Skip to main content

Overcooked food and cancer

Monday 3 December 2007

“A common chemical caused by frying, roasting or grilling can double the risk of cancer in women,” The Daily Telegraph reported today. The story warns that acrylamide, a chemical contained in cooked foods including bread, coffee and breakfast cereals, and meat and potatoes that have been fried, baked, roasted, grilled or barbecued, has been directly linked to the incidence of ovarian and womb cancer.

This news story is based on a recent study conducted in Holland, which found that women who consumed foods that contained a lot of acrylamide (more than 40 micrograms/day) were at greater risk of endometrial and ovarian cancer. The study adds to a growing body of knowledge about this chemical. It is the first study of its kind that has found an actual (rather than hypothetical) association between consumption of acrylamide and cancer in humans. However, the study has some weaknesses because of the way it is designed and, as the authors admit, further studies will be needed before “far-reaching conclusions can be drawn”.

This new study supports the theory that acrylamide causes cancer in humans, but as yet, scientists and regulators don’t know enough to make recommendations about how much can be consumed safely.

Acrylamide was first found in food by Swedish scientists in 2002. It is naturally produced by cooking food at a high temperature. Foods that are carbohydrate rich, such as chips and crisps, contain the highest levels. Acrylamide is a proven carcinogen in laboratory animals, and its risk to humans has long been suspected. A government advisory committee says, “Exposure to DNA-damaging cancer-causing chemicals, such as acrylamide, should be as low as reasonably practicable.”

An EU commissioned project, Heatox, reported in November 2007 that there was increasing evidence that acrylamide might be a cancer risk factor. The study reported that while there is no practical way that its consumption can be eliminated, exposure to it can be reduced. Importantly, the study estimates that the amount of acrylamide obtained from home-cooked food is relatively small when compared with “industrially or restaurant-prepared foods.” Its presence in home-cooked food was primarily in potato products and toasted and homemade bread.

General advice from the project includes avoiding eating over-cooked baked, fried or toasted carbohydrate-rich food. The report also advises: “By following the general dietary recommendations (i.e. a balanced diet without excessive fat or calorie intake) a further reduction of the acrylamide intake can be achieved.”

Where did the story come from?

Dr Janneke Hogervorst and colleagues from Maastricht University, the Food and Consumer Product Safety Authority and the Department of Food and Chemical Risk Analysis carried out this study. The Dutch Food and Consumer Product Safety Authority funded the research.

The study was published in the (peer-reviewed) medical journal Cancer Epidemiology Biomarkers and Prevention .

What kind of scientific study was this?

This was a nested case-control study of women aged 55 to 69. The women were all enrolled in a large cohort study which began in 1986: the Netherlands Cohort Study (NCS) on diet and cancer. The diets of those who got endometrial, ovarian or breast cancer during 11 years of follow-up were compared with a control group of women who did not get cancer drawn randomly from the same NCS population.

As part of the NCS, the women had completed a baseline questionnaire in 1986 on their diet and other risk factors. The researchers were particularly interested in answers to questions about intake of foods that are known to contain high concentrations of acrylamide. Such foods included crisps, bread, coffee, cookies, pastry, peanut butter, breakfast cereals, nuts, pastry, etc. The quantity of acrylamide in each of the foodstuffs had been analysed by the Dutch Food and Consumer Product Safety Authority between 2002 and 2005 when people began to be concerned about this substance in foods.

The researchers compared the women who developed cancer with those who did not in terms of the differences in their diets (specifically intake of acrylamide). They took into account (i.e. adjusted their analyses for) other well-known factors that may have had some effect on the risk of cancer. Such factors included age at menarche, use of oral contraceptives, smoking status, physical activity, energy intake, and alcohol intake. As cigarettes contain so much acrylamide, they also performed some analyses in non-smoking women, to have a better idea of how acrylamide intake through diet affects cancer risk while excluding the effect of smoking.

What were the results of the study?

The researchers analysed the results separately for endometrial, ovarian or breast cancer. They found no statistically significant relationship between intake of acrylamide and endometrial cancer when other risk factors for endometrial cancer were taken into account.

