"[The LD50] is now an anachronism.... I do not think the LD50 test provides much useful information about the health hazards to humans."

—David Rall, Ph.D., Former Director
National Toxicology Program

The Lethal Dose 50 (LD50) test involves the administration of a substance to a group of animals at increasing doses in order to determine the dose that kills 50 percent of the test subjects within a set time frame. Typically, administration of the test substance is via a tube inserted down the esophagus into the stomach. Other routes of administration include inhalation and applying the substance to the animals’ skin. The test is typically allowed to proceed for 14 days, at which time all the animals who have not died from the test substance are killed.

Animals who have not died within the test period may be sick or near death. The LD50 provides no information on what system failure led to the death of the animals. Some deaths may be due to the quantity of the test substance causing gastric rupture or other morbidity unrelated to the toxicity of the test substance.

The designers of the LD50 test in 1927 acknowledged its serious inadequacies, intending it only for certain narrow medical purposes.¹ Nevertheless, use of the LD50 test has become widespread as a general measurement of chemical toxicity. The LD50 has been challenged for decades as both unreliable and uninformative.

The LD50 Is Highly Unreliable

Small changes in test conditions can produce wildly varying outcomes. It has been well documented that species,² strain,³ and age⁴ have marked effects on LD50 results, as do weight,⁵ sex,⁴ health,⁶ diet,⁴ whether the animals are deprived of food before the test,⁷ the method by which the chemical is administered,⁸ ambient temperature,^9,10 and housing conditions of the animals.^6,11 These factors lead to LD50 measurements that differ by orders of magnitude. It is likely that other factors, including humidity, weather, noise, the light-dark cycle, and the dexterity of the laboratory personnel, can also affect the outcome of the test. A study arranged by the Commission of the European Communities found that LD50 values, based on tests of the same substances performed in different laboratories, differed by as much as a factor of 12. A second trial which attempted to standardize conditions across laboratories still yielded results differing by as much as a factor of eight from one laboratory to the next.¹² It is clear that the quantity measured by the LD50 test is not a biological constant, and that the value therefore has little significance in assessing toxicity.

The LD50 Has Little Relevance for Human Toxicity

Even if the LD50 were reliable, the information it provides is of little use to humans for several reasons.

• Species-to-species differences in sensitivity give the LD50 test little predictive capability for assessing toxicity in humans (see chart). Acetaminophen, for example, is fatal to mice at 250-400 mg/kg due to liver necrosis, while the LD50 for rats is about 1,000 mg/kg with little evidence of liver damage.¹³ With such profound differences between mice and rats, extrapolation to humans has little meaning. Indeed, a comparison of the toxicities of various chemicals for humans and animals found large differences to be typical.¹⁴ A recent multi-center study found that even under the most standardized conditions, the correlation between animal LD50 values and acute toxicity in humans was only 63 percent.¹⁵ As an Institute for Toxicology scientist has commented, "[E]ven if the LD50 could be measured exactly and reproducibly, the knowledge of its precise numerical value would barely be of practical importance, because extrapolation from the experimental animals to man is hardly possible."¹⁶
• The LD50 measures only lethality, ignoring other adverse effects which often correlate poorly with mortality. Thus a chemical can have extremely harmful but nonlethal effects at doses far short of the LD50 dosage.
• Pretreatment with small doses of some chemicals (e.g., cadmium chloride) raises the LD50 level, and other substances are lethal at 1/100 the LD50 value when taken daily.
• For pharmacologically inert compounds, the LD50 may measure properties of no significance to human exposure. For example, inosic acid, a flavor enhancer added to food in trace amounts, was found lethal at doses of 20 g/kg, not from true toxicity, but by raising stomach acidity high enough to cause corrosion of the gastrointestinal lining. An equivalent dose in humans would flavor six tons of food.¹⁷
• Roughly 80 to 90 percent of poisonings involve children under five years of age, who commonly react very differently from adults to chemical substances. A study comparing toxicity in newborn and adult animals found large variations due to species-specific developmental patterns that cannot be readily extrapolated to human infants.⁴
• In practice, 50 percent of adult overdoses and 90 percent of narcotic overdoses involve mixtures of drugs, and often the substances ingested are not known. The LD50 test does not account for drug interactions, and is therefore of little use in such cases.¹⁸

