Unbound-Fraction Effects in Hepatic Clearance Models

Why Unbound Fraction Matters

Hepatic enzymes act on drug molecules available in the unbound pool. Consequently, a change in the fraction unbound can change the amount of drug available for elimination even when enzyme abundance and the drug’s intrinsic metabolic capacity have not changed. The effect is not always a matching change in measured total hepatic clearance: predicted clearance also depends on hepatic blood flow and on the liver model used.

The key modeling step is to combine unbound fraction with intrinsic clearance before calculating extraction. This tutorial considers the well-stirred, parallel-tube, and dispersion models, assuming that all flow and clearance quantities use a consistent concentration basis. Blood-to-plasma conversion is not applied here.

Definition: Let $f_u$ be the fraction of drug unbound in the relevant incoming blood, $CL_{int}$ the intrinsic clearance referenced to unbound drug concentration, and $Q_H$ hepatic blood flow. The effective intrinsic clearance is $f_uCL_{int}$, and the dimensionless capacity-to-flow ratio is $$ x=\frac{f_uCL_{int}}{Q_H}. $$ The hepatic extraction ratio is $E_H=1-C_{out}/C_{in}$, and total hepatic clearance is $CL_H=Q_HE_H$.

These definitions rely on a specific set of assumptions. Binding is at rapid equilibrium, the unbound fraction is constant over the concentration range being modeled, and only unbound drug is available to the eliminating enzymes. Intrinsic clearance is linear in unbound concentration and is otherwise held fixed when $f_u$ changes. The equations also assume steady flow through the liver and use matched units for $Q_H$ and $CL_{int}$, such as liters per hour. If binding is concentration-dependent, if the unbound fraction differs between relevant compartments, or if uptake and efflux transport alter enzyme access, the simple substitution $CL_{int}\mapsto f_uCL_{int}$ may not be adequate.

How Unbound Fraction Enters Each Model

In each formulation, $f_u$ enters through $x$. The three models differ in how they translate capacity relative to flow into extraction.

Model equations: For the well-stirred model, $$ E_H=\frac{x}{1+x}, \qquad CL_H=Q_H\frac{x}{1+x} =\frac{Q_Hf_uCL_{int}}{Q_H+f_uCL_{int}}. $$ For the parallel-tube model, $$ E_H=1-\exp(-x), \qquad CL_H=Q_H\left[1-\exp(-x)\right]. $$ For the dispersion model, with dispersion number $D_N>0$, define $$ a=\sqrt{1+4D_Nx}, \qquad \frac{C_{out}}{C_{in}}= \frac{4a\exp\left(\frac{1-a}{2D_N}\right)} {(1+a)^2-(1-a)^2\exp\left(\frac{1-a}{2D_N}\right)}. $$ Then $E_H=1-C_{out}/C_{in}$ and $CL_H=Q_HE_H$.

For all three equations, increasing $f_u$ while holding $CL_{int}$, $Q_H$, and (for the dispersion model) $D_N$ fixed increases $x$. It therefore increases extraction and total hepatic clearance. The size of the increase, however, depends on the model and on the starting value of $x$. A rise in $f_u$ should not be interpreted as necessarily producing the same proportional rise in total clearance.

Worked Example: Converting Binding Information to Effective Capacity

Suppose a hypothetical drug has $CL_{int}=120$ L/h and $Q_H=90$ L/h. Its unbound fraction changes from $f_u=0.10$ to $f_u=0.20$. Calculate the effective intrinsic clearance and $x$ at each value.

At $f_u=0.10$, the effective intrinsic clearance is $$ f_uCL_{int}=0.10(120\ \text{L/h})=12.0\ \text{L/h}, \qquad x=\frac{12.0}{90}=0.1333. $$

At $f_u=0.20$, $$ f_uCL_{int}=0.20(120\ \text{L/h})=24.0\ \text{L/h}, \qquad x=\frac{24.0}{90}=0.2667. $$

The unbound fraction and the effective capacity have both doubled, so $x$ has doubled. This does not establish that total clearance will double; that requires applying the chosen liver model.

