Few numbers carry more weight than the first dose given to the first human. Set the Maximum Recommended Starting Dose too high and you expose healthy volunteers to unacceptable risk; set it too low and you burn cohorts climbing a needless dose-escalation ladder. It is the calculation a toxicologist reaches for constantly, and the FDA has told us exactly how to do it — by converting an animal dose to a Human Equivalent Dose through body surface area, then applying a safety factor to arrive at a defensible starting point for a first-in-human trial.
The starting point is the No Observed Adverse Effect Level, the NOAEL, established in your most relevant toxicology species. The instinct is to scale it by body weight — milligrams per kilogram, straight across. That instinct is wrong, and the FDA's 2005 guidance on estimating the maximum safe starting dose is explicit about why. Physiological processes that govern how a drug is cleared — metabolic rate, cardiac output, glomerular filtration — do not scale linearly with body weight. They scale far more closely with body surface area, which is roughly proportional to body weight to the two-thirds power. A mouse metabolises on a very different clock from a human, and normalising by surface area is what corrects for it.
That normalisation is captured in a single, elegant device: the Km factor, the ratio of an animal's body weight to its surface area. The FDA tabulates these values, and they are the engine of the whole conversion — the human Km is 37, the mouse 3, the rat 6, the rabbit and monkey 12, the dog 20. The Human Equivalent Dose follows directly: you multiply the animal dose in mg/kg by the ratio of the animal Km to the human Km. In practice, a mouse NOAEL is divided by roughly 12 to reach the HED, a rat NOAEL by about 6, a dog NOAEL by under 2. The body-surface-area Km ratio is doing the translational work that a naïve mg/kg conversion silently gets wrong — usually by overestimating the human dose from small rodents, exactly the direction that endangers volunteers.
One HED is never the answer on its own, because you will usually have toxicology in more than one species. The guidance directs you to convert the NOAEL from every tested species to its HED and then, absent a specific scientific reason to prefer one, to carry forward the lowest — the most sensitive species. This is a deliberately conservative choice. The species that yields the smallest Human Equivalent Dose is the one warning you loudest, and in the absence of data telling you humans will behave like a less sensitive species, that warning governs.
The HED is a ceiling of biological equivalence, not a starting dose. To reach the Maximum Recommended Starting Dose you divide the HED from the most sensitive species by a safety factor, and the default is 10. That factor is not a ritual; it is an adjustable acknowledgement of uncertainty. You raise it — to 25, 50, or higher — when the dose-response curve is steep, when toxicities are severe or irreversible or hard to monitor, when the target is novel, or when the animal models are poor predictors of human effect. You may, more rarely and with justification, lower it. The result, expressed as a dose for a standard 60 kg adult, is the number that anchors your first cohort.
Two checks keep that number honest. First, the MRSD should sit comfortably below the pharmacologically active dose — a starting dose that is already pharmacologically active defeats the purpose of a cautious escalation. Second, for biologics, immunostimulatory agents, and drugs acting on high-risk targets, a NOAEL-driven MRSD can itself be too high; here the Minimum Anticipated Biological Effect Level, the MABEL, derived from pharmacology and receptor occupancy, may set a lower and more appropriate starting dose. The lesson of past first-in-human incidents is that surface-area scaling alone is not always conservative enough.
None of this is conceptually hard, but it is procedurally unforgiving. The errors that reach a protocol are almost always mechanical: scaling by mg/kg instead of surface area, applying the wrong Km, forgetting to select the most sensitive species, or overlooking MABEL for a molecule that demanded it. Because the calculation is run so often, those small mechanical slips are precisely where risk accumulates.
The MolWard platform was built to take that mechanical risk off the table. Our HED & MRSD tool converts an animal NOAEL to the Human Equivalent Dose using the FDA body-surface-area Km ratios, compares species to identify the most sensitive, applies your chosen safety factor, and returns the Maximum Recommended Starting Dose as a defensible, documented number for a 60 kg adult — the calculation toxicologists reach for constantly, done consistently every time. Pair it with the PDE / Exposure Limits tool for the parallel safe-exposure question and the ICH M7 Toxicology tool for genotoxic risk, and turn dose translation into a reproducible step rather than a recurring source of error. Explore the MolWard platform.