Dose Conversion Across Species

Overview

Dose conversion estimates equivalent drug doses between species — typically rodent to human, but also between laboratory species. Naïve mg/kg scaling assumes metabolism, distribution, and elimination scale linearly with body weight, which is rarely true for peptides. More sophisticated approaches account for differences in body surface area, basal metabolic rate, plasma protein binding, and clearance pathways.

A closely related glossary entry on dose extrapolation explains the underlying pharmacokinetic principles; this article focuses on the practical formulas and decisions.

The Problem with mg/kg

Mg/kg dosing assumes constant AUC per body weight, which over- or underestimates exposure because:

• Clearance scales with body surface area (or metabolic rate), not mass
• Smaller animals have higher metabolic rates and clear drugs faster on a mass basis
• Volume of distribution often scales with body weight, but clearance does not
• Plasma protein binding and receptor occupancy may differ

Consequently, a 10 mg/kg dose in a mouse typically produces much lower AUC in humans and may require only 0.8 mg/kg for equivalent exposure.

Body Surface Area (BSA) Scaling

The FDA-recommended approach for first-in-human dose selection converts animal mg/kg doses to human equivalent dose (HED) using the Km factor:

HED (mg/kg) = Animal dose (mg/kg) × (Animal Km / Human Km)

Standard Km values (kg / m²):

• Mouse (0.02 kg): 3
• Rat (0.15 kg): 6
• Rabbit (1.8 kg): 12
• Dog (10 kg): 20
• Monkey (3 kg): 12
• Human (60 kg): 37

Example: a no-observed-adverse-effect-level (NOAEL) of 10 mg/kg in mice converts to:

HED = 10 × (3 / 37) ≈ 0.81 mg/kg

For a 60 kg adult, that's ~48 mg absolute dose. A typical safety factor of 10 is then applied to choose the starting clinical dose (~4.8 mg).

Allometric Scaling

General allometric equation:

Y = a × BW^b

where Y is the parameter (clearance, volume, half-life), BW is body weight, and b is the allometric exponent. Typical values:

• Clearance (CL): b ≈ 0.75 (Kleiber's law)
• Volume of distribution (Vd): b ≈ 1.0
• Biological half-life: b ≈ 0.25

From these, the predicted human dose for a target exposure is:

Dose(human) = Dose(animal) × (BW_human / BW_animal)^0.75

Allometric scaling works best for small molecules with passive clearance. For peptides cleared by receptor-mediated endocytosis or proteolysis, the exponent can differ substantially.

Exposure-Based Conversion

Better than either mg/kg or BSA is matching target AUC or Cmax:

1. Determine AUC (see AUC) associated with efficacy or toxicity in animal studies
2. Measure or model peptide PK in the new species (in vitro clearance prediction, physiologically based PK modeling)
3. Calculate dose required to achieve the target AUC in the new species

Exposure-based dosing aligns with modern FDA guidance (including the Critical Exposure Framework and Project Optimus) and handles species differences in plasma protein binding, bioavailability, and clearance more faithfully.

Peptide-Specific Considerations

Proteolytic clearance

Peptides are cleared mainly by protease activity in plasma, kidneys, and tissues. Protease repertoires differ between species, so clearance rates do not scale neatly by body weight. In vitro stability in species-matched plasma provides a first estimate.

Renal clearance

Small peptides (<5 kDa) clear renally. Glomerular filtration rate scales roughly with body surface area. Large peptides (>50 kDa) or PEGylated conjugates clear much more slowly and require different scaling.

Target-mediated clearance

Some peptides are cleared by binding and internalization with their target receptor — target-mediated drug disposition (TMDD). In this case, clearance depends on target expression levels, which may differ between species. Mechanistic PK/PD modeling is required.

Bioavailability by route

Subcutaneous absorption efficiency differs between species. Injection volume tolerance also varies: mice tolerate ~0.1–0.2 mL SC, humans multi-mL.

Worked Examples

Example 1: efficacy dose

A peptide produces 50% efficacy at 1 mg/kg in mice (plasma AUC = 500 ng·h/mL). In vitro and in silico PK modeling predicts human clearance 4× lower per kg than mouse. Target human AUC = 500 ng·h/mL.

Human dose = mouse dose × (mouse CL / human CL) × (BW_mouse / BW_human)^0 adjustments Roughly: 1 mg/kg × (1/4) = 0.25 mg/kg in humans.

Example 2: NOAEL-to-HED

Rat NOAEL = 50 mg/kg. HED = 50 × (6 / 37) = 8.1 mg/kg. Safety factor 10 → starting clinical dose 0.81 mg/kg.

Example 3: BSA-based pediatric dose

Adult dose for the same peptide = 10 mg total. Child BSA = 0.8 m², adult BSA = 1.73 m². Pediatric dose = 10 × (0.8 / 1.73) = 4.6 mg.

When Conversion Fails

Scaling assumptions break down when:

• Target receptor density differs markedly between species
• Species-specific metabolism (e.g., cytochrome P450) activates or inactivates the peptide
• Anti-drug antibodies develop in one species but not another
• The peptide's mechanism depends on tissue-specific transporters or receptors

In these cases, species-specific PK/PD modeling and early adaptive clinical design are required rather than naive scaling.

Summary

Dose conversion is not a single formula but a decision tree. BSA scaling provides a conservative first estimate for safe first-in-human dosing. Allometric scaling works for many small molecules. Exposure-based conversion is most rigorous for peptides. Always check assumptions, and treat the converted dose as a starting point for adaptive clinical design rather than a fixed prescription.