Navigation bar--use text links at bottom of page.

(Looking at the Science on Raw vs. Cooked Foods--continued, Part 1E)

Heterocyclic amines and cancer
risk: lab vs. real-world conditions

Doses in lab animals many times higher than normally encountered in food. In evaluating the potential real-world effects here, note that the preceding discussion and tables concern lab results that utilized extremely high doses to produce easily measurable effects. An important proviso to be kept in mind is that cancers in laboratory rats and mice are induced by doses (per unit of body weight) several orders of magnitude higher than what are usually ingested in normal meals by humans. (Typically, the amounts of HCAs are over a million times higher, based on calculations from data furnished in Augustsson [1999].) In fact, by itself, the ingested daily amount of heterocyclic amines is very probably too small to explain the development of human cancers, and the same is true for numerous other carcinogens. Thus, the simultaneous presence of heterocyclic amines with other genotoxic [i.e., inducing DNA damage] carcinogens and with tumor-promoting agents or tumor-promoting conditions makes it very difficult to make a numerical calculation for risk estimation.

Epidemiological studies done so far remain inconclusive. Some attempts have been made to determine whether heterocyclic amines have a consistent carcinogenic effect in humans by conducting epidemiological studies (such as trying to correlate the rate of colon cancer with the preference for well-done meat), but are not conclusive (yet?): e.g., Sinha et al. [1997], Probst-Hensch et al. [1997], Ward et al. [1997], and Augustsson et al. [1999]. Let's take a look at the latter study by Augustsson et al.

Authors of the study randomly selected 553 controls (that is, people who did not have cancer), 249 cases of rectal cancer, 273 cases of bladder cancer, and 138 cases of kidney cancer. Cancer risk calculations were based on previous laboratory analyses of HCA concentrations observed under given cooking conditions, and on an extensive food-frequency questionnaire including 188 food items, asking for the type of meat and fish ingested, frequency of consumption, portion size, cooking methods, and degree of surface browning. (Participants were asked to compare the color of the foods eaten with a few color photographs showing dishes fried at four different temperatures.)

The population was then divided into five quintiles according to total HCA intake, where quintile 1 consisted of the 20% who had the lowest intake, and quintile 5 consisted of the 20% who had the highest, and so on. The result obtained is summarized in the table below:


Expressed as risk relative to quintile 1,
where quintile 1 = the lowest (reference level) HCA intake
in the study population, and quintile 5 = the highest.


RELATIVE RISK (95% confidence interval)

Quintile 2

Quintile 3

Quintile 4

Quintile 5


1.1 (0.7-1.7)

0.8 (0.5-1.3)

0.7 (0.5-1.1)

0.6 (0.4-1.0)


1.3 (0.8-2.0)

0.8 (0.5-1.4)

0.7 (0.4-1.1)

0.7 (0.4-1.1)


1.5 (0.9-2.6)

1.5 (0.8-2.6)

1.4 (0.8-2.4)

1.2 (0.7-2.1)


1.3 (0.7-2.3)

1.2 (0.6-2.2)

1.1 (0.7-2.2)

1.0 (0.5-1.9)

These numbers were adjusted for age, sex, daily energy intake (and smoking status for bladder and kidney cancer). Quintile 1 (low HCA intake) here is the "reference category," in the sense that the risk for cancer in this quintile was normalized at 1. Thus, the entry "0.6" for quintile 5 concerning colon cancer means that people who cook their meat and fish at the highest temperatures have a decreased risk (by 40%) compared to those who cook their meat and fish at the lowest temperatures; on the other hand, HCAs increase risk for bladder cancer.

Overall impact, if any, appears to be small. Numbers inside parentheses represent the 95% Confidence Interval (CI). For instance, in quintile 5 of the general population, there is a 95% probability that the factor by which the risk for bladder cancer due to elevated HCA intake is increased (compared to those who have a low intake) lies in the interval 0.7-2.1. We see that the results lack precision; nevertheless, it is quite apparent that HCA intake cannot have more than a very small impact on the incidence of cancer in the colon, rectum, bladder, or kidney.

