The previous article covered the basic concept of the Minimum Anticipated Biological Effect Level (MABEL) and how preclinical pharmacology, pharmacokinetics (PK), and pharmacodynamics (PD) fit together when supporting first-in-human (FIH) dose selection.
Predicting human PK from nonclinical data sits at the center of that process. For therapeutic antibodies, simple allometric scaling from cynomolgus monkey PK is one of the more tractable routes: measure clearance in monkeys, apply a scaling exponent, and arrive at an initial human estimate.
This article works through the calculation of human clearance (CL) from a published analysis of 13 therapeutic monoclonal antibodies (mAbs). It covers where the scaling exponent came from, how the approach was validated, and a step-by-step example.
Allometric scaling is not a stand-alone method for setting a MABEL dose. It supplies one component of the human PK prediction, which is then integrated with PK/PD and pharmacological information.
Why Use a Monoclonal Antibody as the Example?
Monoclonal antibodies illustrate interspecies PK scaling unusually well, because their disposition differs substantially from that of many small-molecule drugs.
Therapeutic IgG antibodies are large, distribute mainly within plasma and extracellular fluid, and persist for long systemic half-lives. FcRn-mediated recycling, proteolytic catabolism, and - depending on the target - target-mediated drug disposition (TMDD) all shape the observed PK profile.
When elimination is predominantly non-target-mediated, or when the target-mediated pathway is saturated, PK approaches a linear range over the dose interval under study. Simple allometric scaling has been evaluated most extensively in exactly this setting.
Deng et al. analyzed 13 therapeutic mAbs that showed linear PK across the tested dose ranges. Where antigen-mediated clearance was substantial, the analysis drew clearance values from doses at which that pathway was saturated. Under those conditions, cynomolgus monkey PK tracked human clearance using a fixed allometric exponent of 0.85.
That qualifier matters. A fixed allometric exponent does not transfer automatically to every antibody or every dose level. Where target-mediated clearance remains strongly nonlinear, a mechanistic or nonlinear PK model is the better tool.
The Basic Principle of Allometric Scaling
Allometric scaling describes the relationship between a physiological or pharmacokinetic parameter and body weight as a power function:
Parameter = a × Body Weightᵇ
where a is a proportionality coefficient and b is the allometric exponent.
For interspecies prediction, the equation is written as:
CLhuman = CLmonkey × (BWhuman / BWmonkey)ʷ
where:
- CLhuman is the predicted human clearance;
- CLmonkey is the observed cynomolgus monkey clearance;
- BWhuman and BWmonkey are the body weights used for scaling;
- w is the interspecies scaling exponent.
The exponent is not a universal constant. It varies with the PK parameter, molecule class, species, and underlying biology. For therapeutic antibodies, reported clearance exponents cluster in the approximate range of 0.8–0.9, while volume parameters sit closer to 1.0.
The widely used value of 0.85 for cynomolgus monkey-to-human clearance scaling is therefore an empirical parameter drawn from specific mAb datasets. It is not a universal physiological law.
How Was the 0.85 Exponent Established?
A frequently cited analysis by Deng et al. examined 13 therapeutic mAbs with linear PK and compared several approaches to predicting human clearance, including conventional multi-species allometric scaling and scaling from cynomolgus monkey PK alone.
The simplified approach used the relationship:
CLhuman = CLcyno × (BWhuman / BWcyno)ʷ
Four antibodies served as the training dataset. Using observed cynomolgus monkey and human clearance values with typical body weights, the investigators calculated an antibody-specific scaling exponent for each molecule.
Those exponents ranged from 0.776 to 0.875, with a mean ± SD of 0.831 ± 0.042. A nonlinear mixed-effects analysis of the same four antibodies estimated a population mean exponent of 0.826, with a 95% confidence interval of 0.805–0.845.
Those values supported an exponent close to 0.85.
Independent Validation
Fitting the exponent was the easy part. The more consequential step was testing it against antibodies that had not been used to derive it.
For the remaining nine mAbs, the exponent was fixed at 0.85 and human clearance was predicted from cynomolgus monkey clearance. Prediction error was calculated as:
PE (%) = [(Predicted CL − Observed CL) / Observed CL] × 100
Absolute prediction error stayed below 50% for all nine antibodies in the validation dataset. Across the full set of 13 antibodies, the estimated exponent was 0.847 ± 0.07, with a range of 0.733–0.970. Those results supported 0.85 as a fixed exponent for this simplified scaling approach.
The point is not that 0.85 is the correct exponent for antibodies in general. The value was supported by a defined dataset and then evaluated in an independent validation set. That sequence is what gives it weight.
A Worked Example
Consider a hypothetical therapeutic antibody with a measured cynomolgus monkey clearance of:
CLcyno = 1 mL/h
Assume a representative cynomolgus monkey body weight of 3.5 kg and a human body weight of 70 kg.
