# Dose Response Relationships

Most of the quantitative information concerning the toxicity of substances is obtained from experiments performed by administering doses of the substances to small animals, although recently tests on certain bacteria have been used to determine whether a substance is likely to be a carcinogen or not. Owing to practical considerations including cost and time, most experiments involve acute rather than chronic toxicity, even though it is the latter that usually is of primary interest in environmental science. To determine directly the effects of continuous, low-level exposures over long periods would require a very large number of test animals and a long project time. The practical alternative is to evaluate the effects using high doses—at which point the effects are substantial and clear-cut—and then extrapolate the results down to environmental exposures. Unfortunately, there is no assurance that such extrapolations are always reliable, since the cellular mechanisms that produce the effects at high and low doses could differ.

The dose of the substance administered in toxicity tests is usually expressed as the mass of the chemical, usually in milligrams, per unit of the test animal's body weight, usually expressed in kilograms, thus giving units of milligrams per kilogram, mg/kg. The division by body weight is necessary because the toxicity of a given amount of a substance usually decreases as the size of the individual increases. (Recall that the maximum recommended doses for medicines such as headache remedies are smaller for children than for adults, primarily because of the difference in body masses.) It is also assumed that toxicity values obtained from experiments on small test animals are approximately transferable to humans provided the differences in body weight arc taken into account. Normally the toxicity of a substance increases with increasing dose, although exceptions are known.

### PROBLEM 10-6

If a dose of a few tenths of a microgram of a certain substance is sufficient to kill a mouse, approximately what mass of the substance would be fatal to you? What average level of the substance would have to be present in the water you drink if you were to receive a fatal dose from this source in a week? Note that your weight in kilograms is that in pounds divided by 2.2.

Individuals differ significantly in their susceptibility to a given chemical: Some respond to it even at very low doses, whereas others require a much higher dose before they respond. It is for this reason that scientists created dose-response relationships for toxic substances, including environmental agents. A typical dose-response curve for acute toxicity is illustrated in Figure 10-5a. The dose is plotted on the (horizontal) x-axis, and the cumulative

C1GURE 10-5 Dose-response curves for (a) linear dose scale; (b) logarithmic dose scale.

(a) Linear dose scale

6 8 10 12 14 Dose (linear scale in mg/kg)

(b) Logarithmic dose scale

Threshold

Dose (logarithmic scale)

(b) Logarithmic dose scale

Threshold

Dose (logarithmic scale)

6 8 10 12 14 Dose (linear scale in mg/kg)

(a) Linear dose scale percentage of test animals that display the measured effect (e.g., death) when administered a particular dose is shown on the (vertical) y-axis. For example, in Figure 10-5a, about 60% of the test animals were affected by a dose of about 4 mg/kg.

Because the range of doses on such graphs often exceeds an order of magnitude, and because the effects at the low end of the concentration scale are often important in environmental decision making and cannot be seen clearly using linear scales, the dose-response plot is often recast by using a logarithmic scale for doses. Usually an S-shaped or sigmoidal-shaped curve results from this transformation—see Figure 10-5b.

Most often, the response effect on test animals that is used to construct dose-response curves is death. The dose that proves to be lethal to 50% of the population of test animals is called the LDS0 value of the substance; its determination from a dose-response curve is illustrated by the dashed lines in Figure 10-5a. The smaller the value of LD50, the more potent (i.e., more toxic) is the chemical, since less of it is required to affect the animal. A chemical much less toxic than that illustrated in Figure 10-5b would have a sigmoidal curve shifted to the right of the one shown.

Many sources quote values for the LOD50, the lethal oral dose, when the chemical has been administered orally to the test animals, as opposed to dermal or some other means of delivery. For example, the LOD50 value for DDT for rats is about 110 mg/kg. As mentioned previously, the presumption is usually made that LD50 and LOD50 values are approximately transferable between species. In the case of DDT, for example, humans are known to have survived doses of about 10 mg/kg, so presumably the LOD50 value for humans is greater than 10 mg/kg. However, we have no direct evidence that the 110 mg/kg value for rats is also valid for humans.

Of much more concern than the acute toxicity of DDT is its ability to cause chronic effects in humans, such as cancer. Although DDT is not traditionally considered to be a human carcinogen, some small-scale epidemiological studies found that the higher the concentration of DDE in a woman's blood, the more likely she was to have contracted breast cancer. However, later full-scale studies in the United States have failed to confirm this association between breast cancer and DDE. Thus it does not seem that lifetime exposure to DDT is an important cause of breast cancer, but these studies do not address the issue raised recently of whether exposure during the teenage years, when breasts are developing rapidly, might be <i factor in developing breast cancer decades later.

