So we know products applied to the skin can be, and are, absorbed. What kind of evidence de we have on the carcinogenicity of ingredients in cosmetics and personal-care products, and just how reliable is this evidence?
The data we use in this book has been compiled from numerous sources. These include World Health Organization International Agency for Cancer Research monographs; U.S. National Toxicology Program (NTP) reports, based on rodent tests, of some 600 chemicals; and NTP’s infrequent Annual Reports on Carcinogens (initiated in 1978, with twelve reports published by 2006). These summarize evidence on a range of carcinogens identified through animal tests or human studies.


Additional sources of information on carcinogenic ingredients include the 1980 Science Action Coalition’s Consumer’s Guide to Cosmetics ; my own books, including the 1974 Legislation of Consumer Product Safety and the 1995 The Safe Shopper’s Bible; the 1998 The Breast Cancer Prevention Program; the 2005 Unreasonable Risk of Cosmetics and Personal Care Products; and the press releases and petitions of the non-profit Cancer Prevention Coalition, an association of which I am the current chairman. Finally, evidence on the carcinogenicity of relatively few ingredients and contaminants is admitted, though usually trivialized or dismissed, in the industry’s own Cosmetic Ingredient Review Compendium, published annually by the Personal Care Products Council.


I’ll be talking largely about carcinogens rather than other toxins, especially when it comes to particular ingredients. This is because cancer data is the hardest (which is to say, the most convincing) of all health data collected over the past thirty years. Other categories we have good health data on are allergens and hormone-disruptive chemicals. This information is based on two kinds of evidence: evidence from laboratory studies on animals, and evidence from epidemiological studies (studies on humans). In the majority of cases where chemicals that cause cancer in humans were identified by epidemiological studies, animal studies first predicted their toxicity. So we know that mice and rats, the standard test animals, are effective in enabling us to predict which chemicals will be carcinogenic to humans.


There is an overwhelming consensus, in fact, in the informed independent scientific literature—confirmed by expert bodies, including the International Agency for Research on Cancer (IARC)—that positive results in well-designed animal tests create the strong presumption of human cancer risk.3 It’s a consensus that is also reflected in a wide range of U.S. and international legislative and regulatory precedents.
About 800 industrial chemicals in current use have been shown to be carcinogenic in standard rodent tests. The results of most of these positive tests were initially dismissed or challenged by the industry concerned, and in some cases these results are still being challenged, to protect the profitability of products. The industry claims that the products’ effects on human beings are the best proof of safety. But doing epidemiological studies on individual carcinogenic ingredients in cosmetics and personal-care products is a largely impossible task to perform.


Epidemiological studies depend on the ability to identify population groups exposed to a particular carcinogen or carcinogens, and then to compare their cancer rates with those in unexposed groups. For instance, large-scale epidemiological studies involving millions of people have been conducted on the effects of tobacco use, based on the comparison of lung cancer rates in people who smoke from one to four packs of cigarettes daily for varying periods of time to lung cancer rates in non-smokers. Smaller-scale studies have been conducted on cancer rates in workers exposed to carcinogenic products or processes in a range of industries, based on comparisons to unexposed workers in the same or other industries, or to unexposed groups in the general population.


Doing epidemiological studies on individual carcinogenic ingredients in cosmetics and personal-care products is so difficult because the vast majority of people are exposed, in varying degrees, to the same products and chemicals. In two notable cases, however, such studies have been successfully performed. Epidemiological studies demonstrated excess risks of ovarian cancer in pre-menopausal women who frequently dusted their genital areas with talc or used talc-dusted tampons. Other studies showed the excess risks of a wide range of cancers in women who used black or dark brown permanent or semi-permanent hair dyes for prolonged periods.


Unlike air pollution and water pollution, where it is difficult to get a handle on the sources of our exposure to individual carcinogens, what we put on our bodies in the form of cosmetics and personal-care products is something that we can control. But there is no way of isolating individual ingredients’ effects so as to identify their impacts on health.


Tests for carcinogenicity must be conducted on the products’ individual ingredients and contaminants rather than on the products themselves. The effects of carcinogenic ingredients in a single product are too small to detect reliably, let alone be effectively matched to individual ingredients. So it’s important to note that when you see labels proclaiming “Not Tested on Animals,” these claims relate to the whole product and generally involve irritation or allergy tests in rabbits or guinea pigs rather than toxicity tests on individual ingredients.


The majority of cosmetics and personal-care products manufactured and sold by mainstream companies are veritable witches’ brews of carcinogenic ingredients and contaminants. For instance, many baby soaps, baby shampoos, and bubble baths hide a carcinogenic contaminant called 1,4-dioxane in a range of ingredients known as ethoxylates. This contaminant is not intentionally added to these products, but is rather created, as many contaminants are, during the manufacturing processes.


Because we do have clear evidence for the identity of chemicals that induce carcinogenic effects in rodents and are thus likely to be carcinogenic in humans, our inability to do these kinds of tests is less important. We don’t actually need any more studies or scientific data to feel confident about these chemicals’ effects. We have reached the point where doing more studies becomes a superb excuse for inaction or delay on the part of industry, when they should be dealing with the problems that have already been identified.