Why Reference-Parameter Errors Matter
A PBPK model can use plausible-looking physiological values and still describe no coherent person or animal. A table may combine an adult organ volume with a pediatric body mass, count a tissue both as a named organ and inside a residual compartment, or mix a whole-organ measurement with a value reported per unit tissue. Each number might be reasonable on its own; the problem is that the numbers do not share compatible definitions.
The practical task is therefore not just to find a published value. It is to establish what population, anatomical definition, measurement basis, and species that value represents, then check whether those match the model being built. The earlier material on reference profiles, source evaluation, uncertainty, and data-source conflicts provides useful context for this audit. Here, the emphasis is on errors that can be caught by inspecting the definition and arithmetic of selected inputs.
A Four-Part Audit
Before putting a candidate value into a parameter table, record four things: who or what was measured; what anatomy was included; how the quantity was expressed; and how directly the value applies to the modeled species. These checks are related but not interchangeable. Correcting a unit does not correct a population mismatch, and a good allometric fit does not resolve a disagreement about which tissues were measured.
Record relevant characteristics such as species, age group, sex where reported, body-size range, and physiological condition. Compare these with the intended reference subject. A value for one group is not automatically a suitable value for another.
Determine whether the source reports a named organ, a combined organ group, or a residual category such as “other tissues.” Establish whether any named compartment is already contained in that residual category.
Write the full units, including every denominator. Distinguish, for example, a whole-organ amount from an amount per unit tissue mass, and an absolute amount from an amount per kilogram of body mass.
State whether the value is measured in the target species or predicted from another species. For a prediction, identify the scaling relationship and its evidence; do not treat a convenient numerical conversion as evidence that the biological relationship is valid.
Population and Tissue-Definition Checks
Population mismatch is easy to overlook when a source table gives a single number without a detailed description. A “reference adult” value, for instance, may represent a particular body-size range or demographic group rather than every adult. If a model targets a different group, the value should be treated as a candidate requiring justification, not as a universal constant. Body-size scaling may be appropriate for some quantities, but it does not by itself establish that age, physiological state, or the original measurement definition is transferable.
Anatomical coverage is especially important when a model uses both named compartments and a lumped remainder. If the remainder already includes a separately modeled tissue, adding both values counts that tissue twice. A total that exceeds the subject's body mass or a known whole-body quantity is a warning, but a plausible total does not prove the definitions are disjoint. The source's stated inclusion rules must also be checked.
Worked Example: A Population Mismatch in an Organ-Volume Input
Suppose a hypothetical PBPK model is intended to represent a 50 kg adult. A candidate source reports a liver volume of 1.50 L for a 70 kg reference adult. A modeler proposes to use 1.50 L unchanged. For this example only, assume the model specification calls for a provisional body-mass-proportional adjustment; this assumption is a bookkeeping choice, not a general rule that liver volume is always proportional to body mass.
Under that stated assumption, the adjusted value is calculated as:
The kilograms cancel, leaving liters. The unadjusted value, 1.50 L, is 0.429 L above this provisional value. That difference identifies the consequence of using the 70 kg value without adjustment; it does not establish that 1.071 L is the true liver volume for every 50 kg adult. Before adopting either value, the modeler should confirm the target population and whether this scaling assumption is justified for the intended use. If it is not, a closer-matched reference or a defensible alternative relationship is needed.
Worked Example: Detecting a Duplicated Tissue
Consider a hypothetical 70 kg reference-human table. It lists liver mass as 1.50 kg, combined kidney mass as 0.30 kg, skin mass as 3.00 kg, and “other tissues” as 68.20 kg. The table notes that “other tissues” includes skin. A modeler adds all four entries as separate compartments.
The resulting sum is:
This exceeds the stated 70 kg body mass by \(73.00-70.00=3.00\) kg, exactly the separately added skin mass. The note shows why: skin is counted once inside “other tissues” and again as its own compartment.
There are two consistent repairs. Keep “other tissues” as a skin-inclusive remainder and omit the separately listed skin compartment, giving \(1.50+0.30+68.20=70.00\) kg. Or retain skin as a separate compartment and redefine the remainder to exclude it: \(68.20-3.00=65.20\) kg, giving \(1.50+0.30+3.00+65.20=70.00\) kg. The correct choice depends on the intended model structure and the source definitions; the entries should not be silently adjusted without documenting which definition is being used.
Measurement Bases and Species Transfer
A unit label is part of a physiological value's definition. For example, an organ-wide amount and an amount per 100 g of tissue cannot be substituted for one another without using the relevant tissue mass. Likewise, a value per kilogram of body mass is not an absolute value for a particular subject. Conversion is valid only when the denominator and the quantity being converted are known and compatible.
Species transfer adds another layer. The earlier material on allometric scaling established that a relationship such as \(Y=aM^b\) can be used to relate a physiological quantity \(Y\) to body mass \(M\) when the relationship and exponent \(b\) are supported for the quantity and species range in question. A single source value does not establish that relationship. Nor does a successful calculation establish that the source and target measurements used the same anatomical or experimental definition.
