The second approach these researchers took was to only look at sibling comparisons in the NLSY79 data. In this approach one would look at a sibling pair in which one child was breastfed but the other was formula fed. In this scenario, one could argue, both children were raised in approximately the same conditions, to the same parents. If one was formula fed and the other was breast fed for some reason, then that comparison would eliminate many confounders. Cynthia Colen and David Ramey, the study’s authors, did just that. They focused on sibling data in situations where each child was fed differently.

The authors describe what happens when they specifically look only at the sibling comparison. Compared to looking at the entire population of the study, siblings or not, the results are stunning. When they looked at “between-families” (i.e. Mrs. Jones’ kids compared to Mrs. Smith’s kids), the data showed that breastfeeding was “associated with beneficial long-term child outcomes” This trend was evident for 10 out of the 11 outcomes examined here.

But when they flipped the statistical sibling switch on and accounted for all the confounders, they reported that “when we more fully account for unobserved heterogeneity between children who are breastfed and those who are not, we are forced to reconsider the notion that breastfeeding unequivocally results in improved childhood health and well being.”

So in essence, what we know about breast feeding is based on many studies, but those studies suffer from significant quality issues. This doesn’t mean that women shouldn’t try to nurse when and if they desire to. One thing for sure is that breast milk is free and always available. In addition, for some women, nursing can be a joy and offers a break in the day when mother and child can be physically close and bonded.

However, realizing that many of the reported benefits of human milk are based on flawed data does mean something important: children who are formula fed can have excellent health outcomes, often not any worse than breast fed children. The pressure put on mothers to nurse their babies can sometimes be a major source of grief after the babies are born and as we can see, is generally unwarranted. In the past 10 years or so, breast feeding has been touted for having numerous health advantages for children. However, when better designed studies are performed, the advantages of breastfeeding appear less impressive.

_ _

Cover Photo: Pexels/CCO license 
About the Author /

Dr. Emiliano Tatar received his undergraduate degree from Temple University and he went to medical school at Drexel University. He went on to pursue his residency at St. Christopher's Hospital for Children and for a year worked at the Children's Hospital of Philadelphia in the Pulmonary Department. He currently works in primary care at the Einstein Healthcare Network. He writes for the Philadelphia Inquirer Healthy Kids blog. In his spare time, he enjoys foreign films, cheese, and playing in a band. He lives in suburban Philadelphia with his wife and two kids.

Post a Comment

Scroll Up