The article also includes a table (not reproduced here) where the individual effects of IQ, MeIQ, MeIQx, DiMeIQx, and PhIP were recorded. (These HCAs are suspected to be the main possible carcinogens in humans.) Again, each of the HCAs was found to slightly decrease the risk for some cancers, and slightly increase the risk for others.

The authors note that:

Heterocyclic amines have been shown to be carcinogenic in animals. This carcinogenic effect is induced by high doses, such as 10-400 mg/kg of body weight. The lack of carcinogenic effect of heterocyclic amines in our study may be due to the much lower intake in the study population (median 1 ng/kg of body weight).

In other words, doses required to induce cancer in laboratory animals are typically 10 million times higher (or even more) than those obtained by ordinary cooking.

Extreme cooking times/temperatures may represent the main risk. While the median HCA intake was less than 100 ng, it was observed that seven participants had intakes higher than 1,900 ng--and all of them were cancer cases. A sample of only seven people is certainly too small to provide statistically significant results, but this finding suggests that HCAs may induce cancer at the very highest doses that could conceivably be encountered in food (i.e., in people who consume well-done, fried/roasted meats very frequently).

As a final note, the authors also observe that: "[T]he effect of heterocyclic amines may be strongly modified by genetic factors, such as acetylation status (Roberts-Thompson et al., 1996), that vary between different populations." In other words, since most of us do not know what our acetylation status is, it is preferable to be moderate with regard to intake of fried or roasted meat and fish (just in case one may be genetically predisposed to HCA sensitivity). At the same time, in light of the paper by Augustsson et al. [1999], we also note that one needn't fear cooked meat like violent poisons, as many raw purists seem to do.

Further complicating the picture is that many other substances that can contribute to cancer exist, some of which are found in raw plant food sources themselves [Ames 1990]. Thus, an environment totally free of carcinogens is not possible.

Finally, it should be pointed out that some mechanisms of metabolism and detoxification of heterocyclic amines are known to exist and have been described [Turesky 1990].

Confounding factors regarding composition of domesticated meats vs. wild game are not accounted for in current HCA studies. Aside from cooking, one confounding effect that afflicts the interpretation of current epidemiological studies with regard to meat in the diet is the widely divergent fat profile of domesticated/feedlot meat in Westernized diets compared to the wild game in hunter-gatherer diets on which humans evolved. Among other differences, domesticated meat has roughly 5 times the fat, and the fat present is 5-6 times as saturated [Eaton 1996, p. 1735]. In addition, the overall balance of foodstuffs and nutrients in the omnivorous diets of hunter-gatherers also differs greatly from that in omnivorous modern Western diets.

Failing to account for such differences--especially given that high levels of saturated fats are under investigation as a potential trigger for cancer--represents a widespread but unappreciated fallacy/oversight often involved in studies of the health effects of meat and/or omnivore diets in general, whether or not cooking is involved. For more in-depth information on these points, see the articles Hunter-Gatherers: Examples of Healthy Omnivores; The "Omnivorism = Western Diet" Fallacy; and Logical Fallacies and the Misinterpretation of Research elsewhere on the site.

Mutagenicity/carcinogenicity of Maillard reaction products

Maillard molecules are not carcinogenic, and instead appear to have antioxidant properties. In contrast with heterocyclic amines, Maillard molecules do not have carcinogenic properties. On the contrary, Maillard reaction products seem to have an antioxidative effect in vivo [Chuyen et al. 1990]. From Aeschbacher [1990] (see also Powrie et al. 1986]), Maillard reaction products that occur predominantly in heat-processed carbohydrate-rich foods (mainly bakery products, coffee, caramel) are predominantly of the melanoidin type and some are compounds involved in aroma formation. Studies in vivo show that these compounds are not involved in cancer induction (unlike heterocyclic amines). They might even protect from cancer due to their antioxidant properties.

One of the reasons why heated carbohydrate-rich food products don't induce cancer in vivo is that some liver enzymes, called "catalases," are able to abolish the (weak) mutagenic effect of these foods. In contrast, the mutagenic activity of heterocyclic amines (found in heated protein-rich foods) is activated by these catalases.

