If someone told you certain things would increase the risk of developing breast cancer, would you avoid it?  Some of you would say yes regardless of what it was, some would say it depends, and some wouldn’t care.  As adults and humans, we’re bound to play the odds.  Is the convenience or the other benefits worth the increase in risk?  If it isn’t, we’ll avoid it, but sometimes we accept that is.  Smokers know that they stand a higher chance of developing lung cancer if they smoke, but to some of them, the cost of quitting is greater than the reduction in risk they’ll have.  Why?  Because despite all the ads and public health warnings, cancer isn’t as deadly and pervasive as health officials make it out to be.  Now, don’t get me wrong, the risk is real and it’s there, but many of us go our lives without thinking of it.  And those we know who get it may not do so until later in life and many will get better.  So we weigh the odds.  This is part of the reason why, despite a plethora of studies outlining the fact that not breastfeeding serves to increase the risk of breast, cervical, uterine, and endometrial cancer[1][2][3][4], not all women choose to breastfeed.  Now some don’t do so because they are unaware of the benefits those two boobs can provide for them, but some know of the risk and choose not to breastfeed anyway because in their view, the risk is so minimal that it’s not worth their perceived hassle of breastfeeding and that’s okay – that’s their educated choice to make.  (Some don’t have a choice and have to accept the risks and that can suck, but there are times in all our lives we’re forced to accept risks we don’t want to take.)

The question about risks also has to be asked about childhood cancer.  It has long been listed as one of the things for which there is an increased risk associated with formula use, but many people suggest the evidence is equivocal.  If parents are going to make an informed decision about the risks, then we need to know what the research is saying and why it can seem so darn confusing.

The Evidence 

(Notably here I will revert to using the language of the research which uses a “reduced risk” model for breastfeeding, despite it being the other way around as breastfeeding is the norm.  Part of this is so people who read the original sources are not confused, but also because when I provide odds-ratios, I can just take them from the articles and not have to work out what the flip side would be.)

First, let’s be clear that not all findings support the notion that breastfeeding reduces the risk of childhood cancer, but many methodological limitations make conclusions difficult to come by.  For example, data from the Oxford Survey of Childhood Cancers (OSCC) in the UK looked at the data from thousands of children to try and establish a link between breastfeeding and cancer.  The authors found no relationship at all by any length of breastfeeding (examined at < 1 month, 1-6 months, and >= 7 months)[5].

What were the problems?  A couple things about this study are worth mentioning: a) they removed cases in which symptoms of cancer occurred in the first nine months or the child died in the first year, which may be where the effects of breastfeeding are strongest (though I certainly don’t know this for a fact, but it is reasonable); and b) they did not examine exclusive breastfeeding, but rather any breastfeeding (this is important as the effects of breastfeeding are known to be strongest when breastfeeding is exclusive and it is also difficult to ascertain how little or how much women were breastfeeding when any breastfeeding is the variable of interest).

In China, researchers found a slight reduction in risk for lymphoma for those infants ever breastfed which was slightly greater for longer-term breastfeeding.  Notably, the effects pertained primarily to Hodgkin’s disease and there was no relationship to acute leukemia[6].  Interestingly, breastfeeding rates in China in the mid-1990s were relatively high compared to the US (though it did seem to vary by region) (any: 94%, 82%, and 72% at 1, 4, and 6 months)[7].  Returning to the UK, research examined various factors predicting incidents of childhood cancer in a national cohort.  They examined breastfeeding, but only breastfeeding in the first week of life.  Despite the short time-frame, 46% of controls received breastmilk in the first week while only 28% of children who developed cancer received any breastmilk[8].  This early influence of breastmilk highlights the potential flaw in the OSCC data which excluded children who were symptomatic in the first year of life; if exposure to breastmilk in the first week of life is related to a decreased risk of developing cancer later on, it may be particularly salient in the first year.

