The Big Pharma

Various species of animals react differently to the same drug. Not only do the variations in the metabolism of a drug make it difficult to extrapolate results of animal experiments to man but they create a serious obstacle to the development of new therapeutic drugs.-- Dr. Barnard B. Brodie in Clinical Pharmacology & Therapeutics, 1976

In toxicology and risk assessment of drug toxicity, mathematical models are used to extrapolate the risk to humans from tests on animals. However, assessing the assumed toxicological threshold dose (or NOAEL, for no observed adverse effect level) is difficult for many reasons. The models assume linear dose-response effect, and linear relationship between the percentage of chemical and the percentage of binding. In practice, age, polymorphism, diet, species, strains, sex, health conditions, route of administration, make the assumption invalid. Furthermore, the suitability of a model over another may not be justified.

For instance, multiple a priori toxic substances in people (cadmium, mercury, dioxins, pesticides /herbicides and many drugs) can increase longevity in plants or animals. Many antibiotics, anti-viral drugs, may be harmful at low doses rather than at therapeutic concentration. Anticholinesterase agents may enhance cognitive functions at low doses but decrease them at higher doses. This implies the necessity of carefully monitored clinical trials and a reduced reliance on animal testing.

Also, there is the importance of biological discrepancies among species that animal researchers tend to minimize by making the assumption that biochemical reactions in animal species are similar. Many chemical reactions within the animal body are complicated and little understood. Binding to plasma proteins and tissue constituents varies considerably among chemicals and animals. The prediction of clearance (elimination of a drug/ time/volume of blood), and elimination half-life (time needed to eliminate half of the drug) is often not feasible and is complicated by variable mechanisms of excretion among species.

Toxicologists use mathematical models that lack actual anatomical, physiological and biochemical relevance. The accuracy of such models will depend on the precise measurement and integration of many factors, including organ volume, tissue size, blood flow rates, lipid solubility, binding, permeability, transport, thermodynamic reactions, clearance, site and mechanism of action. Other inputs may be unknown. This process is very long, costly, difficult and do not fit the real data obtained during clinical trials but only theoretical curves. Thus, the data derived from animal models has poor predictive value and huge differences are noticed once human data become available.

DRUG TESTING: A GAME OF CHANCE

Between 1976 to 1985, the U.S. Food and Drug Administration (FDA) approved 209 new compounds, 102 of which were either withdrawn or relabeled because of severe and unpredicted side effects including heart attacks, kidney failure, liver failure and stroke. Many new drugs provide no advantage over existing compounds: in 1977, the FDA released a study of 1,935 drugs introduced up to April 1977, which showed that 79.4 percent of them provided little or no gain. (1) About 80 percent of new introductions in the U.K. are reformulations, or duplications of existing drugs.