Correlations between Breastfeeding and Autism in Recent Decades

and Probable Causes of those Correlations

DP Meulenberg


Both diagnosed autism and breastfeeding rates have been greatly increasing in developed countries in recent decades. But it has not been merely a matter of general increases of both; instead, there have been numerous cases of major locally-varying levels of autism that have correlated closely with locally-varying levels of breastfeeding.  Specific lows, as well as highs, of autism have correlated well with specific lows and highs of breastfeeding, internationally, among U.S. states, and among demographic groups.


There is ample evidence from authoritative sources (and no apparent disagreement) that certain recognized neuro-developmental toxins are ingested in many-times greater quantities by breastfed infants than by formula-fed infants.  Average daily exposure of a breastfed infant to dioxin toxicity over the period of a year, as estimated by the EPA, is over 80 times higher than the reasonably-safe upper threshold of dioxin exposure estimated by the EPA.(1)   For considerable detail about developmental toxins known to be concentrated in breast milk, including authoritative sources of this information, go to www.breastfeeding-toxins.info)


   a) In every developed country in which breastfeeding is known to be high, autism has also been found to be high (the Scandinavian countries, Germany, Switzerland, Austria, Australia, New Zealand, Japan, and South Korea) (for sources, go to Section 1.2.s.5 at this link); in the European countries with low breastfeeding rates, autism rates have been found to be less than half as high (Ireland, France, Belgium, and U.K.) (Section 1.2.s.4 at preceding link); childhood cancer rates, also, show a close correlation with breastfeeding rates (see www.breastfeeding-and-cancer.info ).

   b)  The four U.S. states that are highest in autism (Minnesota, Oregon, Maine and Utah) are also high in breastfeeding rates (Sections 1.2.s.7 and 1.2.x.1 at this link);  of the seven U.S. states that have the lowest rates of breastfeeding, every one of these seven also has an unusually low rate of autism (Kentucky, West Virginia, South Carolina, Ariizona, Alabama, Louisiana and Mississippi)  (Section 1.2.x.5 at preceding link); 

   c)  There is a 50% higher rate of breastfeeding among U.S. whites than among blacks, corresponding with a roughly 50% higher rate of autism among whites than among blacks, and Hispanics are in between in both respects; however, blacks who breastfeed at about the same rate as whites apparently have autism at about the same rate as whites (Section 1.2.s.3 at www.pollutionaction.org/breastfeeding-and-autism-and-cancer.htm);

   d)  Older mothers are more likely to breastfeed, and to do so for longer periods, and toxins built up in their bodies (and therefore in their milk) increase with their years of exposure to the environment; and there is a 20% increase in risk of autism with each 10-year increase in parents' ages (Section 1.2.s.1.b at this link);

image031.jpg   e)   Autism rates have generally been found to be increasing in recent decades, while breastfeeding rates have also been increasing.  A single exception to that generalization has been found, and that exception merits special attention. Notice the conspicuous stability in the rate of breastfeeding in the UK in the years leading up to 2000, in contrast with the general upward trend or high levels in most countries.  A study in the UK found that there had not been an increase in incidence of autism among children born in the years preceding 1999 (3).  A study that failed to find increasing prevalence of autism for any time period during the last two decades appears to be unusual if not unique; and it may be more than coincidental that this study was done in a country in which the prevalence of breastfeeding had also (contrary to the general pattern) not been increasing during the relevant period.


Increasing breastfeeding is associated with increasing autism (most countries).   Stable breastfeeding is associated with stable autism (U.K. during the applicable period.). 


    f)  A 2011 study that investigated data from all 50 U.S. states and 51 U.S. counties found that "exclusive breast-feeding shows a direct epidemiological relationship to autism," and also, "the longer the duration of exclusive breast-feeding, the greater the correlation with autism." 4


Note that, according to the EPA, Epidemiologic studies of exposed human populations provide the most convincing evidence of human health effects.”5a  Also, a dose-response relationship between an exposure and a health outcome is considered to be especially significant evidence to support a finding of cause and effect.  One example of a dose-response relationship, as found in an epidemiological study by a well-published scientist (R.J. Shamberger), was quoted in the previous paragraph.  This finding was even more significant in that it was based on investigation of a very large, diversely-populated geographic area (all 50 U.S. states), and it also applied in relation to numerous smaller-scale units (51 counties). 