The research from China and the UK is enhanced by more conclusive evidence from the US and India.  A case-controlled study in the US examined the role of breastfeeding by dividing groups into > 6 months breastfeeding, 6 months breastfeeding, and artificial feeding.  Compared to the > 6 months group, there was an increased risk of all types of cancers for the 6 months breastfeeding and artificial feeding groups.  Despite the results being significant across all types of cancer, the results seemed to be driven primarily by a large significant protection for lymphoma[9].  An Indian case study examining childhood cancer rates and the effects of breastfeeding also found a significant effect of average duration of breastfeeding and an even larger effect for exclusive breastfeeding on distinguishing between the two groups.  Again, while the results translated across the entire group, they were driven by the highly significant relationship between breastfeeding and lymphoma, as an examination of breastfeeding and all other types of cancer was non-significant[10].

This starts to look like some relatively strong evidence for a role of not breastfeeding in lymphomas; however, these are not so common.  The data for the most common type of childhood cancer – leukemia – is more controversial.  While not looking at breastfeeding directly, one UK study examined how infections in the first year of life predicted later development of ALL.  The authors found that children who later developed ALL had a significantly greater number of clinically-diagnosed infections in the first year of life (i.e., we’re not talking about the common cold)[11].   Given the relationship between breastfeeding and a reduction in infections[12], it seems plausible that breastfeeding may also reduce the incidence of ALL, despite previous non-significant results, but this is simply guesswork.  Although the previous studies failed to find evidence of protection for ALL and AML childhood cancers, Xiao Shu and colleagues examined data from the Children’s Cancer Group (which collects data from the United States, Canada, and Australia) for possible breastfeeding benefits.  They found no significant association between ever breastfeeding and AML status and only a very slight protective benefit for ALL.  However, when duration of breastfeeding was examined, children who had breastfed for longer than six months had a 43% reduction in risk for AML and a 28% reduction in risk for ALL[13].  Already we start to see the issue of breastfeeding duration being one that requires greater examination.

Why would this study find a relationship that none of the aforementioned studies did?  It seems that even within ALL and AML, there are subtypes that are differentially affected by exposure to breastmilk.  Shu and colleagues also performed an analysis by subtype of ALL and AML, something previous research had failed to do.  What they found was that the reduced risk in AML cases only existed for certain subtypes, namely myeloblastic with no maturation, myeloblastic with minimal maturation, and myeloblastic with maturation.  Similarly, extended breastfeeding only protected against Early Pre-B-Cell ALL and having ever breastfed only provided protection against Pre-B-Cell ALL.  This means that while breastfeeding may not provide protection for all types of leukemia, it does offer protection against some of the subtypes of acute leukemia; thus some of the previous research looking at all types of ALL and AML would be biased towards non-significance.

In line with this research, and perhaps more conclusively, researchers at UC Berkeley performed a meta-analysis of 14 studies examining the relationship between breastfeeding and two types of childhood cancer: ALL and AML.  [For those who are unaware, a meta-analysis is considered the gold standard of conclusiveness for research as it combines the data from many studies and analyzes them together.  This means that any individual flaw or bias in any one article gets weighed out by the many other studies in the meta-analysis.  Of course, a meta-analysis is only as good as the collective data included and so one must be sure that stringent criteria were met to decide what to include. ]  They divided the data by short- (<= 6 months) and long-term (> 6 months) breastfeeding, which is important because of the findings to do with breastfeeding duration mentioned above, and found significant negative relationships between breastfeeding and both ALL and AML.  While results were significant for both short- and long-term breastfeeding, the results were much stronger for long-term breastfeeding[14].

Given this data, we must ask ourselves why so few parents are made aware of this information.  One of the problems may be that, for unknown reasons, researchers seem hesitant to believe or support their findings.  In a different meta-analysis the authors reported having found no effect of breastfeeding to most types of childhood cancers, except that wasn’t what their data found.  If you look at their data, they found significant protective effects for acute lymphoblastic leukemia, Hodgkin’s disease, and neuroblastoma (with a 9%, 24% , and 41% reduction in risk, respectively).  Furthermore, their analysis for breastfeeding was only a comparison of “ever breastfed” versus “never”[15], which as we know tends to lower the likelihood of finding anything significant.  If the researchers who find these results don’t support them, how are parents supposed to know about and utilize this information?  (I personally wonder if these scientists are towing a political line that doesn’t want to acknowledge these benefits because then they might actually have to support policies that allow for women to breastfeed instead of forcing them back to work too early or providing proper support in the form of lactation consultants and other specialists.)

