Three Comparisons, Three Questions
A cross-species comparison can start with dose, measured drug exposure, or a specified response. These are related quantities, but they answer different questions. A dose-per-kilogram comparison describes administered dose relative to body weight. An exposure comparison asks what concentrations or exposure summaries occurred. A response-matched comparison asks what dose or exposure is needed to achieve the same defined response. Treating any one as a substitute for the others can lead to an incorrect translational conclusion.
The appropriate comparison depends on the translational estimand: the quantity of interest, the population or species being compared, and the conditions under which the comparison is made. For example, a dose-per-kilogram ratio can describe the size of two administered doses relative to body weight, but by itself it does not establish equivalent exposure or equivalent pharmacological effect. An exposure-normalized comparison may help describe exposure per unit dose, while a response-matched comparison is needed when the question is what dose in each species achieves a specified effect.
A meaningful comparison must specify the administered dose and route, the exposure metric and time interval, and the response endpoint and assessment time. Species, study conditions, and units should be explicit. If those details differ, a numerical ratio may still be calculable, but it may not represent the intended translational contrast.
Dose per Kilogram: A Dose Description, Not an Equivalence Rule
Dose per kilogram is useful for describing how much drug was administered relative to body weight. It does not account, by itself, for bioavailability, clearance, distribution, or pharmacodynamic differences. Even equal doses in mg/kg can yield different exposure in two species; unequal doses in mg/kg can yield similar exposure. Dose normalization is therefore a dose comparison, not proof of matched pharmacokinetics or response.
Worked Example: Comparing Administered Dose per Kilogram
Consider a hypothetical study in which Species A weighs 20 kg and receives 80 mg, while Species B weighs 70 kg and receives 140 mg. Calculate the dose per kilogram and compare the two quantities.
For Species A:
For Species B:
The ratio of dose per kilogram is \(d_A/d_B=4/2=2\). Species A received twice the dose per kilogram of Species B. The absolute-dose ratio is instead \(80/140=0.5714\): Species A received less total drug. These statements are not contradictory because they use different denominators. Neither ratio establishes which species had greater exposure or response; those require exposure or response data.
Exposure-Normalized Comparisons
An exposure-normalized quantity can describe exposure per unit of administered dose. For example, if AUC is measured over a specified interval, dividing it by dose per kilogram gives a quantity with units \((\text{mg}\cdot\text{h}/\text{L})/(\text{mg}/\text{kg})\), or kg·h/L. This can be compared descriptively across species, provided the AUC intervals, routes, and dose definitions are compatible.
Under a simple linear PK model after an intravenous dose, AUC over the complete concentration-time profile is \(D/CL\), where \(CL\) is clearance. After a dose with bioavailability \(F\), the corresponding expression is \(F D/CL\). Dividing by dose per kilogram gives:
This algebra shows why dose-normalized AUC is not automatically the same across species: even under linear PK, it depends on bioavailability and on clearance relative to body weight. For oral dosing, differences in \(F\) can contribute; for either route, the simple relationship requires appropriate linearity and compatible AUC definitions. A partial AUC or a nonlinear dose-exposure relationship need not obey this expression.
Worked Example: Comparing AUC per Unit Dose
In a hypothetical comparison, Species A receives 5 mg/kg and has an AUC of 40 mg·h/L over the defined interval. Species B receives 2 mg/kg and has an AUC of 18 mg·h/L over the same type of interval. Calculate the observed AUC ratio and each dose-normalized AUC.
The observed AUC ratio, Species B relative to Species A, is:
Species B's observed AUC is 45.0% of Species A's in these hypothetical data. The dose-normalized values are:
The dose-normalized ratio, Species B relative to Species A, is \(9.00/8.00=1.125\). Thus, B has 12.5% greater observed AUC per mg/kg, even though its observed AUC is lower at the administered dose. This is a descriptive comparison of the stated measurements. One dose per species cannot establish that exposure is proportional to dose; that assumption would need support from suitable data or a PK model.
The interpretation also depends on whether the goal is to compare exposure at the administered doses or exposure normalized by dose. The first is directly relevant to what was observed under those regimens. The second helps describe how much exposure was associated with each unit of dose, but its extrapolation to a new dose requires additional assumptions. If exposure is nonlinear with dose, a single dose-normalized value may not predict exposure at another dose.
Response-Matched Comparisons
A response-matched comparison begins by defining the response and its measurement conditions. The target might be a specified biomarker change at a fixed time, or a response level predicted by a concentration-response model. The doses that achieve that target can then differ across species. Matching the numerical value of a response is meaningful only if the endpoint has the same interpretation and measurement scale in both species.
