Breastfeeding and Childhood Obesity
Obesity is of especially great concern because it is closely connected with greater likelihood of various serious diseases, including diabetes. It is with good reason that the CDC refers to an "obesity epidemic," pointing out that "obesity costs this country about $150 billion a year, or almost 10 percent of the national medical budget.... Obesity is epidemic in the United States today and a major cause of death, attributable to heart disease, cancer, and diabetes." (2b)
The cause of this epidemic is unknown. There are speculations, including one that is supported by seemingly authoritative sources, to the effect that "children who were not breastfed are at increased risk of obesity" (as stated on the website of the American Academy of Family Physicians). U.S. Surgeon General Regina Benjamin shares this opinion, but she acknowledges that this is only an inference, based on observational studies.(2) The U.S. Agency for Healthcare Research and Quality points out that such studies are subject to false conclusion.(3)
Section 1: Historical data on this subject
How false the conclusion was in this case can be observed by looking at historical health data for the period covering the transition from low rates of breastfeeding to high rates, from the 1960's to at least the early 2000's.
Notice in the left-hand chart that breastfeeding rates in the U.S. before 1972 were stable and not increasing. Then a rapid increase began in 1972.
Childhood obesity data from the U.S. Centers for Disease Control and Prevention are shown in Figure 2 below. Notice that
- - obesity levels were low and not increasing among 6-to-11-year-olds as of the second time period shown, in comparison with the earlier period;
- - a rapid increase began as of the very next time period shown (the third one); and obesity continued to increase rapidly after that.
- - Looking at Figure 1, above, in relation to the sharp increase in obesity seen in Figure 2, observe that a major increase in breastfeeding occurred 6 to 8 years before the period of the major increase in obesity among the 6-to-11-year-old age group.
It is worthwhile to stop and think about that: Stable, low obesity rate while breastfeeding was stable and low, then a major increase in obesity in the very first measurement period in which a large number of the children in that age group would have gone through their infancies during a time of greatly increased breastfeeding.
- - As the infancies of a higher percentage of that age group came to be within the period of higher breastfeeding, and as breastfeeding rates during their infancies had come to be higher, their obesity percentages rose still farther.
Among 12-to-19-year-olds, the increases in obesity were very similar to the above, but with a lag that was commensurate with the older age of this group. Children in this group as of 1976-1980 would not have been infants during the period of increased breastfeeding, but those in this age group as of 1988-1994 would have been infants during times of much greater breastfeeding. Observe the doubling of the obesity rate of this age group as of the 1988-1994 period, precisely at the time of the transition from infancies without increasing breastfeeding to infancies with greatly increased breastfeeding. And children in this age group as of the next period after that (with 16% obesity) would have been infants during a period of a still higher rate of breastfeeding.
A very similar pattern is visible among 18-to-29-year-olds, as shown in this chart. Between 1971 and 1980, when that age group's infancies would have been unaffected by the increase in breastfeeding, their obesity percentage remained a flat and relatively low 7.9%. As with the other age groups, in the very first period by which time a significant percentage of this age group's infancies could have been affected by the increase in breastfeeding (1988-1994), their obesity percentage changed from flat to a major increase (nearly doubling in this case). And obesity continued to increase as a higher proportion of this group's infancies came to coincide with the period of higher breastfeeding, until the group's obesity percentage had tripled, as of 2005-2006.(4)
The three age groups for which data is shown above had increases in obesity between 1971-74 and 2003-2006 of 325%, 289%, and 206%, in order from youngest to oldest. Notice that these percentages go from high to lower in correlation with the groups' varying completeness of exposures (during their infancies) to the higher levels of the increases in breastfeeding.
Obesity during this period also increased among the older age groups, but the increases among the older people were (a) much smaller than those of the age groups affected by the increases in breastfeeding, and (b) largely explainable by the transition from farm and factory work to sedentary work. Agricultural employment in the U.S. declined 42% between 1960 and 1990,(4c) and manufacturing employment also decreased greatly as a percentage of the labor force (see red line in chart).
