Diagnosing Implausible Obesity-Related PBPK Inputs

Why Input Audits Matter

A PBPK simulation can be numerically stable and still represent an implausible person. In obesity-related models, this can happen when inputs assembled from different sources use incompatible body-size descriptors, when a flow is counted twice, or when an organ volume is scaled as if every tissue grew in direct proportion to body weight. These errors affect the model before any drug-specific parameter or exposure prediction is considered.

A useful audit asks several distinct questions. Do the volumes fit the model’s body-volume bookkeeping? Do blood flows leaving the arterial supply sum to cardiac output? Are tissue-specific flows plausible relative to the sizes of those tissues? Finally, do the changes with body weight agree with the physiological scaling assumptions used to construct the individual? A failed check is a signal to investigate, not an automatic diagnosis of which input is wrong.

The denominator and compartment definitions matter. A sum of organ fractions can be expected to equal one only when the listed compartments are an exhaustive, mutually exclusive partition of the model’s stated volume. Some implementations include additional tissues, blood pools, or residual compartments; others use a different volume convention. Likewise, flows should be summed over destinations that partition cardiac output. A total liver flow and one of its component inflows cannot both be counted as separate destinations.

Deriving the Two Balance Checks

Suppose a model represents the body as \(k\) mutually exclusive tissue compartments that together exhaust its defined body-volume denominator. Then \(V_{\mathrm{body}}=\sum_{i=1}^{k}V_i\). Dividing both sides by \(V_{\mathrm{body}}\) gives the volume-fraction identity:

This is a bookkeeping identity under the stated partition assumption, not a universal rule that every list of organ volumes must sum to an anatomically exact whole-body volume. If the modeled list is incomplete, the remainder must be represented explicitly or the expected sum must reflect the model’s convention.

For a steady circulation in which the listed, mutually exclusive destinations receive all cardiac output, conservation of flow gives \(Q_{\mathrm{CO}}=\sum_{i=1}^{k}Q_i\). Dividing by \(Q_{\mathrm{CO}}\) yields \(\sum_i q_i=1\). If a row is duplicated with flow \(\delta\), the apparent sum becomes \(1+\delta/Q_{\mathrm{CO}}\). Thus the size of a flow-balance discrepancy can help identify a duplicated or misclassified component, although it does not prove that this is the cause.

Worked Example: Auditing a Body-Volume Partition

Checking Cardiac Output and Tissue Perfusion

The flow audit begins with a clear destination list. Confirm whether each row represents a complete organ flow or a component of another row’s flow. Then sum the destination flows and compare them with the cardiac output assigned to that individual. Record the discrepancy in both absolute units and as a fraction of cardiac output. A total near the target can still conceal offsetting errors, so row definitions and units need review as well.

A flow fraction and tissue-specific perfusion answer different questions. \(q_i=Q_i/Q_{\mathrm{CO}}\) describes how much of total cardiac output is assigned to tissue \(i\). \(Q_i/V_i\) describes flow relative to that tissue’s volume. A large tissue can receive a substantial share of total flow without having an unusually high flow per unit volume; conversely, a small tissue can have a small flow fraction but high tissue-specific perfusion. Compare the latter with the model’s intended physiological values and units, not with an arbitrary universal cutoff.

Testing Scaling With Body Weight

A common input error is to multiply every reference volume and flow by the same ratio of body weights. That operation assumes direct proportionality to total body weight for every quantity. Obesity-related physiology does not generally justify that blanket assumption: adipose tissue, lean tissues, individual organs, and cardiac output need not change in the same proportion. The earlier material on selecting body-size descriptors and constructing consistent physiological scaling provides the modeling context; here the aim is to test whether the assembled values obey their stated scaling rules.

For each quantity, compare the obese individual with the reference individual using the ratio \(R_i=V_{i,\mathrm{obese}}/V_{i,\mathrm{reference}}\) for volumes or \(R_Q=Q_{\mathrm{CO,obese}}/Q_{\mathrm{CO,reference}}\) for cardiac output. Compare these ratios with the specified physiological model, not automatically with the body-weight ratio. Also check that related inputs come from compatible individuals and definitions: an organ volume from one body-size method paired with a cardiac output from another can create a plausible-looking but incoherent set.

A Practical Diagnostic Sequence

When an audit fails, avoid changing values until the source of the discrepancy is understood. A useful sequence separates arithmetic checks from physiological judgment:

Several approaches can support this sequence. A strict balance check is effective for detecting bookkeeping inconsistencies when the model’s compartments are exhaustive. A comparison with tissue-specific physiological inputs is more informative when the model intentionally includes a residual compartment or uses a nonstandard denominator. A reference-ratio comparison helps reveal inappropriate scaling, but it does not establish that an individual value is wrong unless the model’s scaling assumptions are known. These checks complement one another; none replaces review of definitions and provenance.

Diagnostics also have limits. Biological variation means that an unusual value is not automatically an error. A balance can close even when two inputs are both wrong in offsetting directions. Conversely, a sum below one may reflect an omitted modeled compartment rather than a faulty recorded value. These audits assess internal consistency and plausibility, not whether the PBPK model predicts clinical exposure accurately.

Common Mistakes and Limitations

• Forcing a sum to one. Adjusting one volume or flow simply to close a balance hides the source of the inconsistency. First establish that the list is exhaustive and that the quantities have compatible definitions.
• Counting components and totals together. A vascular component such as hepatic arterial flow must not be added as a separate destination when the listed organ flow already includes it.
• Treating body weight as a universal scaling factor. The same weight ratio need not apply to adipose tissue, organs, muscle, and cardiac output. Use the specified tissue-specific assumptions.
• Confusing flow fraction with perfusion. A tissue’s percentage of cardiac output does not describe flow per unit tissue volume. Calculate each quantity with its own denominator and units.
• Using a balance as proof of physiological truth. A perfectly closed set of inputs can still be implausible. Independent physiological comparisons and input provenance remain necessary.