Deriving Consistent Physiological Scaling Across Body Sizes

Why Scaling Needs a Whole-Body Check

Scaling each organ independently can produce plausible-looking values that are inconsistent as a set. Organ volumes may add to more or less than the target volume, and organ blood flows may sum to more or less than cardiac output. In a PBPK model, those discrepancies are not just bookkeeping errors: they affect the amounts of tissue available to contain drug and the blood delivered to each compartment.

The earlier material on body-size descriptors, organ volumes, and cardiac-output reallocation provides the physiological context for choosing what changes with size. Here the emphasis is on deriving a consistent set of scaling equations. The key distinction is between a raw allometric weight, which describes relative scaling, and a constrained allocation, which also satisfies a whole-body total.

The descriptor \(B\) must be chosen for the model's purpose and used consistently: it might be body mass, lean mass, or another defined measure. A ratio such as \(r\) is dimensionless, so \(w_i^V\) retains volume units and \(w_i^Q\) retains flow units. The exponents \(b_i\) and \(c_i\) need not be equal across compartments. They are assumptions or estimates about relative scaling, not guarantees that the resulting values meet global constraints.

Before applying a constraint, define exactly what is being summed. For volume, use mutually exclusive compartments that partition the specified target volume \(V_T(r)\). If modeled organs do not account for the entire body, include an explicit residual compartment or define a smaller target such as total modeled-organ volume. Do not add a blood volume separately if it is already included in the compartments' stated volumes. For flow, include the systemic destinations that partition the cardiac output being used; do not count the same flow twice by including both a parent vessel and its downstream organs.

Constrained Allometric Equations

For \(m\) disjoint volume compartments with positive raw weights, define the constrained volume of compartment \(i\) as its normalized share of the target:

This retains the relative pattern specified by the raw weights while making the total equal to \(V_T(r)\). For systemic flows to \(k\) destinations, use the corresponding equation with a specified cardiac output \(Q_{\mathrm{CO}}(r)\):

These equations require positive target totals and positive raw weights. A reference model is recovered at \(r=1\) if the reference compartments sum to the reference target and the reference flows sum to reference cardiac output. If not, normalization still gives the specified total, but it changes the reference allocations; that is a signal to reconcile the reference data or target definition rather than quietly treating the discrepancy as harmless.

Why Normalization Satisfies the Constraints

The volume identity follows directly by summing the allocated volumes. Since the denominator is common to every compartment and the raw weights sum to that denominator:

The same argument gives \(\sum_{i=1}^{k}Q_i(r)=Q_{\mathrm{CO}}(r)\). Normalization also preserves ratios: \(V_i/V_j=w_i^V/w_j^V\), and \(Q_i/Q_j=w_i^Q/w_j^Q\). In particular, for volumes,

Thus a common multiplier applied to every raw weight cancels out. The exponents control how the relative allocation changes with body size; the target total controls the overall scale. This separation is useful, but it also makes a modeling choice explicit: enforcing a total can alter each compartment from its unnormalized allometric prediction.

Keeping Selected Flows Fixed

Sometimes a model specifies one or more flows directly, rather than allowing every flow to change in the same normalized way. Let \(F\) be the set of fixed destinations, with fixed flows \(Q_i^{\mathrm{fixed}}\), and let \(R\) be the remaining destinations. First calculate the residual cardiac output:

If \(Q_{\mathrm{res}}>0\), allocate it over the remaining destinations using their positive raw weights:

This construction preserves the fixed values and makes the total exact. It is feasible only if the fixed flows are nonnegative and do not exceed cardiac output, and if the remaining weights have a positive sum when residual flow is positive. If the residual is zero, all remaining flows must be zero. A negative residual is an incompatibility between the fixed inputs and the cardiac-output target, not a value to be repaired by normalization.

Assumptions, Diagnostics, and Alternatives

The normalization equations are accounting constructions, not physiological laws that determine the correct exponents or target totals. They assume that the compartments in each sum are defined consistently, their raw weights are meaningful and positive, and the target \(V_T(r)\) or \(Q_{\mathrm{CO}}(r)\) is itself appropriate for the subject. The equations enforce one total; they cannot establish that an individual organ's resulting value is biologically plausible.

A useful check is to calculate both the raw and constrained allocations at the reference size and at selected sizes. Confirm the following before using the values in a PBPK model:

• Reference recovery: at \(r=1\), do the baseline compartment values sum to the declared target? If not, identify whether the target, residual compartment, or reference values need reconciliation.
• Balance: does the sum of disjoint volumes equal \(V_T(r)\), and does the sum of systemic destination flows equal \(Q_{\mathrm{CO}}(r)\), within numerical tolerance?
• Units and domains: are volume terms in common volume units, flow terms in common flow units, and all weights and target totals positive where required?
• Trend and plausibility: do organ shares change in the directions implied by the exponents, and are the resulting absolute volumes and flows plausible over the intended size range?
• Definition of totals: have residual tissues and relevant flow destinations been included once, without overlapping compartments or omitting a route?

If the equations are evaluated in software, retain adequate precision internally and check the sum before display rounding. A small discrepancy at machine precision can be acceptable; a discrepancy caused by rounded inputs or omitted compartments should be investigated. Extreme values of \(r\) can also expose problems: even a mathematically valid positive allocation may become physiologically implausible when extrapolated far beyond the range for which its exponents were chosen.

The main alternative is to use unnormalized allometric predictions \(V_i=V_{i,0}r^{b_i}\) and \(Q_i=Q_{i,0}r^{c_i}\) directly. This preserves the specified absolute scaling of each compartment but does not generally satisfy a whole-body total. It may be appropriate when the component predictions are independently supported and the totals are not constrained to a common target, but it requires explicit accounting for any mismatch. A normalized allocation guarantees the total but can shift every compartment to compensate for changes in the others. A hybrid approach fixes well-supported values and normalizes only the remaining compartments, as in the residual-flow example. These approaches answer different modeling questions; none removes the need to justify the descriptor, exponents, targets, and compartment definitions.

Common Pitfall

A frequent error is to treat a raw allometric weight as an already constrained organ volume or flow. For example, if raw volumes sum to \(105.1826\) L while the target is \(100.0\) L, reporting the raw values as compartment volumes violates the specified whole-body total. The correction is to define the target and the complete set of non-overlapping compartments, divide each positive weight by the sum of all weights, and multiply by the target. The reverse error is also possible: normalization may be applied automatically even when the raw sum is meaningful and the model did not intend a fixed total. State which interpretation is intended before calculating.