Other anticarcinogenic effects of Maillard products in vivo include: inhibition of the formation of (carcinogenic) nitrosamines; scavenging of active oxygen (it is known that oxidative damage to cells, in particular--but not only--to DNA, contributes to carcinogenesis); and changes in chemical structures of carcinogens.

From Shibamoto [1989], no significant mutagenic activity from MRPs is obtained below 130°C (266°F). Boiling alone doesn't produce mutagenic activity.

On the other hand, one may argue that heating destroys some natural antioxidants such as vitamin C, but it appears that the two effects compensate each other, so that, in an experiment of Nicoli et al. [1997] on tomatoes, the overall antioxidant properties of the food products were maintained or even enhanced by the development of Maillard reaction products.

Other factors influencing carcinogenesis


Most food allergens are natural proteins whose allergenic properties are either destroyed or unaffected by heating [Yunginger JW, 1991]. Examples: Persons allergic to fresh tuna or fresh salmon may be able to tolerate ingestion of canned tuna or canned salmon (the prolonged cooking undergone by these products may denature the allergenic components). The major egg (white) allergen, ovomucoid, is heat-stable, and people sensitive to this component may react to foods containing cooked eggs as well to raw eggs. Many peanut allergens are both present in raw and heated peanuts, but high-temperature processed peanut oil is not allergenic, contrary to cold-pressed peanut oil.

Lactose/protein combination an exception. However, beta-lactoglobulin allergenic activity is enhanced by the Maillard reaction between lactose and protein [Matsuda T et al. 1990]; elevated skin reactivity of heated cow's milk proteins had earlier been demonstrated by Bleumink E [1966, 1968, 1970].

A case of allergy to heated and/or aged pecan nuts has been reported by Malanin et al. [1995]. However, such reports (cases of people having more acute allergic reactions to a given food when it has been heated but not when it is raw) are quite rare in the scientific literature.

In summary, except for dairy products, cooked foods generally are less than or equally allergenic when compared to raw foods.

Risks for the vascular system and the kidney?

Let's investigate here the possible effects of Maillard molecules on the vascular system and the kidney.

Rats fed 5% to 10% Maillard reaction products (MRPs) show various pathologies (including of the kidney) [O'Brien et al. 1988]. Rats fed for 8 to 10 weeks with a mixture of casein and glucose which had been previously heated for 4 days at 65°C (149°F) showed abnormalities of the kidney [Erbersdobler et al. 1984].

There is evidence that MRP clearance by the kidney of diabetics is impaired, which causes elevated serum AGEs (advanced glycation endproducts) [Koschinsky et al. 1997]. This may constitute an added chronic risk for renal-vascular injury in diabetes: previously it had been shown [Vlassara et al. 1995] by injecting AGEs in nondiabetic rabbits that elevated serum AGEs contribute to atherogenesis (clogging of the arteries).

Dietary AGEs are much more problematic for diabetics than for nondiabetics because (a) there is an accelerated formation of AGEs in the body, and (b) AGE clearance is impaired in diabetics. However, note that in the above-mentioned experiments [O'Brien et al. 1988, Erbersdobler et al. 1984], rats were fed a considerable amount of MRPs: no one cooks their food for 4 days, and formation of MRPs is faster in artificial mixtures of casein + glucose than in whole foods.

No evidence of risks in nondiabetics. More to the point, in the concentrations at which they appear in traditionally cooked foods, we find no evidence that MRPs can cause any risks in nondiabetic people. (And it is doubtful diabetes would exist but in rare cases if people were to consume an evolutionarily congruent diet with limited carbohydrate levels so as not to promote or exacerbate insulin resistance in the first place.)


(Polycyclic Aromatic Hydrocarbons and Sensitivity to Cooking Method)

Return to beginning of article



GO TO PART 1 - Is Cooked Food "Toxic"?

GO TO PART 2 - Does Cooked Food Contain Less Nutrition?

GO TO PART 3 - Discussion: 100% Raw vs. Predominantly Raw

Back to Research-Based Appraisals of Alternative Diet Lore

   Beyond Veg home   |   Feedback   |   Links