Additional support for a causal connection between breastfeeding and autism was provided by three additional, relatively recent studies, with a dose-response relationship being apparent in the overall view of those three studies, with different degrees of correlation with autism according to the different durations of breastfeeding.

-- In a 2011 Canadian study of a population of over 125,000, using discharge from the hospital as the dividing line for breastfeeding exposure, there was a 25% higher autism rate among the breastfed children than among non-breastfed children.5b

-- In a 2009 U.K. study, the duration of breastfeeding that was assessed was four weeks versus less than that, with 65% of the autism cases having received breastfeeding for at least four weeks; that should be compared with only about 28% of the general U.K. infant population receiving that much breastfeeding; that meant a 130% higher odds of having had that much breastfeeding, among those with autism.5c

-- In a 2010 American study in Kentucky by two MD’s, the duration of breastfeeding used for comparison was six months, and 37% of autism cases had received that much breastfeeding, compared with 13% of the controls, indicating an approximately 185% (37%/13%) greater likelihood that the autism cases would have had more breastfeeding; the p-value was .003, meaning three chances out of a thousand that the finding was a result of chance occurrence.5

-- In addition to the above, there are several other relevant studies, which can be found in Section 4 of www.pollution-effects.info


The above correlations should make one think carefully about (a) what could be causal connections, and (2) what else could be causes of the highs and lows in autism and childhood cancer, aside from the known highs and lows in breastfeeding and/or concentrations of toxins in breast milk in those same locations and social groups.  For a causal connection, there are many randomized, controlled, high-quality studies that verify the neuro-developmentally-toxic and carcinogenic effects of specific chemicals of kinds that are known to be contained in typical breast milk. (see www.breastfeeding-toxins.info)


In addition, responsible U.S. government agencies acknowledge that the chemicals mentioned above are concentrated in breast milk and are (a) ingested by breastfed infants in far higher quantities than by bottle-fed infants and (b) ingested in doses many times higher than the level that the EPA has determined to be safe (see the link just above for sources).  This author invites and will publish here any thoughts submitted by readers as to what else, besides toxins in breast milk, could be causes of the associations noted above as well as those to follow. (Please write to dm@pollutionaction.org)




There are many people who feel that considerations such as the above should not cause a mother to decide against breastfeeding, given what are believed to be benefits of breastfeeding.  But a close look at a complete list of the alleged benefits to full-term infants (provided in the U.S. Surgeon General's Call to Action to Support Breastfeeding 2011) reveals that the presumed benefits prove not to be real.   Regarding the diseases and conditions that the Surgeon General has declared to be "risks" of not breast-feeding, there is revealing historical health data from CDC, NIH, and other respected sources.  Meaningful comparison is made possible by considerable government health data for the period of transition from low breastfeeding (before the 1970's) to high breastfeeding (late 1970's and later).  Health data that applies to those born after the transition to higher breastfeeding rates shows the following:  Outcomes have become substantially worse with respect to all but one of the conditions alleged to be "risks" of not breastfeeding, following the transition to higher rates of breastfeeding.  "Epidemics" among children with some of those disorders have been declared following the increase in breastfeeding:  obesity, diabetes, allergies, and asthma,(6) and possibly also autism.  Cancer increased among children during the transition to higher breastfeeding while it declined in the general population; it has been higher in the countries, U.S. states, and demographic groups in which breastfeeding has been higher, and it has been lower in the countries, U.S. states and demographic groups in which breastfeeding has been lower. (see http://www.breastfeeding-and-cancer.info)   Childhood diabetes has been many times higher in high-breastfeeding countries than in low-breastfeeding countries in recent decades; and it has risen especially rapidly among specific demographic groups and in many specific countries in precise correlation with the time periods when those particular demographic groups and countries were undergoing great increases in breastfeeding rates; but it has declined in the only region in which breastfeeding rates were declining. (see www.breastfeedingprosandcons.info/breastfeeding-and-diabetes.htm)  Another claim about breastfeeding is that it reduces likelihood of obesity in the infant, but major historical evidence over the decades (provided by the CDC) has shown that the opposite has occurred, in precise correlation with increases in breastfeeding. (see www.child-obesity.us .)