Think about the fact that the 6-to-11-year-olds were obviously not affected by the dramatic reduction in manufacturing and agricultural jobs that affected the older groups, and yet their obesity incidence went up 2.4 times as much as the obesity increase among the older groups. (325% compared with an average of 137% during that same period for the 35-to-74-year-olds).(4c1) There had to have been something that has specially affected the people who were born in recent decades. Nobody seems to think it has been due to some kind of microbial exposure. As indicated earlier in this article, the timing of the increases in obesity has been remarkably compatible with the timing of the increases in breastfeeding, in all of the age groups that have been born since breastfeeding started increasing greatly.
For information about chemicals in breast milk that are linked with increased obesity following increases in breastfeeding, see Section 4 below.
Having seen (earlier) how accurately increases in obesity have followed after increases in breastfeeding, remember the opinions of the American Academy of Family Physicians and Surgeon General Benjamin that "children who were not breastfed are at increased risk of obesity." Is it possible that there has been a bandwagon effect of people with emotional feelings in favor of breastfeeding, who have arrived at conclusions based on what they expect to see rather than on the basis of good evidence? This is only the beginning of the evidence indicating that the claims about benefits of breastfeeding are completely at odds with historical fact. For additional evidence along such lines, go to www.breastfeedingprosandcons.info.
. . . . . . . . . . . . . . . . . . . . .
Section 2: Scientific studies on breastfeeding and obesity that avoid the weaknesses of the other studies
As mentioned briefly in the introduction above and in more detail at www.breastfeeding-benefits.net, the studies that are quoted by breastfeeding's promoters have serious weaknesses, including that they are all observational studies and are therefore very subject to confounders that could be the underlying actual causes of the "associations" that are found. As indicated in the Surgeon General's Call to Action to Support Breastfeeding and elsewhere, mothers of higher socio-economic strata in the U.S. and much of Europe are far more likely to breastfeed than mothers of lower socio-economic strata. And obesity is well-known to be far less prevalent in children of mothers of higher socio-economic strata than in children of lower socio-economic strata.(4c2) An "association" has been found in those countries between more breastfeeding and less obesity, causing many to think that the reduced obesity results from breastfeeding, rather than from the confounding fact that the breastfed children are disproportionately of the class that always has less obesity.
There are several studies, including two especially large ones, that have taken effective steps to avoid those U.S.-European confounders; and, not surprisingly, those studies have arrived at findings that are the opposite of the studies that are selectively quoted by breastfeeding's proponents.
a) First note that (1) size of a study is a very important factor in determining its accuracy, and (2) randomization is considered to be the "gold standard" in scientific studies as a means of avoiding the effects of confounders. The largest study yet conducted on the health effects of breastfeeding, published in the Journal of the American Medical Association, was also the only such study that has utilized randomization: the PROBIT study conducted in Belarus, with over 17,000 mother-infant pairs, following the children's outcomes to age 11.5. The researchers gathered data concerning an experimental group whose breastfeeding rates were greatly increased by major promotion of breastfeeding, comparing that with data from a control group whose breastfeeding was relatively minimal. The odds ratio for overweight/obesity in the experimental group was 1.18, and the odds ratio for obesity alone only was 1.17. (4d) In other words, there was a 17% increase in likelihood of becoming obese among those children who had been more highly breastfed. The authors pointed out that their analysis "may underestimate the effect of the true exposure of interest (breastfeeding exclusivity and duration), owing to overlap in breastfeeding between the randomized groups (many intervention mothers did not exclusively breastfeed for 3 or 6 months, whereas some control mothers did)." Using other statistical methods to analyze their data in different ways (shown in their Table 4), the researchers arrived at alternative figures, in this case comparing children breastfed for less than 3 months with children breastfed for two different periods of greater duration. Utilizing one of their two means of adjusting their data, they estimated that children breastfed for between 3 and 6 months were about 51% more likely to become obese than those breastfed for less than 3 months, and children breastfed for over 6 months were 98% more likely to become obese than those breastfed for less than 3 months. Using the other of their two ways of adjusting the data for this comparison, the increases linked with greater durations of breastfeeding were 45% and 80%.