A Cross-Cultural Story

While the rates of childhood cancer are 14.3 per 100,000 worldwide, according to the American Cancer Society, rates and types vary by location.  It is the variety in types that interested me most and that I wanted to share.  Based on the above evidence, it seems fair to say that breastfeeding holds its power to reduce risk on leukemias, lymphomas, and neuroblastomas.  Data from the American Cancer Society’s Global Cancer Facts and Figures, 2nd Edition tells us that leukemia is the most common form of cancer in most parts of the world, and especially in Western nations, but is a rarity in Africa.  In fact, if we examine the predominance of types of cancers in Europe versus Africa, an interesting picture emerges.  While leukemia, lymphoma, and neuroblastomas make up 49% of childhood cancer cases in North America, they make up 21% of cases in Africa[16] (not including Burkitt lymphomas as they are linked to the Epstein-Barr virus and thus occur nearly exclusively in Africa and seem unrelated to other types of lymphomas[17]).

A comparison of the rates of cancer is impossible as the diagnosis and treatment of cancer in Africa is qualitatively different than that in North America, it is interesting to note that on a continent where breastfeeding and extended breastfeeding are common—only 3.9% of babies receive no breastmilk in the first six months, 8.2% receive no breastmilk from 6 months to a year, and only 30.1% of infants do not breastfeed in the second year of life[18]—the rates of childhood cancers that are affected by breastfeeding are much lower than in other areas of the world.  Of course, none of this is causal and more research would need to be done to make firm conclusions, but combined with the evidence we see on childhood cancer and breastfeeding, it’s hard not to think something’s going on.

Implications

Is the data absolutely clear cut?  No.  But given what we have, the evidence falls more in line with an increased risk of certain types of childhood cancers (not all) with not breastfeeding, and particularly, not breastfeeding six months.  More research is needed to clarify these relationships, but as we delve deeper and into more specific findings, hopefully that becomes possible.  Of course, to note again, the risk of these cancers is still quite small and so although there is an increased risk for not breastfeeding, the absolute risk remains quite small.  This is where risk-benefit analysis will become very personal.  For some people, this matters greatly.  For others, the influence isn’t nearly as great.  That’s a personal decision.

I would be remiss to not discuss the issue of women who cannot breastfeed.  Taking on risks you don’t desire to take on can lead to immense grief and pain.  It’s not fair, but it’s life.  What we can do is work to make donor breastmilk more available and make sure the emotional support is there for these women.  The goal in talking about research is not to try and make people feel bad but to share information and find ways to create better scenarios for those who have to unwillingly accept the risks.

Many mums are unable to physically feed their baby for various reasons, and as such there is always ample need for a mother’s breastmilk to help those mums who can’t breastfeed their child on their own. 

If you live in Australia, it seems that there are 4 milk banks: Mothers’ Milk Bank in northern New South Wales (www.mothersmilkbank.com.au), the PREM Bank in Perth, Western Australia (www.kemh.health.wa.gov.au/services/PREM_Bank/index.htm), the Mercy Health Breastmilk Bank in Victoria (www.mercyhealthbreastmilkbank.com.au) and a new one just opened in the last few weeks at the Royal Brisbane and Women’s Hospital in Brisbane, Queensland (http://www.rbwhfoundation.com.au).

If you live in New Zealand, there is currently no milk bank, but there is a non-profit group, Mothers Milk NZ, working towards the development of a national milk bank. 

If you live in Canada, there are now two milk banks.  The first is run by Women’s Hospital in Vancouver, BC.  Although located on the west coast, women from all over the country can both utilize and donate to it.  For more information, please go tohttp://www.bcwomens.ca/Services/PregnancyBirthNewborns/HospitalCare/Breastfeeding.htm#Milkbankfor a list of FAQs and contact information.  The second is in Toronto, ON, and is a partnership amongst three hospitals in Toronto.  They accept donations from all over and will cover the cost of shipping milk to the bank.  For more information, please go to http://www.milkbankontario.ca/ for all the information you need!

If you live in the United States, several states have milk banks which you can contact to see if they have milk you can purchase.  I don’t know how each of them sets their standards or if they mail milk out-of-state (if you live in a state without a milk bank), but a list of all the milk banks in the US, with contact information, can be found here: http://www.hmbana.org/milk-bank-locations.

