Quantal experiments record whether each animal shows a defined all-or-none response, and the ED50 is the amount at which half would respond. The figures below are invented for an invented compound, Drug X; no real substance or real study is described.
Practice maths for exams only. Not clinical guidance; always follow your local protocol and prescriber.
Fictional data. Five groups of 10 animals receive Drug X at 2, 4, 8, 16 and 32 mg/kg, and the number showing the defined response is counted.
| Drug X (mg/kg) | log₁₀ amount | Responders out of 10 | Percentage responding |
|---|
| 2 | 0.301 | 1 | 10% |
| 4 | 0.602 | 2 | 20% |
| 8 | 0.903 | 4 | 40% |
| 16 | 1.204 | 7 | 70% |
| 32 | 1.505 | 9 | 90% |
Interpolating on the log scale. 50% lies between 8 mg/kg (40%) and 16 mg/kg (70%).
| Step | Working | Result |
|---|
| Fraction of the way from 40% to 70% | (50 − 40) ÷ (70 − 40) | 0.333 |
| log ED50 | 0.903 + 0.333 × (1.204 − 0.903) | 1.0034 |
| ED50 | 10^1.0034 | 10.08 mg/kg |
| Therapeutic index, if a fictional TD50 of 60 mg/kg was found in the same species and route | 60 ÷ 10.08 | 5.95 |
What to add in a report. Linear interpolation between two points is a quick estimate; probit or logistic regression, or the Litchfield-Wilcoxon graphical method in older manuals, uses every group and gives confidence limits. Explain that classical lethal-dose designs used large groups, and that current OECD acute toxicity guidelines use fixed-dose, acute toxic class or up-and-down procedures, which need far fewer animals: Reduction and Refinement in practice.
Educational only. Invented compound, invented numbers, for exam practice.