To summarize the data from this uniquely large study, published in a prestigious medical journal, which was also the only study to utilize the recognized best means of avoiding the error-producing effects of confounders: The findings were analyzed in various different ways, and the only disagreement among their various estimates was as to whether more extensive breastfeeding led to large increases in obesity or extremely-large increases in obesity. A high proportion of the estimates fell into the extremely-large category. And their estimates showed a dose-response relationship, with progressively greater durations of breastfeeding being linked with progressively greater incidence of obesity.
b) A study of 4442 children in Brazil (Pelotas area) considered the contrasting findings in different studies of the relationship between breastfeeding and obesity; it appeared that the contrasting findings depended on the particular characteristics of the societies in which the studies are conducted. Those differing social characteristics affect the nature of the confounders in the different societies, the confounders that can cause what the AHRQ calls "false conclusions" in typical (observational) studies. The chart below shows some data from the ALSPAC study in England, indicating a strong positive relationship between breastfeeding rates and income as that relationship is typically found in the U.S. and much of Europe; it also shows the lack of such a relationship as was the case in Pelotas. As mentioned, higher socio-economic status is a strong predictor of lower obesity incidence. Low rates of obesity in the ALSPAC study were found to be associated with high rates of breastfeeding; and many observers saw this as evidence that obesity is reduced by breastfeeding, overlooking the link between low obesity and the higher income levels that are disproportionately characteristic of the high breastfeeders. Higher income levels are undisputedly a strong predictor of lower obesity,(4c2) and it is entirely likely that higher income levels were the actual determinant of the reduced obesity among the high breastfeeders. For the Pelotas children, in a region where there was no confounding by income levels in this regard (see chart below, showing no relationship between income and breastfeeding rates), the "breastfeeding association effect size" was a positive .14, compared with a negative .16 in the ALSPAC study. (4f)
c) A study of over 8000 children was conducted in Hong Kong, where the greater likelihood of breastfeeding on the part of more-educated mothers did not exist; the follow-up at age 7 was of over 7000. That study calculated the relationship between breastfeeding and overweight among children in various different ways, and none of their different calculations found beneficial associations of breastfeeding with obesity incidence. One of their calculations (results shown in their Table 2, bottom) found that "presence of overweight" among children who had been exclusively breastfed for three or more months was 19% greater than the overweight incidence among children who had never been breastfed. Another calculation (their Appendix Table A1) found that overweight was 27% greater among children who had been exclusively breastfed for three months or more than among never-breastfed children.(4e) In what is a very understated way of discussing the problem of false conclusions arrived at in the U.S.-European studies on effects of breastfeeding, the authors of the Hong Kong study concluded with a "key message:.... Studies in populations with a different confounding structure may be valuable in clarifying and reconciling potentially confounded epidemiological associations." Again, in a region in which the confounder of income-influenced breastfeeding rates was not present, breastfeeding was associated with significantly increased obesity.
d) Another study analyzed sibling pairs as a means of reducing effects of confounders. Comparing different health outcomes among members of sibling pairs minimizes confounding due to socio-economic factors, such as the far higher rates of breastfeeding among higher-income mothers and the lower rates of obesity among children of higher income mothers. As also occurred in the other cases (above) in which effective means were utilized to reduce effects of confounding, the results went in the same direction: When comparing siblings both of whom were overweight, the heavier sibling was almost four times as likely to have been breastfed as the lighter sibling.(4h)
It is worth reviewing the four above studies and seeing how similar their results are, all finding substantially increased childhood obesity where breastfeeding is substantially increased, as long as measures are taken to avoid the typical confounders that are present in the U.S. and European studies. The various means of avoiding confounding were (a) randomization (the Belarus study), (b) and (c) by conducting the study in an area where the socio-economic confounders are not the same as they are in the U.S. and Europe, and (d) by means of sibling comparisons. All were effective in varying degrees and all led to findings of increased obesity linked with increasing breastfeeding. It is noteworthy that the study that was by far the largest and that used the "gold standard" for preventing confounding also came up with some results showing extremely high increases in obesity linked with increasing breastfeeding.
The above should be seen in relation to the close, detailed correlations seen in Section 1 between major increases in breastfeeding in the U.S. and major increases in obesity.
. . . . . . . . . . . . . . . . .