Surgeon General Benjamin has acknowledged (in p. 33 of her Call To Action) that the evidence that found benefits of breastfeeding consists only of observational studies, merely yielding inferences; the U.S. Agency for Health Research and Quality says that such studies are subject to false conclusion,(7) given the inherent likelihood of "confounding factors" not appropriately considered.  It should not be surprising if false conclusions were reached in this area, given the recognized confounders of low income and smoking known to prevail among bottle feeders, which confounders are known to lead to the same disorders that are typically attributed to bottle feeding.(8)





OR_PCBs      Regarding this map of PCB exposure in the western U.S., bear in mind that (1) dioxins and PCBs from the environment become concentrated in breast milk (www.breastfeeding-toxins.info) (PCBs are normally considered a variation of dioxins, and both are neuro-developmental toxins); and (2) Oregon (the western part of which is shown in red) has one of the very highest rates of breastfeeding in the U.S.  See the text below the map for what may be the outcome of the high rate of breastfeeding in combination with high levels of PCBs in an area.   For more on the subject shown in this map, and for the complete map, see Section 1.2.x.1 at this link.




Some metrics that apply:

1/800,000th of an ounce was the amount of a PCB-containing product administered per day to pregnant or lactating female monkeys that was found to cause the offspring to be hyperactive and retarded in learning ability.  Compare that with 600,000 tons, which is the amount of PCBs produced in the U.S. between 1930 and 1977, which became part of considerable electrical equipment and appliances that are still in use (often leaking) and that are heavily present in landfills, from where the toxins can be released to the atmosphere or water supplies. (see www.babyfeeding.info/toxins-in-breastmilk-and-formula.htm)


The effective concentration of PCBs taken into infants’ bodies in breast milk was 30 times higher than the concentrations that entered the mothers’ bodies, according to one study.  The toxicity-equivalent concentrations of dioxins in formula-fed infants at 11 months of age was ten times lower than those in infants that had been breast-fed for six to seven months, according to a German study. (see www.breastfeeding-toxins.info)  Note that the first year or so of life includes the “window” during which critical development of the brain takes place, if it will ever take place. (Section 1.2.b.1 at this link.)


Regarding the possible benefits to mothers of breastfeeding, those benefits could be real, but for reasons that should make the mother think carefully.  A U.S. study found a 70% decrease in a mother’s body burden of dioxins over a two-year period while she was breastfeeding twins, and another study praised the "decontamination" of the mother resulting from breastfeeding.  That excellent decontamination of the mother results from her excreting toxins from her body into her infant, at what is almost certainly the period of the infant's life of greatest vulnerability to neuro-developmental toxins, such as those concentrated in breast milk.


A fourth child’s average risk of autism is half as high as that of a firstborn.  And the odds of being diagnosed with autism continuously decrease from first to later children.  Infants later in birth order are less likely to be breastfed, they are breastfed for shorter periods on average, and the milk they receive has toxin levels that have been reduced as a result of excretion to earlier-born infants during previous breastfeeding.  (Section 1.2.s.1.c at this link)  What a surprise that later-born infants have progressively lower rates of autism than firstborns, given the known extraordinarily high levels of neuro-developmental toxins in breast milk.


