Intended for healthcare professionals


Diet, transplacental carcinogenesis, and risk to children

BMJ 2015; 351 doi: (Published 28 August 2015) Cite this as: BMJ 2015;351:h4636
  1. Denis L Henshaw, scientific director1,
  2. William A Suk, director, superfund research programme2
  1. 1Children with Cancer UK, London WC1N 3JQ, UK
  2. 2National Institute of Environmental Health Sciences, National Institutes of Health, USA
  1. Correspondence to: D L Henshaw d.l.henshaw{at}

Dietary carcinogens readily cross the placenta, but we don’t know whether this leads to childhood cancer

With an estimated worldwide annual incidence of 175 000 and a rate below 200 cases per million, cancer in children is rare. Although survival rates have steadily improved, especially for leukaemia, the prognosis for other cancers remains poor, with brain tumours now the major cause of death from cancer in childhood. The aetiology of individual childhood cancers remains largely unknown, but interest continues in the possible role of environmental and lifestyle factors, especially given that incidence seems to be increasing.1 The rarity of childhood cancer, however, creates severe challenges for epidemiology in that even national studies often lack statistical power to investigate possible environmental causes.

Childhood leukaemia is the most investigated in terms of potential causes. Ionising radiation is an established cause. Estimates from the Japanese atomic bomb survivors, supported by epidemiological evidence, suggest that 15-20% of incidence could be due to largely unavoidable exposure to natural background radiation.2 Other environmental exposures linked to increased risk include magnetic fields from the electricity supply,3 pollution from motor vehicle exhausts,4 5 and pesticides.6 Infections seem to be associated with both increased and decreased risk depending on the mode and time of exposure.7

Perhaps half of childhood cancers are initiated in utero. Acute lymphoblastic leukaemia is often associated with a particular chromosomal translocation that creates a fusion gene; this is present at birth and produces a persistent but clinically covert pre-leukaemic clone. However, only a small proportion of children born with this gene rearrangement (perhaps 1 in 100) go on to contract leukaemia.8 Brain tumours may also originate in utero, but whether this occurs in cancers arising in postnatally persistent embryonal cells, such as Wilm’s tumour and neuroblastoma, is unclear.9

Could environmental carcinogens reach the fetus by transplacental transfer from maternal diet and thereby initiate carcinogenesis? Although it is often assumed that the placenta protects the fetus, carcinogenic substances do cross the placental barrier, entering the fetal bloodstream.10 Indeed, there is epidemiological evidence in children for such transplacental carcinogenesis.11

In a linked Analysis article, Jos Kleinjans and colleagues describe their series of studies on transplacental carcinogenesis (the NewGeneris project12), which overcomes the problems of statistical power in conventional epidemiological studies. It marks an important step forward in our understanding of the topic.

They took paired blood samples from mothers and the umbilical cords of 1151 newborns from Denmark, Greece, Norway, Spain, and UK. They measured biochemical and cytogenetic biomarkers of exposure to carcinogens in the maternal diet and present in the fetal bloodstream: acrylamide (associated with deep fried potato products such as French fries and crisps), polycyclic aromatic hydrocarbons (associated with grilled and barbecued meats and smoked foods), oxidative fat metabolites (associated with vegetable oils), nitrosamines (in processed and smoked meats and fish), and dioxins and polychlorinated biphenyls (PCBs) (in meat, fish, and dairy products).

In addition, laboratory measurements were made of transfer across placental samples from 93 mothers obtained at term. Although these measurements are not necessarily indicative of transfer rates throughout gestation, all of the above carcinogens readily crossed the placental barrier. The group used various assays to quantify the concentrations of carcinogens in cord blood. Though large variations in levels were present, fetal exposure was confirmed in most infants.

The group also determined the frequencies of micronuclei in lymphocytes in 467 newborns. Micronuclei represent a cytogenetic biomarker for chromosomal damage. Though the link is not confirmed in children, in adults occupationally exposed to carcinogenic compounds, micronuclei have been prospectively associated with cancer risk. Statistically significant associations were found between features of micronuclei and exposure markers for oxidative fat metabolites and levels of dioxin and PCB related chemically activated gene expression.

Some children may be genetically predisposed or more susceptible to dietary carcinogens. They may lack DNA repair genes or display higher metabolic transformation. While the evidence remains limited, the NewGeneris data suggest a role for genetic predisposition in childhood cancer risk in association with fetal exposure to dietary carcinogens.

Whole genome gene expression levels were assessed in 120 newborns. Sex specific responses were found to acrylamide and dioxin exposure in boys, which may explain the higher leukaemia and overall cancer incidence among boys. Selected genes were examined for all 1151 newborns. Higher gene expression relating to generic processes involved in carcinogenesis was found, which seemed to be associated with markers of exposure to dietary carcinogens.

In other measurements, gestational age was shorter by about half a week in the highest compared with the lowest exposure levels. Some evidence was found for an association between exposure to acrylamide and polycyclic aromatic hydrocarbons and lower birth weight, a possible risk factor for the onset of cardiovascular disease, type 2 diabetes, and osteoporosis in later life. Maternal intake of dioxin and PCBs was also associated with immunotoxic events during early childhood.

The NewGenersis findings provide explicit evidence of fetal exposure and biologically relevant responses to carcinogens present in maternal diets in a large sample. This is an important advance in our understanding of potential early stage, chemically induced carcinogenesis in children. We cannot yet define specific risk factors for carcinogens in maternal diets because we do not know what proportion, if any, of childhood cancer may be so attributed.

Prudent avoidance of consumption of cured meats in pregnancy is one area for consideration given a possible link with childhood brain tumours.10 Food safety agencies may wish to consider these findings and appropriate advice to food manufacturers and expectant mothers. However, many of the dietary carcinogens investigated are hard to avoid given their widespread presence in foods both nationally and globally.


Cite this as: BMJ 2015;351:h4636


  • Analysis, doi:10.1136/bmj.h4501
  • Competing interests: We have read and understood BMJ policy on declaration of interests and have no relevant interests to declare.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract

Log in

Log in through your institution


* For online subscription