For those who believe that the increased obesity found to be "associated" with low breastfeeding in the U.S. is anything other than misleading, the following is of relevance: even that increased obesity, of questionable origin, is apparently only relatively temporary, anyway. A study that investigated the association between infant feeding and obesity among 35,526 women (participants in the Nurses' Health Study II) found the following: "Exclusive breastfeeding for more than 6 months was associated with leaner body shape at age 5 ... compared to women who were not breastfed or breastfed for less than 1 week, but this association did not persist during adolescence or adulthood."(4g)
. . . . . . . . . . . . . . . . .
Section 3: Apparent obesity-related effects of milk from diabetic mothers
There appear to be strong effects of breast milk from diabetic mothers in increasing risk of overweight in the breastfed child, as found in two German studies. In one of the studies, a positive correlation was found between volume of breast milk consumed by a child of a diabetic mother and the risk of the child's being overweight at age 2. Among children of diabetic mothers, the percentage of overweight among those who had been fed no breast milk was 13%, the percentage of overweight among those who had been fed "some" breast milk from their mothers was 25%, and the risk of overweight among children who had been fed "only" their mothers' milk was 38%.(4i) In another German study, breastfed offspring of diabetic mothers were again found to have an increased risk of overweight, with an odds ratio of 1.98,(4j) an increased risk of 98%. Most studies of the relation between breastfeeding and obesity in Germany have apparently found that higher breastfeeding is associated with lower obesity, implying that Germany has the same confounders that are characteristic of the U.S. and much of Europe; therefore these findings of greatly increased overweight among breastfed children of diabetic mothers mean that the effect of the "diabetic breast milk" were probably very strong, since those effects still stood out strongly even after confounders would have pulled the results in the opposite direction.
. . . . . . . . . . . . . . . . . . . . .
Section 4: Obesogens in breast milk:
For those who think that obesity results simply from eating too much or eating too much of the wrong foods, studies have found that there are "genetic mechanisms that cause extreme obesity" which "are similar to those that cause milder forms of overweight and obesity."(4b) Development of such genetic mechanisms in the body is presumably subject to the effects of developmental toxins, such as those that are known to be concentrated in breast milk. According to a major study on this matter, "exposure to toxicants during the organogenesis of tissues involved in metabolic homeostasis, e.g. adipose, liver, skeletal muscle, pancreas and brain, may play an important pathophysiological role in the development of childhood obesity. While much of organogenesis occurs prenatally, adipose, skeletal muscle, pancreas, and brain continue to develop postnatally" (5) The first toxicants mentioned by the authors as possible obesogens were dioxins and dioxin-like PCBs, pointing out, "Cross sectional studies of adult PCB exposures have consistently shown a positive association with measures of obesity." Bear in mind the very high concentrations of PCBs in human milk and the 4˝-times-greater concentrations of PCBs found in breastfed children than in formula-fed children even at 3 1/2 years of age; this is as summarized by the U.S. Agency for Toxic Substances and Disease Registry, quoted in Section 1.a of www.breastfeeding-toxins.info . Also note in Section 1.a of that same website, quoting from a major study, "Much higher doses of organochlorine compounds (from 10 to 20 times higher) penetrate the infant's body via the (mother's) milk than via the transplacental route." (Note that PCBs and dioxins are organochlorine compounds.)
The authors of the above-cited study on chemical toxins that contribute to obesity also point out, "There are a number of anti-psychotic and anti-diabetic drugs that also increase obesity in children and adolescents…." Note in the Section 1.c of www.breastfeeding-toxins.info the many drugs about which the American Academy of Pediatrics expresses concern with regard to their transmission from a mother to a nursing infant.
But there is also reason to see a possible connection of obesity with infant feeding in a non-chemically-related respect. In earlier decades, feeding infants on schedule, as opposed to feeding on demand, was apparently the norm.(6) Feeding on demand became the norm at some time during recent decades, and it is possible that this may have helped lead to increases in obesity. It may be that an infant who never has unsatisfied hunger pangs for long does not develop tolerance for mild hunger. It seems to be common knowledge that a child who is accustomed to getting what he wants is less likely to develop a tolerance for not having his wishes fulfilled (the "spoiled child"). It would not be surprising if there is a similar effect resulting in some children failing to develop a level of comfort with going without a full stomach for a suitably long time.