Accumulated dioxin toxic equivalency exposure in infants that had been breastfed for one year was 6 times higher than that in infants that had not been breastfed, in an EPA study.  The average daily exposure of a breastfed infant to dioxin toxicity over the period of a year's breastfeeding was estimated by the EPA to be 86 times higher than the reasonably-safe upper threshold of dioxin exposure estimated by the EPA in 2012.  (60 pg of TEQ/kg bw/day vs. 0.7 pg of TEQ/kg bw/day)  The peak body-weight-based dose received by a breastfeeding infant is estimated in an EPA study to be 242 pg TEQ/kg-day.(1)


For greater detail on the above subject, but with a good introductory summary including links, and also including a discussion of origins of childhood cancer that have much in common with origins of autism, go to www.pollutionaction.org/breastfeeding-and-autism-and-cancer.htm.


The above has been pointed out to high officials of the American physicians' associations and of the U.S. Department of Health and Human Services, and none of them have responded with any criticism of the above.  They point to the "recognized" benefits of breastfeeding and "risks" of not breastfeeding, as presented by U.S. Surgeon General Benjamin.(1.0)  However, looking at historical health data (from CDC, NIH and other authoritative sources) regarding the diseases and conditions referred to as risks of not breastfeeding, it turns out that there has been significant worsening in child health in all but one of those areas during the years of the transition from low rates of breastfeeding to high rates. And this has not merely been a general worsening over decades, while many toxins have been increasing in the environment.  The specific times of the increases of disorders for particular age groups correlated fairly precisely with the specific times of those age groups' earlier increased exposures to breastfeeding; in cases in which breastfeeding rates of particular ethnic, socio-economic and national groups increased especially rapidly, diseases later increased extraordinarily rapidly just for those specific groups.  In the cases of four of the disorders alleged to be improved by breastfeeding, epidemics have been declared due to actual increases in those diseases or conditions, all of which major increases occurred following the transition to higher rates of breastfeeding.  For a complete look at what the historical record shows regarding the alleged benefits of breastfeeding, go to www.breastfeedingprosandcons.info.



Disproportionately high among children of the less-educated: General intellectual disability.(1a) No surprise.

Disproportionately high among children of the more-educated: The recently increasing autism. This is surprising, until one considers the fact that breastfeeding, increasing greatly in recent decades and imparting neuro-developmentally toxic dioxins to infants in doses scores of times higher than the EPA-estimated safe dose, is twice as frequent among college graduates as among high school graduates. (see Section 1.2.s.1.a)

Additional surprising findings about autism, correlating with the increased ingestion of developmental toxins by breastfed infants, can be found at www.pollutionaction.org/surprises.htm .




A question that should be addressed to those who are recommending breastfeeding, but which they probably won't want to answer:


Given (a) the inconclusiveness of the studies that support breastfeeding,** (b) the known concentrations of environmental toxins in recent human milk,** and (c) the many close correlations between variations in breastfeeding levels and similar variations in incidence of many important childhood diseases (seen in national health data**):  Why should we believe that breastfeeding is more beneficial than harmful?


** Supporting information and references to authoritative sources regarding matters raised in this question are included in a one-page printable version of this question, to be found at www.pollutionaction.org/Q.pdf .


We have good reason to say that those who recommend breastfeeding probably will not have an answer to the above question.  A slightly different version of essentially this same question was mailed to four different high officials at the U.S. Department of Health and Human Services, who are heads of divisions that are involved in promoting breastfeeding.  As of 9 and more weeks after mailing those letters, no reply has been received.  Several months earlier, each of those officials had sent one response to an earlier letter that brought up the matters above, and none of their responses said anything in criticism of any of those points.  Those points are all well substantiated.  


The same applies to the doctors' organizations on which the Department of Health and Human Services relies for support of its position.  The director of Pollution Action mailed two or more letters to the offices of each of the American Academy of Pediatrics, the American Academy of Family Physicians, and the American Congress of Obstetricians and Gynecologists, challenging their evidence on which they base their advocacy of breastfeeding, and not one reply has been received as of two and nine months after mailing those letters.