Calculating and Comparing Model Predictions

A comparison is informative only if the same $Q_H$, $CL_{int}$, and $f_u$ are supplied to each model. The following example uses the inputs above at $f_u=0.10$, with $D_N=0.500$ for the dispersion calculation. The numbers are hypothetical.

Worked Example: Clearance at One Unbound Fraction

For $Q_H=90$ L/h, $CL_{int}=120$ L/h, and $f_u=0.10$, the effective capacity is 12.0 L/h and $x=12/90=0.1333$.

Well-stirred model. $$ E_H=\frac{0.1333}{1+0.1333} =\frac{12/90}{1+12/90} =\frac{12}{102}=0.1176, \qquad CL_H=90(12/102)\approx10.59\ \text{L/h}. $$

Parallel-tube model. $$ E_H=1-\exp(-12/90) =1-\exp(-0.1333) \approx1-0.8752=0.1248, \qquad CL_H=90(0.1248)=11.23\ \text{L/h}. $$

Dispersion model. Using the unrounded ratio $x=12/90$ in the displayed substitution, $$ a=\sqrt{1+4(0.500)(12/90)} =\sqrt{1.2666667} \approx1.12546. $$ Then $\exp((1-a)/(2D_N))=\exp(1-a)\approx0.8821$, giving $$ \frac{C_{out}}{C_{in}} \approx \frac{4(1.12546)(0.8821)} {(2.12546)^2-(0.12546)^2(0.8821)} \approx0.8817. $$ Thus $E_H\approx1-0.8817=0.1183$ and $CL_H\approx90(0.1183)=10.64$ L/h.

At these inputs, the predicted clearance is about 10.59 L/h in the well-stirred model, 11.23 L/h in the parallel-tube model, and 10.64 L/h in the dispersion model. The differences reflect the model equations, not a change in the input unbound fraction.

A short derivative shows why an increase in $f_u$ raises clearance in the well-stirred model, while also showing why the increase is not generally proportional. Since $x=f_uCL_{int}/Q_H$, $dx/df_u=CL_{int}/Q_H$. Differentiating the well-stirred clearance gives

\frac{dCL_H}{df_u} =Q_H\frac{1}{(1+x)^2}\frac{dx}{df_u} =\frac{CL_{int}}{(1+x)^2}>0.

For a fixed $CL_{int}$, this derivative is positive, so increasing $f_u$ increases predicted total clearance. As $x$ grows, the derivative becomes smaller: clearance approaches the flow limit $Q_H$, and further increases in unbound fraction have less effect on total clearance. This is a statement about the well-stirred model; the other models have their own response curves.

Worked Example: Effect of Doubling Unbound Fraction

Continue with the hypothetical parameters $Q_H=90$ L/h and $CL_{int}=120$ L/h. Compare predictions at $f_u=0.10$ and $f_u=0.20$, using $D_N=0.500$ for the dispersion model.

At $f_u=0.10$, $x=12/90=0.1333$. The preceding example gives well-stirred, parallel-tube, and dispersion clearances of approximately 10.59, 11.23, and 10.64 L/h, respectively.

At $f_u=0.20$, $x=24/90=0.2667$. For the well-stirred model, $$ E_H=\frac{24}{90+24}=\frac{24}{114}=0.2105, \qquad CL_H=90(0.2105)=18.95\ \text{L/h}. $$ For the parallel-tube model, $$ E_H=1-\exp(-24/90) =1-\exp(-0.2667)\approx0.2341, \qquad CL_H=90(0.2341)=21.07\ \text{L/h}. $$ For the dispersion model, $$ a=\sqrt{1+4(0.500)(24/90)}\approx1.23828, $$ and substitution into the outlet-to-inlet ratio gives $C_{out}/C_{in}\approx0.7861$. Therefore $E_H\approx0.2139$ and $CL_H\approx19.25$ L/h.

