Translating Clearance and Rate Parameterizations

Why Parameterization Translation Matters

A pharmacokinetic model can describe the same disposition process using parameters with different meanings. One representation uses clearance and volume terms, such as \(CL\), \(V_1\), \(Q\), and \(V_2\). Another uses micro-rate constants, such as \(k_{10}\), \(k_{12}\), and \(k_{21}\). Translating between these representations is essential when moving a model between software implementations or comparing parameter estimates.

Matching parameter names is not enough to establish that two implementations match. The units, equations, compartment assignment, initial conditions, and observation calculation must also agree. This tutorial develops the conversion for a linear two-compartment model and uses a hypothetical example to check equivalence.

The Two-Compartment Model and Its Conversion

Assume a linear, time-invariant model with two well-mixed compartments. Drug leaves the central compartment by systemic elimination and by distribution to the peripheral compartment; drug returns from the peripheral compartment to the central compartment. There is no peripheral elimination in this model. Let \(A_1\) and \(A_2\) be the amounts in the central and peripheral compartments. For an intravenous dose, the amount equations are:

The central concentration predicted by the model is \(C_1=A_1/V_1\). The rate constants follow from dividing each clearance by the volume of the compartment from which that clearance acts. In particular, systemic elimination and movement from central to peripheral both act on \(A_1\), while return acts on \(A_2\).

Substituting these definitions into the amount equations gives the micro-rate form:

The conversion can also be reversed if the volumes are known. The same intercompartmental clearance \(Q\) is obtained from either direction of exchange; the rate constants differ because the source compartment volumes differ.

These relationships require positive volumes and compatible units. For example, if clearance is in L/h and volume is in L, the resulting rate is in h\(^{-1}\). The two forms are algebraically equivalent only when their parameters satisfy the conversion equations.

Worked Example: Converting Clearance and Volume to Micro-Rates

Justifying Equivalence

Equivalence follows directly by substitution, rather than by relying on software-specific labels. For example, the total central loss coefficient in the clearance-volume equation is \((CL+Q)/V_1\). Applying the definitions gives:

The coefficient on \(A_2\) in the central equation is \(Q/V_2=k_{21}\). In the peripheral equation, the coefficient on \(A_1\) is \(Q/V_1=k_{12}\), and the coefficient on \(A_2\) is \(-Q/V_2=-k_{21}\). Thus every coefficient in the two amount equations matches. With the same initial amounts, the two formulations are the same linear differential equation system, so they produce the same amount trajectories. If both implementations calculate central concentration as \(A_1/V_1\), they also produce the same central concentration predictions.

Worked Example: Recovering Clearances from Rates

Representing the Same Model in Four Tools

The mathematical conversion is independent of software, but the available model definitions and parameter names differ. The table summarizes common ways to express this model; exact syntax depends on the selected model function or user-defined implementation.

For the first worked example, the equivalent clearance-volume parameter set is \(CL=5\) L/h, \(V_1=20\) L, \(Q=3\) L/h, \(V_2=30\) L. Its micro-rate set is \(k_{10}=0.25\) h\(^{-1}\), \(k_{12}=0.15\) h\(^{-1}\), and \(k_{21}=0.10\) h\(^{-1}\). A NONMEM implementation using the clearance-volume form and one using the corresponding TRANS1 rates should therefore generate matching amount and concentration trajectories, provided the dose, initial conditions, and observation calculation also match.

Diagnostics, Alternatives, and Limitations

A useful translation check has several layers. First, check units: clearance divided by volume must have units of inverse time. Second, check the mapping identities \(Q=k_{12}V_1=k_{21}V_2\). Third, compare the equations term by term, including the signs of losses and gains. Finally, run both implementations with the same hypothetical dose and compare the predicted central concentrations at selected times. Matching initial derivatives is a particularly quick check for a bolus into the central compartment.

A discrepancy may arise even when the parameter conversion is correct. Examples include different dose placement, different initial amounts, a different volume used to calculate concentration, or a mismatch in time units. Numerical solver tolerances can also create small differences in simulated values; they do not explain a large, persistent difference in the trajectories. Inspect the model equations and event setup before attributing disagreement to numerical precision.

The clearance-volume form often makes physiological quantities easier to interpret and is convenient when clearance and volume are the parameters of interest. The micro-rate form can make the differential equations compact and can be convenient when the rates themselves are specified or compared. Neither form is universally preferable: the appropriate choice depends on the intended interpretation, estimation setup, and software implementation. These forms are not interchangeable by merely renaming parameters.

The equivalence shown here is limited to the stated linear two-compartment structure. It assumes constant parameters, first-order transfer, elimination from the central compartment, and no additional processes such as time-varying clearance or nonlinear elimination. It also does not establish that the parameters can all be estimated precisely from a given data set. It establishes only that the two parameterizations represent the same model when the conversion identities hold.