The body-weight ratio is therefore:
70 / 3.5 = 20
Using the fixed clearance exponent of 0.85:
CLhuman = 1 × 20⁰·⁸⁵
This gives:
CLhuman ≈ 12.8 mL/h
Converting to a daily clearance:
12.8 mL/h × 24 h ≈ 307 mL/day
or approximately:
0.31 L/day
The exponent is the whole story of this calculation. A 20-fold difference in body weight does not produce a 20-fold difference in clearance. At an exponent of 0.85, predicted clearance rises by approximately 12.8-fold.
That is the arithmetic consequence of a sublinear allometric relationship.
What Does the Scaling Exponent Mean?
The exponent sets how fast a PK parameter changes with body weight.
At an exponent of 1.0, a 20-fold increase in body weight would correspond to a 20-fold increase in the parameter.
At 0.85, the corresponding increase is approximately 12.8-fold.
For therapeutic antibodies, volume parameters have often been found to scale roughly in proportion to body weight, whereas clearance generally follows a somewhat lower exponent. Distribution volume and clearance are governed by different physiological processes, which is where the difference originates.
Treat the exponent as an empirical scaling parameter. Reading it literally as a measure of the "efficiency" of individual clearance organs overstates what it can support.
When Is This Approach Appropriate?
Simple allometric scaling from cynomolgus monkey PK is useful when the objective is an initial human PK estimate for a therapeutic antibody with approximately linear PK over the relevant dose range.
The approach fits when:
- cynomolgus monkey is a pharmacologically relevant species;
- the antibody shows approximately linear PK over the dose range used for the analysis;
- target-mediated clearance does not dominate the observed clearance, or the relevant pathway is sufficiently saturated;
- the available monkey PK data are of adequate quality for estimating clearance.
Reliability drops when strong nonlinear PK persists, when species differences in target biology are substantial, or when the antibody's disposition characteristics are not adequately represented by a simple empirical scaling relationship.
For such programs, mechanistic TMDD models, population PK approaches, PBPK models, and other model-informed methods add information. Recent reviews have shown that different antibodies can yield different optimal scaling exponents, which is a reason to weigh molecule-specific biology rather than treating 0.85 as a universal rule.
How Does This Fit Into MABEL Dose Selection?
The calculation earns its place once it goes back into the broader MABEL framework.
MABEL exists to estimate a human dose expected to produce a minimal biological effect, particularly where conventional toxicology-driven approaches do not adequately capture the pharmacological risk of a biologic.
Human PK prediction is one component. Once an estimated human clearance and other PK parameters are available, they are combined with pharmacological information such as target binding, receptor occupancy, concentration-response relationships, and relevant in vitro and in vivo PD data.
The resulting PK/PD relationship then supports an estimate of the human exposure associated with a defined level of biological activity.
The logic runs in one direction:
Cynomolgus monkey PK → Human PK prediction → Human exposure prediction → PK/PD integration → MABEL dose assessment
The 0.85 scaling equation addresses the first step of that chain only. On its own it does not determine the MABEL dose.
That boundary matters most for biologics with potent or potentially high-risk pharmacology. Current regulatory and scientific practice emphasizes integrating pharmacological and PK information when establishing an FIH starting dose. In June 2026, the FDA issued a draft guidance specifically addressing QSP-based approaches to MABEL dose selection, part of a broader move toward model-informed FIH dose selection.
Practical Takeaway
The simplified allometric approach compresses to one line:
Human CL ≈ Cynomolgus monkey CL × (Human body weight / Monkey body weight)⁰·⁸⁵
For the 13 therapeutic mAbs evaluated by Deng et al., this gave a useful empirical relationship between cynomolgus monkey and human clearance for antibodies with linear PK over the analyzed dose ranges.
Its value is simplicity: a well-characterized NHP PK parameter converts into an initial human estimate without a full mechanistic model. The limitation is just as real. The calculation rests on empirical observations from a defined class of antibodies and does not substitute for assessment of target biology, nonlinear clearance, species differences, or other available PK/PD evidence.
The question that matters in translational pharmacology is not whether an allometric equation can be computed. It is whether the underlying PK assumptions hold for the molecule and dose range under evaluation.
References
- Deng R, et al. Projecting human pharmacokinetics of therapeutic antibodies from nonclinical data: What have we learned? mAbs. 2011;3(1):61–70.
- Wang W, et al. Pharmacokinetics of monoclonal antibodies and Fc-fusion proteins. Protein & Cell. 2018;9:15–32.
- Müller PY, et al. The minimum anticipated biological effect level (MABEL) for selection of first human dose in clinical trials with monoclonal antibodies. Current Opinion in Biotechnology. 2009;20(6):722–729.