If you live in the United Kingdom, you have 17 milk banks (go you!) and the primary site with information and how to donate is found at http://www.ukamb.org/.  

And of course, no matter where you live there are the always-wonderful Human Milk 4 Human Babies andEats on Feets networks which are (nearly) global.  Here you will be able to network with other mums who may live in your area and have breastmilk for you.  This is not a regulated endeavor and is running solely on the goodwill of mothers everywhere. 

Good luck and happy feeding!



[1] Rosenblatt KA, Thomas DB, and the WHO collaborative study of neoplasia and steroid contraceptives. Prolonged Lactation and endometrial cancer. International Journal of Epidemiology (1995); 24: 499-503.

[2] Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and breastfeeding: collaborative reanalysis of individual data from 47 epidemiological studies in 30 countries, including 50302 women with breast cancer and 96973 women without the disease. Lancet (2002); 360: 187-95.

[3] Rosenblatt KA, Thomas DB, and the WHO collaborative study of neoplasia and steroid contraceptives. Lactation and the risk of epithelial ovarian cancer. International Journal of Epidemiology (1993); 22: 192-7.

[4] Brock KE, Berry G, Brinton LA, Kerr C, MacLennan R, Mock PA, & Shearman RP. Sexual, reproductive and contraceptive risk factors for carcinoma-in-situ of the uterine cervix in Sydney. The Medical Journal of Australia (1989); 150: 125-30.

[5] Lancashire RJ & Sorahan T.  Breastfeeding and childhood cancer risks: OSCC data.  British Journal of Cancer (2003); 88: 1035-1037.

[6] Shu X, Clemens J, Zheng W, Ying DM, Ji BT, & Jin F.  Infant breastfeeding and the risk of childhood lymphoma and leukaemia.  International Journal of Epidemiology (1995); 24: 27-32.

[7] Xu F, Lui X, Binns CW, Xiao C, Wu J, & Lee AH.  A decade of change in breastfeeding in China’s far north-west.  International Breastfeeding Journal (2006); 1: 22.

[8] Golding J, Paterson M, & Kinlen J.  Factors associated with childhood cancer in a national cohort study.  British Journal of Cancer (1990); 62: 304-308.

[9] Davis MK, Savitz DA, & Graubard AI.  Infant feeding and childhood cancer.  The Lancet (1988); 332: 365-368.

[10] Mathur GP, Gupta N, Mathur S, Gupta V, Pradhan S, et al.  Breastfeeding and childhood cancer.  Indian Pediatrics (1993); 30: 651-657.

[11] Roman E, Simpson J, Ansell P, Kinsey S, Mitchell CD, McKinney PA, et al.  Childhood acute lymphoblastic leukemia and infections in the first year of life: a report from the United Kingdom childhood cancer study.  American Journal of Epidemiology (2007); 165: 496-504.

[12] Duijts L, Jaddoe VWV, Hofman A, & Moll HA.  Prolonged and exclusive breastfeeding reduces the risk of infectious diseases in infancy.  Pediatrics 2010; 126: e18-e25.

[13] Shu XO, Linet MS, Steinbuch M, Wen WQ, Buckley JD, Neglia JP, et al. Breast-feeding and risk of childhood acute leukemia.  Journal of the National Cancer Institute (1999); 91: 1765-1772.

[14] Kwan ML, Buffler PA, Abrams B, & Kiley VA.  Breastfeeding and the risk of childhood leukemia: a meta-analysis.  Public Health Reports (2004); 119: 521-535.

[15] Martin RM, Gunnell D, Owen CG, & Smith GD.  Breast-feeding and childhood cancer: a systematic review with metaanalysis.  International Journal of Cancer (2005); 117: 1020-1031.

[16] American Cancer Society.  Global Cancer Facts & Figures, 2nd Edition. Atlanta: American Cancer Society; 2011

[18] Lauer JA, Betran AP, Victora CG, de Onis M, & Barros AJD.  Breastfeeding patterns and exposure to suboptimal breastfeeding among children in developing countries: review and analysis of nationally representative surveys.  BMC Medicine (2004); 2: 26.