In a 2012 study it was pointed out that “recent studies have highlighted an unexpected implication of POPs (persistent organic pollutants) in the development of metabolic diseases like type 2 diabetes and obesity.”(7) Bear in mind that POPs are the same kind of pollutants that are also described as “persistent organohalogens” in the following authoritative statement:
"Significantly more (10 to 20 times) of a mother's body burden of persistent organohalogens is transferred to the infant via the milk than by the transplacental route,"(8)
and also note that POPs include dioxins, PCBs and PBDEs, which are present in human milk in concentrations scores to hundreds of times higher than in infant formula. (see www.breastfeeding-toxins.info)
In conclusion, to repeat an important message: The above is only the beginning of the evidence indicating that the claims about benefits of breastfeeding are completely at odds with historical fact. For additional evidence along such lines, go to www.breastfeedingprosandcons.info.
Message to health professionals and scientists reading this paper: This author cordially invites you to indicate your reactions to the contents presented here. As of now, new parents almost never hear anything but completely one-sided promotion of breastfeeding, with no mention of possible drawbacks except in cases of serious problems on the part of the mother. If you feel that parents should be informed about both sides of this question and thereby enabled to make an educated decision in this important matter, please write to the author of this paper. Also, if you find anything here that you feel isn't accurately drawn from trustworthy sources or based on sound reasoning, please by all means send your comments, to email@example.com.
Comments from readers:
From this paper's inception in early 2012 until present, the invitation has been extended to all readers to submit criticisms of contents of this paper, asking them to point out how anything written here is not well supported by authoritative sources (as cited) or is not logically based on the evidence presented. As of May 4, 2013, after more than a year, no criticisms of contents of this paper have yet been received in response to that invitation. (That is significant, considering the thousands of visits we receive from readers every month.) We have received some e-mails that have not criticized contents of this paper but which are of interest; several of those comments or inquiries and our responses to them are entered at www.pollutionaction.org/comments.htm . All comments are welcome, especially those that point out any deficiencies in our evidence in relation to conclusions drawn or any lack of quality in the reasoning as presented. Please send comments or questions to firstname.lastname@example.org .
* For information about those providing the above information, Pollution Action, go to http://www.pollutionaction.org
Some of the full articles below are available for free online, but to obtain the full text of some of these articles for free, you may have to visit a university library or ask at the reference desk at your local public library.
(1) "Breastfeeding, Family Physicians Supporting (Position Paper)" -- AAFP Policies -- American Academy of Family Physicians
(2) "Surgeon General's Call to Action to Support Breastfeeding, 2011," p. 33 at http://www.surgeongeneral.gov/library/calls/breastfeeding/calltoactiontosupportbreastfeeding.pdf
(2a) Arenz S, Ruckerl R, Koletzko B, von Kries R. "Breast-feeding and childhood obesity—a systematic review." Int J Obes Relat Metab Disord 2004;28:1247–1256.
(3) Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services, "Systems to Rate the Strength of Scientific Evidence, Evidence Report/Technology Assessment: Number 47" http://archive.ahrq.gov/clinic/epcsums/strengthsum.pdf
(3a) A History of Infant Feeding J Perinat Educ. 2009 Spring; 18(2): 32–39. doi: 10.1624/105812409X426314 PMCID: PMC2684040 Emily E Stevens, et al., at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2684040/
(4) CDC's Health United States, 2008, Data Table for Figure 7. At http://www.cdc.gov/nchs/data/hus/hus08.pdf That specific data table is shown at http://www.pollutionaction.org/HUS08Datatabl.jpg
(4a) Owen CG, Martin RM, Whincup PH, Smith GD, Cook DG. Does breastfeeding influence risk of type 2 diabetes in later life? A quantitative analysis of published evidence. Am J Clin Nutr 2006;84:1043–1054.
