This fiftieth volume of the IARC Monographs comprises monographs on five antineoplastic agents, four antimicrobial agents, two diuretics, ciclosporin (an immunosuppressant), Cimetidine (used in the treatment of gastric and duodenal ulcers), paracetamol (a popular analgesic and antipyretic drug) and dantron (a laxative). Many pharmaceutical drugs were evaluated in previous IARC Monographs (see Table 1), including some of those covered in this volume. Azacitidine and trichlormethine—both antineoplastic agents—and nitrofural—an antibacterial drug—were re-evaluated because new data on carcinogenicity in experimental animals had been published since the earlier evaluations; thiotepa and chloramphenicol were re-evaluated largely because of new data on carcinogenicity in humans.

Table 1

Pharmaceutical agents evaluated in the IARC Monographs.

Derivatives of chloramphenicol without the NO₂ moiety have been developed; of these, thiamphenicol has been used extensively, but florfenicol is not used in man. Thiamphenicol and florfenicol were not considered in this volume, however, because there appear to be no published data with regard to their carcinogenicity. Similarly, ranitidine and famotidine are used therapeutically like Cimetidine; but monographs on ranitidine and famotidine and their nitrosated derivatives were also not prepared due to a lack of relevant published studies.

In clinical use and in formulations, salts, esters and complexes of drugs are often designated by the name of the parent compound; this is the case with ampicillin and chloramphenicol. In the case of nitrofurantoin, products of different crystal size have been synthesized. While the Working Group attempted to distinguish these alternative forms, in some instances insufficient information was available to do so.

The primary source of human exposure to drugs is from their use in therapy. Other types of exposure may also occur, however: persons employed in the manufacture of drugs may be exposed, as well as nursing and other staff responsible for the preparation and administration of compounds and staff responsible for the care of treated patients. Veterinary use of drugs may result in their entry into the human food chain.

For the drugs considered here, as for many others, studies of human carcinogenicity present difficult problems. The symptoms of an undiagnosed cancer may prompt the use of drug, which is subsequently suspected as its cause. Alternatively, the condition for which the drug therapy is prescribed may itself be a risk factor for cancer. An additional problem is that patients commonly receive more than one drug, and determination of the carcinogenicity of any single drug may not be feasible. Repeated reference is made in this volume to hypothesis-generating studies. These refer to sets of data containing information on many drugs and many outcomes, in which multiple comparisons are made. Statistically significant associations (p < 0.05) are noted, but in terms of probability theory many such associations may be due to chance. For this reason, the P values given in the text must be interpreted with caution, and independent examination of associations identified in hypothesis-generating studies is particularly desirable. This situation is substantially different from that in which a prior hypothesis exists before the data are analysed.

An increasing number of agents, including pharmaceutical drugs, have been shown to inhibit cancer development in animal models. Such properties may lead to new possibilities for cancer treatment and prevention. The Working Group noted that in long-term experiments with paracetamol, nitrofurantoin and nitrofural, reductions in tumour incidence were seen at some sites in some animal species, although such reductions may have other interpretations than an inhibition of tumour induction.

Exposure can generally be much more accurately measured for drugs than for other agents suspected or identified as human carcinogens, and therapeutic doses used in humans are often close to those tested in experimental animals. However, as is the policy in the IARC Monographs, no attempt was made to quantify cancer risk at specific dose levels. As stated in the Preamble, the Monographs represent the first stage in carcinogenic risk assessment. Subsequent stages, not attempted in the Monographs, may involve quantitative determinations. By extrapolation of available epidemiological data, and possibly experimental data, estimations of risk may be attempted for specific populations in respect of particular carcinogens. Such information may be a factor in regulatory or legislative processes, but no recommendation concerning these processes is given in the Monographs. However, the Working Group responsible for the present monographs observed that inference of carcinogenic hazard was likely to be a major factor in decision-making regarding the usage of many of the drugs considered.

Many (if not most) regulatory decisions concerning putative carcinogens necessitate consideration not only of perceived hazard but also of the benefit derived from particular chemicals. It is crucial, therefore, that decisions on the availability of drugs include assessment not only of potential carcinogenicity but also the health benefit derived from their usage.