The range of LD50 and LOD50 values for acute toxicity of various chemical and biological substances is enormous, and spans about ten powers of 10. Indeed, all substances are toxic at sufficiently high doses; as the Renaissance-era German philosopher Paracelsus observed, all things are poison, and it is the dose that differentiates a poison from a remedy. The World Health Organization (WHO) has devised descriptors for four broad levels of toxicity for substances, especially pesticides; the range and descriptor for each class is shown in

WHO and U.S. EPA Pesticide Hazard Categorization

TABLE 10-4

WHO and U.S. EPA Pesticide Hazard Categorization

TABLE 10-4

 WHO Examples Category U.S. EPA WHO LOD50t Synthetic "Natural" Number Category* Description (m^kg) Pesticide Pesticide la I Extremely hazardous <5 aldicarb; parathion; methyl parathion; turbufos lb I Highly hazardous carbofuran; dichlorvos nicotine II II Moderately hazardous 50-500 carbaryl; chlorpyrifos; diazinon; dimethoate; endosulfan; fenitrothion; lindane; paraquat; propoxur permethrin; Pyrethrins; rotenone III III Slightly hazardous 500-5000 alachlor; malathion; metolachlor; 2,4-D family allethrin

III IV >5000

* The United States EPA does not distinguish between WHO classes la and lb but uses a single category I. The EPA also defines a fourth category, IV, for substances with LOD50 values greater than 5000 mg/kg.

tThe LOD50 values quoted are for the solid form and are based upon experiments with rats; LD50 ranges are a factor of 2 higher.

Lethal dose ranges for liquids are a factor of 4 larger than their respective LD50 and LOD50 ranges.

III IV >5000

* The United States EPA does not distinguish between WHO classes la and lb but uses a single category I. The EPA also defines a fourth category, IV, for substances with LOD50 values greater than 5000 mg/kg.

tThe LOD50 values quoted are for the solid form and are based upon experiments with rats; LD50 ranges are a factor of 2 higher.

Lethal dose ranges for liquids are a factor of 4 larger than their respective LD50 and LOD50 ranges.

Table 10-4, along with some examples of pesticides that fall in each. All the pesticides in the extremely toxic category la are synthetic, but nicotine—which as a solution of its sulfate salt has been used as an organic insecticide in gardens for many decades—falls in category lb, highly hazardous. The U.S. EPA classifies pesticides in a similar fashion to WHO but does not distinguish between the two sublevels of WHO's category I. Category II substances, moderately hazardous, include many pesticides—both synthetic and organic—still on the market. WHO's category III, slightly hazardous, is subdivided into III and IV by the U.S. EPA.

In the dose-response curves for some substances, there exists a dose below which none of the animals are affected; this is called the threshold, and it is illustrated in Figure 10-5b. The highest dose at which no effects are seen lies slightly below it and is called the no observable effects level (NOEL), although sometimes the two terms are used interchangeably. It is difficult to determine the threshold or NOEL level: It may be that if more animals were involved in a particular study, effects at low doses might be uncovered that are not apparent with only a small number of test animals. Most toxicologists believe that for toxic effects other than carcinogenesis there is probably a nonzero threshold for each chemical. A few scientists hold the controversial view that for some substances the curve in Figure 10-5b actually falls below the zero or NOEL value for very low concentrations before returning to zero at zero dose, indicating that tiny amounts of these substances could have a positive rather than a negative effect on health.

Experiments involving test animals are also used to determine how carcinogenic a compound is. However, the simple Ames test, which uses bacteria, can be used fairly reliably to distinguish compounds likely to be human carcinogens from those that are not.

A parameter that is useful in judging whether a specific chemical is present in an environmental sample in dangerous amounts or not is the lethal concentration, LC, of the substance. Usually this is listed as its LC50, the concentration of the substance that is lethal to 50% of a specified organism within a fixed exposure period. LC50 values may refer to the concentration of a substance in air or in aqueous solution to which the organism is exposed and usually have units of milligrams per liter. For example, the LC50 for rainbow trout for a four-day exposure to endosulfan in water is only 0.001 mg/L; indeed, this insecticide is "supertoxic" to many fish species. The LC50 for a shorter exposure would be a value greater than 0.001.