The serious inadequacies of the LD50 test leave it "only marginally informative, toxicologically inadequate, and misleading."¹⁹

The LD50 Is a Poor Choice of Test

Modifications, such as the up-down and limit tests, are simply refinements of the classic LD50 test, and suffer from the same deficiencies. However, in vitro methods are available that produce highly reliable results and provide more predictive information about the effects of chemicals on human beings. As David Rall, Ph.D., then-director of the National Toxicology Program (NTP), wrote in March 1983, the LD50 "is now an anachronism….I do not think the LD50 test provides much useful information about the health hazards to humans."²⁰ The LD50 is a highly unsatisfactory measure of toxicity in humans.

1. Trevan JW. The error of determination of toxicity. Proc Roy Soc 1927;101B:483-514.
2. Morrison JK et al. The purpose and value of LD50 determinations. Modern Trends in Toxicology 1968; Butterworths, London: p.1.
3. Dieke SH, Richter CP. Acute toxicity to rats in relation to age, diet, strain, and species variation. J Pharmacol Exp Ther 1945;83:195-202.
4. Goldenthal EI. A compilation of LD50 values in newborn and adult animals. Toxicol Appl Pharmacol 1971;18:185-207.
5. Balazs T, Arena E, Barron CN. Protection against the cardiotoxic effect of isoproterenol HCl by restricted food intake in rats. Toxicol Appl Pharmacol 1972;21(2):237.
6. Weil CS, Wright GJ. Intra- and interlaboratory comparative evaluation of single oral test. Toxicol Appl Pharmacol 1967;11:378-8.
7. Quinton RM, Reinert H, cited by Morrison et al.; 1968.
8. Ferguson HC. Dilution of dose and acute toxicity. Toxicol Appl Pharmacol 1962;4:759-62.
9. Fuhrman GJ, Fuhrman FA. Effects of temperature on the action of drugs. Ann Rev Pharmacol 1961;1:65-78.
10. Wiehe WH. The effect of ambient temperature on the action of drugs. Ann Rev Pharmacol 1973;13:409-25.
11. Wilberg HC, Grice. Effect of prolonged individual caging on toxicity parameters in rats. Food Cosmet Toxicol 1965;3:597-603.
12. Hunter WJ et al. An intercomparison study conducted by the Commission of the European Communities on the determination of the single administration toxicity in rats. Communicated by the Health and Safety Directorate (unpublished). 1977, Commission of the EC and United States EPA, 1979.
13. Jollow DJ, Thorgeirsson SS, Potter WZ, Hashimoto M, Mitchell JR. Acetaminophen-induced hepatic necrosis. VI. Metabolic disposition of toxic and nontoxic doses of acetaminophen. Pharmacology 1974;12(4-5):251-71.
14. Müller R. Vergleich der im Tierexperiment und beim Menschen rödlichen Dosen wichtiger Pharmaka. Diss Univ Frankfurt/Main 1948.
15. Multicenter Evaluation of In-Vitro Cytotoxicity Tests Trial, 1989-1999.
16. Lorke D. A new approach to practical acute toxicity testing. Arch Toxicol 1983;54(4):275-87.
17. Zbinden G, Flury-Roversi M. Significance of the LD50-Test for the Toxicological Evaluation of Chemical Substances. Arch Toxicol 1981;47:77-99.
18. Kaufmann SR, Cohen MJ. The clinical relevance of the LD50. Vet Hum Toxicol 1986:29(1):39-41.
19. Sperling F. Nonlethal parameters as indices of acute toxicity: inadequacy of the acute LD50. New concepts of safety evaluation 1976; John Wiley and Sons, NY: p. 177.
20. Spira H. Winning with archimedian principles. ATLA 1985;13:117-22.