So the question above is a logical question to ask.  But the promoters of breastfeeding appear to be unwilling or unable to respond to it.  If they can't or won't answer that question as part of an informed debate on this matter (therefore to dm@pollutionaction.org, as well as to you), should anybody pay attention to their advice?




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 dm@pollutionaction.org


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 its contents, 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 22, 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 dm@pollutionaction.org .


 * About Pollution Action:  Please visit www.pollutionaction.org




(1) Infant Exposure to Dioxin-like Compounds in Breast Milk,  Lorber and Phillips  Volume 110 | Number 6 | June 2002 • Environmental Health Perspectives  http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=54708#Download   Also EPA Home/Research/Environmental Assessment: An Evaluation of Infant Exposure to Dioxin-Like Compounds in Breast Milk, Matthew Lorber (National Center for Environmental Assessment, Office of Research and Development, U.S. Environmental Protection Agency) et al.

U.S. EPA. Estimating Exposure To Dioxin-Like Compounds - Volume I: U.S. Environmental Protection Agency, Washington, D.C., EPA/600/8-88/005Ca., 2002, revised 2005 – http://cfpub.epa.gov/si/si_public_record_Report.cfm?dirEntryID=43870,  Section II.6, "Highly Exposed Populations" (nursing infants are considered to be one of the highly-exposed populations), 4/94 (p. 39)  "Using these procedures and assuming that an infant breast feeds for one year, has an average weight during this period of 10 kg, ingests 0.8 kg/d of breast milk and that the dioxin concentration in milk fat is 20 ppt of TEQ, the average daily dose to the infant over this period is predicted to be about 60 pg of TEQ/kg-d." 

Also http://www.epa.gov/iris/supdocs/dioxinv1sup.pdf  in section 4.3.5, at end of that section, "...the resulting RfD in standard units is 7 × 10−10 mg/kg-day."   In the EPA’s “Glossary of Health Effects”, RfD is defined:  “RfD (oral reference dose): An estimate (with uncertainty spanning perhaps an order of magnitude) of a daily oral exposure of a chemical to the human population (including sensitive subpopulations) that is likely to be without risk of deleterious noncancer effects during a lifetime.”


(1.0)  The Surgeon General's Call to Action to Support Breastfeeding 2011


(2) Autism rates associated with nutrition and the WIC program.  Shamberger RJ., King James Medical Laboratory, Cleveland, Ohio  J Am Coll Nutr. 2011 Oct;30(5):348-53.  At  http://www.ncbi.nlm.nih.gov/pubmed/22081621


(3) Pervasive Developmental Disorders in Preschool Children: Confirmation of High Prevalence  Suniti Chakrabarti, M.D., et al. Am J Psychiatry 2005;162:1133-1141. 10.1176/appi.ajp.162.6.1133


4) Shamberger, R.J., Autism rates associated with nutrition and the WIC program, J Am Coll Nutr. 2011 Oct;30(5):348-53.  Abstract at www.ncbi.nlm.nih.gov/pubmed/22081621   An image of part of the article is shown below, since it may be expensive for many readers to see the complete study.




(5) Breastfeeding and Autism  P. G. Williams, MD, Pediatrics, University of Louisville, and L. L. Sears, MD, presented at International Meeting for Autism Research, May 22, 2010, Philadelphia Marriot  https://imfar.confex.com/imfar/2010/webprogram/Paper6362.html)


(5a) EPA, The Effects of Great Lakes Contaminants on Human Health, Report to Congress, Section III, p. 12 (bottom) and again on p. 16, at http://nepis.epa.gov/Exe/ZyNET.exe/2000BSHI.txt?ZyActionD=ZyDocument&Client=EPA&Index=1995%20Thru%201999&Docs=&Query=&Time=&EndTime=&SearchMethod=1&TocRestrict=n&Toc=&TocEntry=&QField=&QFieldYear=&QFieldMonth=&QFieldDay=&UseQField=&IntQFieldOp=0&ExtQFieldOp=0&XmlQuery=&File=D%3A%5CZYFILES%5CINDEX%20DATA%5C95THRU99%5CTXT%5C00000002%5C2000BSHI.txt&User=ANONYMOUS&Password=anonymous&SortMethod=h%7C-&MaximumDocuments=1&FuzzyDegree=0&ImageQuality=r75g8/r75g8/x150y150g16/i425&Display=p%7Cf&DefSeekPage=x&SearchBack=ZyActionL&Back=ZyActionS&BackDesc=Results%20page&MaximumPages=10&ZyEntry=20

