Научные исследования

International scientists’ opinion on IARC classification

Ken Donaldson, David Bernstein, Allen Gibbs, Fred Pooley, Arthur Langer, John Hoskins, Jacques Dunnigan

The International Agency for Research on Cancer (IARC) is part of the World Health Organization (WHO).  IARC's mission is to coordinate and conduct research on the causes of human cancer, the mechanisms of carcinogenesis, and to develop scientific strategies for cancer control.

The IARC has several research units, one of them: the «Carcinogen Identification and Evaluation Unit».  This research unit works on the rationale that authoritative information about proven and possible human carcinogens is needed to assess the risks posed by exposure to chemical, physical and biological factors.  The sources of such exposures are varied: the workplace, the environment or the individual lifestyles (alcohol drinking, tobacco smoking).  Independent scientific evaluations of the carcinogenic potential of such exposures can be used as a basis for information, regulation and legislation by the research community, national authorities and international organizations.

The main work of the Unit is production of the IARC «Monographs» series on the Evaluation of Carcinogenic Risks to Humans.  Since its inception in 1972, the «Programme» has reviewed more than 885 agents, and IARC Monographs have become well-known for their thoroughness, accuracy and integrity.  With these data, there is an opportunity to provide meaningful information for nations to use in gauging the extent of risk to their populations regarding a substance identified as a hazard.  Information concerning route of exposure to humans, actual conditions during use, cumulative exposure to the agent, pharmacokinetics regarding its fate in the human host, are in many cases available for complete risk analysis.  However, this information is largely unexplored in the IARC classification scheme.  Indeed, one has to ask why does IARC persist in using the term ‘risk’ in the title and text of their monographs when in fact, they agree that they do not perform risk assessment.  Unfortunately, some governments and pressure groups have used hazard identification to advance the cause for banning without conducting the appropriate quantitative risk assessment to prevent unintended consequences.

Misconception of terminology 

As evaluated in the IARC Monographs Volumes 1-83, a list contains all agents, mixtures and exposures circumstances evaluated to date as being in  «Group 1» (carcinogenic to humans) (

The list was last updated April 28, 2004.  It now contains some 90 agents, mixtures and activities classified in «Group 1 – Carcinogenic to Humans»   The 90 entries are divided into three sections: «Agents and groups of agents»; «Mixtures»; «Exposure circumstances».  From each of these three sections, some are indicated below for illustrative purposes.

Agents and groups of agents:




Oestrogen therapy, post-menauposal

Oestrogens, both steroidal and non-steroidal

Oral contraceptives, sequential

Silica (crystalline, inhaled in the form of cristobalite)

Vinyl chloride

X-radiation and gamma radiation


Alcoholic beverages

Analgesic mixtures containing phenacetin

Salted fish (Chinese-style)

Tobacco smoke

Wood dust

Exposure circumstances:

Aluminium production

Boot and shoe manufacture

Furniture and cabinet making

Iron and steel foundry

Painter (occupational exposure)

Rubber industry

Solar irradiation

Tobacco smoking

Question: does the presence on the IARC list of « Group 1 - Agents, Mixtures and Activities » imply that these must be banned?  The answer is obviously NO.

The reason is because the IARC classification covers only the identification and characterization (hazard) of these agents, mixtures and activities.  It does not include the assessment of risk, i.e.: the probability of toxic manifestations under actual conditions of use today.

This is an important distinction: « hazard » is not « risk ».  The IARC classification is about hazard, not risk.  Indeed, characterizing a hazardous substance is not equal to assessing its true risk.

Hazard identification is an essential but insufficient component of risk assessment, which comprises also exposure data over time, and estimation of the likely risk under actual conditions of use.  Because of the conceptual confusion and indiscriminate use of the terms « hazard » and « risk », untoward fear of unwelcome end points such as cancer, in many sectors of the general public, is driven by hazard data misrepresented as risk data.  This misperception often results in political response to perceived fear, sometimes nurtured by media taste for sensationalism, pushing regulatory action to extremes.  The abusive use of the « Precautionary Principle » (PP) is an example of such situations.  Some governments and pressure groups put forward the PP, confident that a ban represents complete hazard elimination.  Such a simplistic view fails to weigh the cost/benefit of a complete ban vs a controlled-approach, and leads to unintended outcomes.  Consider the 90 agents and places of work where exposure to complex mixtures are causally linked to cancer.  Estrogen therapy, chemotherapeutic agents used in the treatment of some cancers, production of rubber, boot and shoe manufacturing, furniture and cabinet making and the house painter’s environment are among agents and places of work where excess cancers have been reported.  Ban of these agents and places of work could cause chaos globally.  For these reasons, IARC should include additional pronouncements in their documents to warn of potential unintended consequences of extreme control or ban.

When dealing with potentially harmful substances, the classical three-pronged approach is used:

1: hazard identification (characterization);

2: risk assessment;

3: risk management.