Doubling $f_u$ increases clearance in all three predictions, but the increases are not identical: approximately 8.36 L/h for the well-stirred model, 9.83 L/h for the parallel-tube model, and 8.61 L/h for the dispersion model. The differences are consequences of how each model represents liver mixing and extraction.

Interpreting Changes, Diagnostics, and Limits

A useful diagnostic is to check the chain of calculation: unbound fraction times intrinsic clearance must have units of volume per time; dividing by flow must give a dimensionless $x$; extraction must lie between 0 and 1; and clearance must equal $Q_HE_H$, so it cannot exceed $Q_H$ under these formulations. If a calculation violates any of these checks, inspect unit consistency, the placement of $f_u$, and the extraction equation.

The models are alternatives for mapping effective capacity to extraction, not alternative definitions of unbound fraction. In the well-stirred model, the liver is represented as completely mixed; in the parallel-tube model, concentration declines along parallel flow paths; the dispersion model allows axial mixing. For a specified $x$, these formulations can yield different extraction values. Their comparison is meaningful only when inputs, concentration basis, and assumptions are aligned. The earlier material on intrinsic-clearance sensitivity provides the related context for changing enzyme capacity; here the parameter being varied is $f_u$.

The equations describe total hepatic clearance based on total incoming concentration and flow, not a universal rule for the effect of protein binding on every pharmacokinetic measurement. In a low-extraction setting, total clearance can be sensitive to $f_u$ because it is governed substantially by effective capacity. In a high-extraction setting, clearance approaches the flow limit, so changes in $f_u$ can have a smaller impact on total clearance even when they alter the unbound concentration. These qualitative patterns do not remove the need to calculate using the selected model and the relevant parameter values.

Common pitfall: “If $f_u$ doubles, total hepatic clearance must double.” This reasoning treats clearance as directly proportional to unbound fraction at every extraction level. In the well-stirred model, for example, $CL_H=Q_Hx/(1+x)$, which is nonlinear in $x$; as $x$ increases, clearance approaches $Q_H$. The correction is to update $f_uCL_{int}$, calculate the new $x$, and then recompute extraction and clearance. Also do not multiply by $f_u$ twice if the supplied intrinsic clearance is already defined as an effective clearance that incorporates binding.

The constant-$f_u$ assumption is a substantive limitation. Saturable binding can make the unbound fraction depend on concentration, and changes in albumin or other binding components can alter $f_u$ without changing enzyme capacity. Likewise, if metabolism depends on transport into hepatocytes, the unbound fraction in incoming blood alone may not capture the concentration available at the enzyme. Such cases require a model that represents the relevant binding, transport, and concentration relationships rather than an unqualified substitution in a clearance equation.

Check Your Understanding

Answer each question using the assumptions and equations in this tutorial. Brief solutions follow the questions.

If $CL_{int}=80$ L/h and $f_u=0.25$, what is the effective intrinsic clearance?
For the well-stirred model with $Q_H=100$ L/h and $x=0.20$, calculate $E_H$ and $CL_H$.
Why does doubling $f_u$ not necessarily double total hepatic clearance?
Name two assumptions required for the substitution $CL_{int}\mapsto f_uCL_{int}$.
A calculated extraction ratio is 1.08. Give one model-based diagnostic conclusion.

Solutions. (1) $f_uCL_{int}=0.25(80)=20$ L/h. (2) $E_H=0.20/(1+0.20)=0.1667$, so $CL_H=100(0.1667)=16.67$ L/h. (3) Extraction is a nonlinear function of $x$ in these models, and total clearance approaches the flow limit rather than increasing without bound. (4) For example, binding is at rapid equilibrium and $f_u$ is constant over the modeled concentration range; the model also assumes only unbound drug is available to enzymes. (5) Since the extraction ratio must lie between 0 and 1, recheck the formula, substitutions, units, and arithmetic.

Unbound-Fraction Effects in Hepatic Clearance Models

What You'll Learn