Encoding Genotype Effects on Intrinsic Clearance

From Enzyme Function to Clearance

A genotype does not directly specify a drug’s total clearance. It informs the function of one or more enzymes, which contribute to particular metabolic pathways. The PBPK translation therefore has two distinct steps: map the genotype-derived activity score to relative function of the relevant enzyme, then apply that function to the enzyme’s contribution to intrinsic clearance. Other metabolic routes and non-genetic sources of variability should not be changed unless there is a separate reason to change them.

This distinction matters because hepatic clearance is a nonlinear function of intrinsic clearance. Even a proportional change in one enzyme’s pathway contribution will generally not produce the same proportional change in total intrinsic clearance, hepatic clearance, or exposure.

The multiplier \(g\) must reflect the functional interpretation of the alleles and the activity-score system being used. It is not automatically equal to the activity score itself. A proportional mapping such as \(g(AS)=AS/AS_{\mathrm{ref}}\) is appropriate only when that relationship is supported for the enzyme and substrate being modeled. A calibrated lookup table or other justified mapping may be preferable when allele function is not proportional to score. The earlier material on rare alleles is especially relevant when a functional assignment is uncertain: uncertainty should be represented in the mapping or assessed in sensitivity analyses rather than hidden by a precise-looking multiplier.

For multiple pathways, write the total as a sum, for example \(CL_{\mathrm{int,total},i}=CL_{\mathrm{int,target},i}+CL_{\mathrm{int,other},i}\). A change in the target enzyme does not, by itself, imply a change in \(CL_{\mathrm{int,other},i}\). If non-genetic variability has already been assigned to the target pathway, preserve that factor when comparing genotypes for the same hypothetical subject. Applying a new, unrelated variability draw in each genotype comparison would mix the genetic contrast with an avoidable change in physiology.

Connecting Intrinsic Clearance to Hepatic Exposure

For the well-stirred liver model, hepatic blood clearance is determined by hepatic blood flow \(Q_H\), the unbound fraction in blood \(f_{u,B}\), and total intrinsic clearance \(CL_{\mathrm{int}}\). The model assumes that blood leaving the liver is in equilibrium with the well-mixed liver compartment, and uses constant flow and binding over the modeled range. The equation below is for hepatic clearance in blood-clearance units.

For a hypothetical intravenous dose \(D\), with linear and time-invariant pharmacokinetics and hepatic clearance as the only elimination route, \(AUC=D/CL_H\). Substitution into the well-stirred equation gives a useful expression for understanding the contrast:

This identity shows why halving intrinsic clearance does not generally double AUC under the well-stirred model. The flow-dependent term remains, and hepatic clearance cannot exceed \(Q_H\). It also gives a direct monotonicity result, with other quantities fixed:

Thus, when the assumptions hold, a decrease in total intrinsic clearance increases AUC. The size of the increase depends on the starting clearance, hepatic flow, binding, and the fraction of total intrinsic clearance affected by the genotype.

Worked Genotype-to-Exposure Calculations

Assumptions, Diagnostics, and Alternatives

The calculations above require more than a genotype assignment. Check that the functional mapping applies to the enzyme and substrate, the target pathway is represented as a distinct contribution, and the other contributions are not inadvertently rescaled. The well-stirred model uses hepatic blood flow and blood unbound fraction; mixing blood-clearance quantities with plasma clearance or using inconsistent units can produce a numerically tidy but physiologically invalid result. For intravenous AUC to equal dose divided by hepatic clearance, assume linear pharmacokinetics and no material non-hepatic elimination. For oral dosing, bioavailability, including hepatic first-pass extraction, may also change, so dose divided by hepatic clearance alone is not the oral AUC calculation.

Useful checks include verifying that \(g(AS_{\mathrm{ref}})=1\), that a null-function assignment affects only the relevant pathway, and that all pathway contributions are nonnegative. Under the stated well-stirred assumptions, \(CL_H\) must be less than \(Q_H\); a computed value above hepatic blood flow signals an equation, unit, or parameterization error. With \(Q_H\) and \(f_{u,B}\) held fixed, decreasing target function should not increase total intrinsic clearance or decrease AUC. A result that violates this direction warrants checking the mapping and arithmetic.

The principal alternative is to assign a genotype-specific multiplier directly to total intrinsic clearance. That approach is defensible only if the multiplier was calibrated for the whole drug clearance and appropriately represents all relevant pathways; it is generally not equivalent to changing one enzyme’s contribution. Another option is to use genotype-group estimates of total clearance directly, but this can be less mechanistic and may not transfer well to a different drug or population. A pathway-specific PBPK representation makes the causal assumption explicit while allowing non-target pathways to compensate.

The mapping also has limits. Activity scores summarize functional assignments and may not capture substrate-specific activity, uncertain allele function, or the full range of enzyme expression. A modeled reduction in one pathway may have little effect when that pathway contributes only a small fraction of total clearance. Conversely, a dominant pathway can produce a larger contrast. These calculations are conditional model predictions, not proof that genotype alone determines an individual’s clinical exposure.