The following calculation uses a simple hypothetical concentration-response model for illustration. Let \(C\) be a representative concentration, \(E_0\) the baseline response, \(E_{\max}\) the maximum change above baseline, and \(EC_{50}\) the concentration producing half of that maximum change:
For a target response \(E^*\), if \(E_0<E^*<E_0+E_{\max}\), solving the equation gives:
This expression follows by subtracting \(E_0\), multiplying both sides by \(EC_{50}+C\), and collecting the terms in \(C\). It gives a target concentration under this model; translating that concentration into a dose also requires a PK relationship. The result is conditional on the model and its parameters, not a universal dose conversion.
Worked Example: Finding Doses for a Matched Response
Suppose two hypothetical species have the same modeled baseline \(E_0=0\) response units and maximum change \(E_{\max}=100\) response units, but different \(EC_{50}\) values: 2 mg/L for Species A and 4 mg/L for Species B. The target is 50 response units. Assume the representative concentration is \(C_{\mathrm{avg}}=\mathrm{AUC}/24\ \text{h}\), and assume the hypothetical exposure-dose relationships over the relevant range are \(\mathrm{AUC}_A=12d_A\) and \(\mathrm{AUC}_B=8d_B\), with AUC in mg·h/L and dose \(d\) in mg/kg.
First, find the concentration required in each species. Since \(E^*=50\), \(E_0=0\), and \(E_{\max}=100\), the target is half of the maximum change. The model therefore gives \(C^*=EC_{50}\): 2 mg/L for A and 4 mg/L for B. Each concentration predicts:
Because \(C_{\mathrm{avg}}=\mathrm{AUC}/24\ \text{h}\), the target AUCs are:
Using the stated exposure-dose relationships, the corresponding doses are:
Under the assumed PK and response models, 4 mg/kg in Species A and 12 mg/kg in Species B each produce the target response. The dose ratio is \(12/4=3\), but the matched response does not imply matched concentration or AUC: the target concentrations and AUCs differ. This calculation is a model-based comparison, and its conclusion depends on the assumed relationships being appropriate in the dose range of interest.
Choosing and Checking the Comparison
The three approaches can be summarized by their target question and their main requirements:
| Comparison | Question answered | Important condition |
|---|---|---|
| Dose per kilogram | What dose was administered relative to body weight? | Use compatible definitions of dose and body weight; do not infer exposure or response equivalence. |
| Exposure-normalized | How much measured exposure occurred per unit of dose? | Use compatible exposure metrics, intervals, and routes; dose extrapolation may require linear PK. |
| Response-matched | What dose or exposure achieves the same specified response? | Define a comparable response endpoint and justify the PK and response models used to identify the target. |
A sound analysis checks both the calculation and the suitability of the comparison. Inspect concentration-time data or model diagnostics for evidence that the selected exposure measure is poorly characterized. Check whether exposure changes proportionally with dose before using dose-normalized values to project beyond observed doses. For response matching, examine whether the model describes the observed response over the range containing the target and whether the target is attainable in both species under the model.
Other approaches may be preferable when the question differs. Comparing exposure at the actual administered regimens is direct when the practical concern is the exposure those regimens produced. A dose-normalized comparison is a compact descriptive summary when the dose ranges and PK conditions support it. A model-based response match is more directly aligned with a target-effect question, but relies on the adequacy of both the exposure-dose and concentration-response models. None of these alone explains why species differ; mechanistic explanations require evidence beyond the ratio itself.
Common failures include comparing AUCs from different time intervals, dividing by the wrong dose (for example, total dose in one species and dose per kilogram in another), and treating a single dose-normalized value as proof of linearity. A response match can also mislead if the endpoint or assessment time differs across species, or if the target lies outside the observed response range and is obtained only by extrapolation. Report the species, dose, route, metric, interval, units, and assumptions alongside any ratio.
Check Your Understanding
Answer each question using the definitions and calculations above. Brief solutions are included after the questions.
- Species X receives 30 mg at 10 kg, and Species Y receives 60 mg at 40 kg. What is each dose per kilogram, and which species received the larger absolute dose?
- What does an AUC-per-dose ratio describe, and what additional assumption is needed before using it to predict exposure at a different dose?
- In the response model with \(E_0=0\) and \(E_{\max}=80\), what response is predicted at \(C=EC_{50}\)?
- Why does matching response not necessarily mean matching AUC?
Solutions. (1) X received \(30/10=3\) mg/kg; Y received \(60/40=1.5\) mg/kg. Y received the larger absolute dose, 60 mg versus 30 mg. (2) It describes measured exposure per unit dose for the specified metric and conditions. Prediction at another dose requires, among other conditions, a justified exposure-dose relationship; proportionality cannot be established from one dose alone. (3) \(E(EC_{50})=0+80EC_{50}/(EC_{50}+EC_{50})=40\) response units. (4) Species can have different exposure-response relationships, so different concentrations and AUCs may produce the same defined response.