(5b) Dodds et al., The Role of Prenatal, Obstetric and Neonatal Factors in the Development of Autism, J Autism Dev Disord (2011) 41:891–902  DOI 10.1007/s10803-010-1114-8, Table 6, at http://autism.medicine.dal.ca/research/documents/2011DoddsetalJAutDevDisord.pdf   This 2010 Canadian study, drawing data from a population-based clinically-rich perinatal database,” investigated a very large population, nearly 130,000 births.  Data from almost 127,000 of those children (those without identified genetic risk of autism) went into the study’s finding that there was a 25% increased risk of autism among children who were breastfed at discharge from the hospital. 


(5c)  Whitely et al., Trends in Developmental, Behavioral and Somatic Factors by Diagnostic Sub-group in Pervasive Developmental Disorders: A Follow-up Analysis, pp. 10, 14  Autism Insights 2009:1 3-17  at https://www.academia.edu/17814363/Trends_in_Developmental_Behavioral_and_somatic_Factors_by_Diagnostic_sub-group_in_pervasive_Developmental_Disorders_A_Follow-up_Analysis.  This study found that 65% of autism cases had been breastfed for at least four weeks; the authors looked at a comparison figure of 54%, but that figure was unrealistically high, since it came from a study (Pontin et al.) of breastfeeding by mothers largely from “more affluent families”, who breastfeed at unusually high rates in the U.K.   For breastfeeding prevalence data that would apply to the general U.K. population, the authors of the Pontin study referred the reader to Infant Feeding 1995 (Foster et al.); examination of the data in that book reveals that a figure in the upper 20%’s would apply for the equivalent period (just after four weeks). That is also as was found in the U.K. Infant Feeding Survey - UK, 2010 Publication date: November 20, 2012, Chapter 2, at http://www.hscic.gov.uk/catalogue/PUB08694/ifs-uk-2010-chap2-inc-prev-dur.pdf


 (6) Type 2 Diabetes in Children and Young Adults:  A “New Epidemic”  Francine Ratner Kaufman, MD  Clinical Diabetes • Volume 20, Number 4, 2002  at http://clinical.diabetesjournals.org/content/20/4/217.full.pdf+html

     Food allergy: Riding the second wave of the allergy epidemic  Susan Prescott  http://onlinelibrary.wiley.com/doi/10.1111/j.1399-3038.2011.01145.x/pdf

     Research needs in allergy: an EAACI position paper, in collaboration with EFA  Papadopoulos et al. Clinical and Translational Allergy 2012, 2:21  http://www.ctajournal.com/content/2/1/21

     CDC web page,  http://www.cdc.gov/CDCTV/ObesityEpidemic/Transcripts/ObesityEpidemic.pdf    Obesity increase data also at CDC's Health United States, 2008, Data Table for Figure 7.

     Examination of historical data with regard to all of the diseases and conditions said by the Surgeon General to be "excess risks" of formula feeding, with sources, can be found at www.breastfeedingprosandcons.info .


(7)   Agency for Healthcare Research and Quality, U.S. DHHS, Systems to Rate the Strength of Scientific Evidence, Evidence Report/Technology Assessment: Number 47  http://archive.ahrq.gov/clinic/epcsums/strengthsum.pdf


(8)  Much explanatory text and many sources are to be found in Appendix 1 at www.breastfeedingprosandcons.info/Appendix.htm




 * About Pollution Action

Please see www.pollutionaction.org