Worked Example: Converting a Tissue-Specific Flow Value
Suppose a hypothetical source reports liver flow as 21 mL/min per 100 g of liver tissue. A model table instead requires an absolute value for a liver weighing 1.50 kg. Assume, for this calculation, that the reported rate applies uniformly to the stated liver tissue mass and that both measurements refer to the same physiological condition.
First convert the liver mass to the denominator used by the source: \(1.50\ \mathrm{kg}=1500\ \mathrm{g}\), which is \(1500/100=15.00\) units of 100 g. Then multiply:
The tissue-mass denominator cancels, leaving a whole-liver rate. Using 21 mL/min as the absolute liver value would be incorrect because it would discard the “per 100 g” basis. The converted result is valid only under the stated matching and uniformity assumptions; the arithmetic cannot repair a mismatch in condition, tissue definition, or measurement method.
Worked Example: Why a Direct Species Transfer Is Not a Scaling Model
Suppose a hypothetical rat has a body mass of 0.30 kg and a measured organ volume of 12 mL. A modeler transfers the rat organ fraction directly to a 70 kg human by assuming proportionality to body mass. This gives:
The calculation is dimensionally consistent, but it assumes the same organ volume per kilogram in rat and human. It does not demonstrate that this assumption is biologically supported. If a hypothetical, independently assembled human reference table gives 1.50 L, the discrepancy is \(2.800-1.50=1.300\) L. That conflict is a reason to investigate the scaling assumption, source populations, and organ definitions—not a reason to choose one number solely because it is closer to the expected answer.
A more defensible transfer would require evidence for a relationship appropriate to the quantity and species range, along with compatible measurement definitions. If only one rat value is available, the data do not identify an across-species scaling exponent. The direct calculation should therefore be labeled as an assumption-based scenario, not presented as a supported human reference value.
Diagnostics, Alternatives, and Limitations
Several checks can reveal trouble early. First, retain the full unit string in the parameter table; if it is unclear whether a value is absolute or normalized, do not enter it as if the distinction were settled. Second, compare compartment sums with the whole-body total and inspect residual-category definitions. Third, place population and species descriptors beside each value so a mismatch is visible rather than buried in a source note. Finally, when two candidates conflict, trace each to its definition and measurement basis before averaging or selecting between them.
| Approach | Useful when | Main caution |
|---|---|---|
| Use a directly matched reference value | The source population, tissue definition, and measurement basis closely match the model. | A label such as “adult” may hide important differences; verify the source description. |
| Convert units or denominators | The source and model refer to the same quantity and population but express it on different compatible bases. | Conversion changes the representation, not the underlying evidence or population applicability. |
| Scale from another body size or species | A supported relationship applies to the quantity and range being used. | An assumed exponent or extrapolation beyond the evidence can create false precision. |
| Keep alternative values as scenarios | Sources conflict or applicability remains uncertain. | Scenarios should be labeled and justified; they are not a substitute for resolving definitions where possible. |
These checks cannot establish that a source measurement is accurate, resolve every biological difference between populations, or prove that a chosen parameter is appropriate for every modeling purpose. Mass balance and unit consistency are necessary checks, not sufficient validation. A table can sum correctly while using mismatched ages, conditions, or species. Conversely, a discrepancy in a total can reflect a legitimate difference in scope rather than a simple arithmetic error. Record the evidence, assumptions, and unresolved uncertainty so later model evaluation can interpret the inputs appropriately.
Common Mistake
Check Your Understanding
Answer each question, then compare your reasoning with the solutions below.
- A value is reported as 18 mL/min per 100 g of tissue. What information is needed before converting it to an absolute organ value?
- A table lists a named tissue separately and also states that the “remainder” includes that tissue. What is the potential error, and what are two consistent repairs?
- A value from a 60 kg reference adult is used unchanged for a 40 kg target. Name two questions to ask before deciding whether the value should be adjusted.
- One rat measurement is converted to a human value by multiplying by the ratio of body masses. What assumption does that calculation make, and what does one rat measurement fail to establish?
- A set of compartment masses sums to the stated body mass. Does this prove that no reference-parameter selection error occurred? Explain.
Solutions. (1) The relevant tissue mass, confirmation that the reported rate applies to that tissue and condition, and a compatible anatomical definition are needed; the denominator must cancel in the conversion. (2) The tissue may be double-counted. Either omit its separate entry while keeping it inside the remainder, or subtract it from the remainder and retain it separately. (3) Ask whether the population and body-size basis match, and whether a justified scaling relationship applies to that quantity; body-mass proportionality should not be presumed. (4) It assumes the same value per kilogram across the species. A single rat measurement does not establish an interspecies relationship or its exponent. (5) No. A correct sum is a useful consistency check, but it does not verify population match, measurement basis, species applicability, or the underlying source definitions.