It must be re-emphasized that the IARC classification scheme refers only to « hazard identification ».  It does not refer to « risk assessment » which, as already mentioned, must include the various components of dose and duration of exposure.  Therefore, the IARC classification is not meant to be used as a « risk management » instrument for regulatory action, without the proper risk assessment step.

Hazard identification: A source of risk that does not necessarily imply a potential for occurrence.  A hazard produces risk only if an exposure pathway exists and if exposures create the possibility of adverse consequences.

Risk Assessment: A process that involves the integration of data, hazard identification, exposure pathways, and dose-response relationships to estimate the nature and likelihood of adverse effects.

The case of asbestos minerals.

It should be realized that the word «asbestos» is a generic, commercial term which encompasses two very different families of fibrous silicates: the serpentine and the amphiboles.  With the growing body of recent evidence regarding the distinct «hazard characterization» of chrysotile asbestos vs that of the amphibole varieties of asbestos, the time has come to better differentiate the characteristic hazards associated with the two families of asbestos.

While the current IARC classification does not make this distinction for the different varieties of asbestos, the various exercises of « risk assessment » carried over several years of investigation between the two families of asbestos have confirmed that the risk associated with the use of chrysotile asbestos is quite different from that of the amphiboles.  In fact, the amphibole asbestos minerals crocidolite and amosite produce perhaps orders of magnitude more disease than does chrysotile asbestos when the fibers are used in the same way.  Recent analyses of tremolite exposure in Libby, Montana, suggest that this amphibole asbestos mineral produces 1000 fold more disease than does chrysotile asbestos when the fibers are used in the same way. 

Finally, it is now is generally accepted that the much longer residence time (biopersistence) in the lung of inhaled amphibole fibers is one of the key factors for their much higher pathogenicity compared to chrysotile (Wagner & Pooley, 1986; Albin et al, 1994).  Recent quantitative reviews that analyzed data from available epidemiological surveys to determine potency of asbestos in relation to fiber types confirmed the difference in risk between chrysotile and the amphiboles (Hodgson & Darnton, 2000; Berman & Crump, 2004).  Recently published experimental biopersistence studies (Bernstein et al, 2003; 2004; 2005) provide strong support for the differences seen epidemiologically between chrysotile and amphibole asbestos.

Additionally, many epidemiological studies (Liddell, McDonald & McDonald, 1997) have shown no evidence of increased cancer risk from chrysotile exposure at presently regulated occupational exposure levels (~1 f/ml, 8-hour time weighted average), as recommended by the Group of Experts convened by the WHO in Oxford (1989). 

In use today, methods for controlling chrysotile exposures in the work place have been greatly improved.  Thus, instead of concentrations of 50 to 100 fibers/cm3 that occurred in the past, today typical applications are below 1 fiber/cm3.  There is no doubt that amphiboles should be avoided.  However, the weight of evidence today indicates that chrysotile asbestos can be used safely.  Chrysotile in cement products such as water pipes and cement boards for housing provide a versatile product which can withstand extremes of temperatures and weather at affordable cost for developing countries which if eliminated would cost lives rather than save lives.


Albin et al (1994). Occup. Environ. Med. 51: 205-211

Berman, DW and Crump KS (2004). Technical Support Document for a Protocol to Assess Asbestos-Related Risk. Washington DC. ATSDR

Bernstein D, Rogers R, Smith P (2003).  Inhalation Toxicology 15: 1247-1274

Bernstein D, Rogers R, Smith P (2004).  Inhalation Toxicology 16: 745-761

Bernstein D, Rogers R, Smith P (2005). Inhalation Toxicology 17: 1-14

Hodgson JT and Darnton A (2000). Occup. Hyg. 44(8): 565-601

Liddell FDK, McDonald JC and McDonald A (1997). Ann. Occup. Hyg. 41:13-35

Wagner, JC and Pooley FD (1986). Thorax 41: 161-166

 Ken Donaldson              (ken.donaldson@ed.ac.uk)
Professor of Respiratory Toxicology

The Medical School 
University of Edinburgh 
 Edinburgh, UK

 David Bernstein           (davidb@iprolink.ch)

Consultant in Toxicology

Geneva, Switzerland

 Allen Gibbs                   (allenrg@btinternet.com)

Department of Histopathology

University of Wales College of Medicine

Cardiff, UK

Fred Pooley                  (pooleyfd@cardiff.ac.uk)

Professor of Engineering

Cardiff University

Cardiff, UK

 Arthur Langer               (artlanger@aol.com)

Environmental Sciences Laboratory

Brooklyn College, City University of New York

New York, USA

 John Hoskins                 (j.a.hoskins@isbe.demon.co.uk

Consultant in Toxicology

Reigate, UK

Jacques Dunnigan         (dunnigan@abacom.com)


University of Sherbrooke

Sherbrooke, Qc, Canada