In non-smoking women however, those who were taking in the highest amount of acrylamide (about 40 micrograms/day) in their diet were about twice as likely to have endometrial cancer than those consuming the lowest amount of acrylamide (about eight micrograms/day).

Similarly, the women consuming the most acrylamide were at higher risk of ovarian cancer whether they smoked or not than women who consumed the lowest amount. As with endometrial cancer, non-smokers seeming to be at greater risk.

There was no link between acrylamide and breast cancer in any group.

The only significant results were seen when women consuming more than 40 micrograms/day were compared with those consuming less than 10 micrograms/day. There was no apparent increase in risk when women were consuming approximately 25 micrograms/day or less.

What interpretations did the researchers draw from these results?

The researchers conclude that their study has shown that consumption of high levels of acrylamide increased the risk of postmenopausal endometrial and ovarian cancer, particularly among women who had never smoked. They are unclear why there was this difference between the total group and the never-smoked group.

What does the NHS Knowledge Service make of this study?

This is a case-control study (nested within a larger cohort) and as such has limitations. The researchers raise some of these:

  • The dietary questionnaire was given at the beginning of the larger study. As the women were followed up over 11 years, their diets are unlikely to have stayed the same over that time. The study could not capture this potential change in the women’s diets and therefore cannot estimate what effect this would have on cancer risk.
  • The rates of cancer were calculated using the total number of women in the NCS (more than 62,000) as the denominator, this is done on the assumption that the 2,438 women chosen as controls are representative of the larger group.
  • Levels of acrylamide are likely to vary greatly between foods depending on how they were cooked. The researchers used quantifications of acrylamide based on an analysis of various foodstuffs that were then assumed to represent the quantity in the food that the women consumed. This is unlikely to have been 100% accurate for all women.
  • The concentration of acrylamide in different foodstuffs was measured between 2002 and 2005, some time after the NCS study began. It is unlikely that foods the women were consuming in 1986 were the same as those tested in 2002. The quantities of acrylamide may have changed one way or the other during that time too.
  • A large proportion (about half) of the acrylamide eaten by the women with the highest intake of the chemical came from a particular Dutch spiced honey cake, which may not be eaten in other countries. The authors refer to unpublished data to support the view that it is not just the spiced cake that is responsible for the observed association.

Acrylamide is present in large amounts in starchy food that has been cooked at high temperatures. There is currently no guidance what is considered to be a safe amount to eat. The study quotes the World Health Association (WHO) that reports that the daily intake of acrylamide for developed countries is 0.3 to 0.8micrograms per kg of body weight. The women in this study who were in the top fifth for consumption ate about 40micrograms per day (equivalent to a daily intake of 0.5micrograms per kg of body weight).  Each day, 20micrograms came  from spiced honey cake, 10micrograms from coffee and the rest from other foods such as cookies, French fries and crisps. Some examples of the average acrylamide content of foods were given such as: 1,249micrograms per Kg for potato crisps; 1,018 for dutch spiced cake; 351 for French fries; 121 for corn flakes.

Research of this type can warn that particular chemicals need further investigation, but there are so many thousands of chemicals in the diet, that it is difficult to untangle direct associations between individual chemicals in food, and the observation of an increase in a condition such as cancer. This is especially true  when the intake of acrylamide may be a more general marker of an unhealthy diet.

General advice to consumers is to follow the usual recommendations to maintain a balanced diet and to avoid overcooking food.

Sir Muir Gray adds...

Very rarely do single studies provide a clear answer. We need to see the results of this study incorporated into a systematic review with other similar studies before a change in diet is advised. However, any process that adds chemicals may entail some risk. Therefore, the message to eat fruit and vegetables five times a day instead of cooked and baked food is further supported by this finding.

Analysis by Bazian
Edited by NHS Website

Links to the headlines

Study finds acrylamide link to cancer in women

The Daily Telegraph, 3 December 2007

Links to the science

Hogervorst JG, Schouten LJ, Konings EJ, et al.

A Prospective Study of Dietary Acrylamide Intake and the Risk of Endometrial, Ovarian, and Breast Cancer.

Cancer Epidemiol Biomarkers Prev 2007; 16:2304-2313