(4b) Findings shed light on overlap between extreme obesity and mild overweight researcher says By Robert Preidt Monday, April 8, 2013 Health Day of Medline Plus, of NIH Citing study published online April 7, 2013 in the journal Nature Genetics, quoting study coordinator Erik Ingelsson, professor at Uppsala University, Sweden URL of this Medline Plus page until July 2013: http://www.nlm.nih.gov/medlineplus/news/fullstory_135660.html
(4c) Monthly Labor Review, Nov. 1993, p. 7, at http://www.bls.gov/mlr/1993/11/art1full.pdf of U.S. Bureau of Labor Statistics website.
(4c1) CDC's Health United States 2008, Table 75, and Data Table for Figure 7.
(4c2) Poor Neighborhoods Home to More Obese Kids: Study Researchers find link between weight and the economic and educational status of the community HealthDay News, Nov., 2012 at http://consumer.healthday.com/public-health-information-30/demographics-news-173/poor-neighborhoods-home-to-more-obese-kids-study-670620.html ; also Dana Hughes, DrPH, Mary Kreger, DrPH, et al.: Reducing Health Disparities Among Children: Strategies And Programs For Health Plans. At http://nihcm.org/pdf/HealthDisparitiesFinal.pdf;
(4d) Effects of Promoting Longer-term and Exclusive Breastfeeding on Adiposity and Insulin-like Growth Factor-I at Age 11.5 Years A Randomized Trial. Richard M. Martin, PhD et al. JAMA. (Journal of the American Medical Assn.) 2013;309(10):1005-1013. doi:10.1001/jama.2013.167. March 13, 2013, Vol 309, No. 10 Found at http://jama.jamanetwork.com/article.aspx?articleid=1667089
(4e) Kwok MK, et al., Does breastfeeding protect against childhood overweight? Int J Epidemiol. 2010;39(1):297-305 found at http://ije.oxfordjournals.org/content/39/1/297.full.pdf
(4f) Brion MJ et al. What are the causal effects of breastfeeding on IQ, obesity and blood pressure? Int J Epidemiol. 2011;40(3):670-680 found at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC314707
(4g) Michels KB, et al, A longitudinal study of infant feeding and obesity throughout life course. Int J Obes (Lond). 2007 Jul;31(7):1078-85. Epub 2007 Apr http://www.ncbi.nlm.nih.gov/pubmed/17452993
(4h) Nelson MC et al., Are adolescents who were breast-fed less likely to be overweight?: Analyses of sibling pairs to reduce confounding , Table 3 Epidemiology. 2005;16(2):247-253 found at http://journals.lww.com/epidem/Fulltext/2005/03000/Are_Adolescents_Who_Were_Breast_fed_Less_Likely_to.15.aspx
(4i) Long-term impact of neonatal breast-feeding on body weight and glucose tolerance in children of diabetic mothers, Plagemann A et al., Table 3 Diabetes Care. 2002 Jan;25(1):16-22. Found at http://www.ncbi.nlm.nih.gov/pubmed/11772895/
(4j) Long-term impact of breast-feeding on body weight and glucose tolerance in children of diabetic mothers: role of the late neonatal period and early infancy. Rodekamp E et al., Diabetes Care. 2005 Jun;28(6):1457-62. Found at http://www.ncbi.nlm.nih.gov/pubmed/15920068
(5) CHILDHOOD OBESITY AND ENVIRONMENTAL CHEMiCALS, Michele La Merrill et al., Mt Sinai J Med. 2011 Jan–Feb; 78(1): 22–48. doi: 10.1002/msj.20229 PMCID: PMC3076189 NIHMSID: NIHMS253603
(7) J. Ruzzin, Public health concern behind the exposure to persistent organic pollutants and the risk of metabolic diseases BMC Public Health. 2012; 12: 298. Published online 2012 April 20. doi: 10.1186/1471-2458-12-298 PMCID: PMC3408385 at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3408385
(8) Jensen, A.A. et al, Chemical Contaminants in Human Milk, CRC Press, Inc., Boca Raton, Ann Arbor, Boston, 1991, p 15., p 15. Findings of above confirmed in animal tests, with even greater contrasts, in Ahlborg et al., Risk Assessment of Polychlorinated Biphenyls (PCBs), Nordic Council of Ministers, Copenhagen. Report NORD 1992; 26