Physiologic Hepatic Clearance Models

Why Relate Hepatic Clearance to Physiology?

A clearance value summarizes how rapidly drug is removed relative to its concentration, but it does not by itself explain which physiological factors govern that removal. For hepatic elimination, the well-stirred model connects three quantities: hepatic blood flow, the fraction of drug unbound in blood, and the liver’s intrinsic ability to eliminate unbound drug. The connection is useful for interpreting why a change in blood flow may matter greatly for one drug but little for another.

The model is a structural approximation, not a universal description of every liver. It treats the liver as a single, uniformly mixed compartment in which drug entering in blood is rapidly mixed before leaving. Its predictions are meaningful only when its assumptions and concentration basis are appropriate.

Definition: In the well-stirred liver model, \(Q_H\) is hepatic blood flow (volume per time), \(f_u\) is the unbound fraction of drug in hepatic blood (dimensionless), and \(CL_{int}\) is intrinsic hepatic clearance referenced to the unbound drug concentration (volume per time). The model predicts hepatic blood clearance \(CL_H\) and the hepatic extraction ratio \(E_H\), the fraction removed from blood during one passage through the liver.

The model assumes steady flow and a well-mixed liver, with the drug concentration in the liver equal to the hepatic venous outflow concentration. Only unbound drug is available for elimination, and its elimination rate is proportional to its unbound concentration, with proportionality \(CL_{int}\). The model also assumes that the relevant flow, binding fraction, and intrinsic clearance are constant over the conditions being modeled. These assumptions lead to a linear relationship; this tutorial uses that relationship rather than describing concentration-dependent elimination.

Concentrations and flow must be defined consistently. Here \(Q_H\) is blood flow and the inlet and outlet concentrations are blood concentrations; \(f_u\) is the unbound fraction in that same blood. If measurements are reported in plasma, a blood-to-plasma relationship may be needed before applying the equations. Substituting plasma clearance or plasma unbound fraction without checking the concentration basis can make the calculation internally inconsistent.

Deriving Clearance and Extraction

Let \(C_{in}\) be the total drug concentration in blood entering the liver and \(C_{out}\) the total concentration leaving it. The amount delivered to the liver per unit time is \(Q_H C_{in}\), and the amount leaving in blood per unit time is \(Q_H C_{out}\). Their difference is the elimination rate. Under the well-stirred assumption, the unbound concentration available for elimination is \(f_u C_{out}\), so intrinsic clearance gives an elimination rate of \(CL_{int} f_u C_{out}\).

Q_H(C_{in}-C_{out})=CL_{int}f_u C_{out}

Rearranging the flow balance gives the outflow-to-inflow concentration ratio. The extraction ratio is one minus that ratio. Hepatic clearance is the elimination rate divided by the inlet concentration. These definitions yield the central relationships:

\frac{C_{out}}{C_{in}}=\frac{Q_H}{Q_H+f_uCL_{int}}, \qquad E_H=\frac{C_{in}-C_{out}}{C_{in}} =\frac{f_uCL_{int}}{Q_H+f_uCL_{int}}, \qquad CL_H=Q_HE_H =\frac{Q_Hf_uCL_{int}}{Q_H+f_uCL_{int}}.

The units provide a useful check. The product \(f_uCL_{int}\) has units of volume per time, as do \(Q_H\) and \(CL_H\); \(E_H\) is dimensionless. For positive \(Q_H\), \(f_u\), and \(CL_{int}\), the extraction ratio lies between zero and one, and \(CL_H\) is less than \(Q_H\). The model cannot predict more drug extracted per unit time than is delivered to the liver.

Worked Example: Calculating Clearance and Extraction

Suppose a hypothetical drug has \(Q_H=90\) L/h, \(f_u=0.10\), and \(CL_{int}=300\) L/h. Calculate its hepatic clearance, extraction ratio, and outflow concentration as a fraction of the inflow concentration.

First calculate the unbound intrinsic-clearance term: \(f_uCL_{int}=0.10(300)=30\) L/h. Then

CL_H=\frac{90(30)}{90+30} =\frac{2700}{120}=22.50\ {\rm L/h}.

Calculate extraction directly from the model terms: \(E_H=30/(90+30)=30/120=0.2500\). The predicted outflow-to-inflow concentration ratio is \(1-E_H=0.7500\), so \(C_{out}=0.7500C_{in}\). For example, if the inlet blood concentration were 4.00 mg/L, the modeled outlet concentration would be \(0.7500(4.00)=3.00\) mg/L. The extraction and clearance are consistent because \(Q_HE_H=90(0.2500)=22.50\) L/h.

Interpreting Low and High Extraction

The product \(f_uCL_{int}\) is the effective intrinsic elimination capacity in the model. When it is small relative to hepatic blood flow, extraction is low and clearance is approximately that product. When it is large relative to flow, extraction is high and clearance approaches hepatic blood flow. These are limiting approximations, not separate equations that replace the full model in intermediate cases.

f_uCL_{int}\ll Q_H: \quad CL_H\approx f_uCL_{int},\quad E_H\approx\frac{f_uCL_{int}}{Q_H}; \qquad f_uCL_{int}\gg Q_H: \quad CL_H\approx Q_H,\quad E_H\approx 1.

The approximations follow by simplifying the denominator \(Q_H+f_uCL_{int}\). In the low-extraction case it is close to \(Q_H\), so clearance is approximately \(Q_Hf_uCL_{int}/Q_H=f_uCL_{int}\). In the high-extraction case it is close to \(f_uCL_{int}\), so clearance is approximately \(Q_H\). Thus low-extraction clearance is especially responsive to unbound fraction and intrinsic clearance, while high-extraction clearance is especially responsive to hepatic blood flow.

Worked Example: A High-Extraction Drug

For a different hypothetical drug, let \(Q_H=90\) L/h, \(f_u=0.80\), and \(CL_{int}=600\) L/h. The effective intrinsic term is \(0.80(600)=480\) L/h, which is much larger than the flow.

The exact model calculation is

CL_H=\frac{90(480)}{90+480} =\frac{43200}{570}=75.79\ {\rm L/h}.

Using the unrounded model terms, \(E_H=480/570=0.8421\), and \(C_{out}/C_{in}=90/570=0.1579\). Clearance is below the flow of 90 L/h, as required, but approaches it because extraction is high. The high-extraction approximation would give \(CL_H\approx90\) L/h; it indicates the governing influence, but it is not as accurate as the exact value of 75.79 L/h here.

Parameter sensitivity can be stated more precisely. Write \(x=f_uCL_{int}\), so \(CL_H=Q_Hx/(Q_H+x)\). Differentiating with respect to \(x\) and \(Q_H\), then expressing the results as proportional sensitivities, gives:

\frac{\partial CL_H}{\partial x} =\frac{Q_H^2}{(Q_H+x)^2}, \qquad \frac{\partial CL_H}{\partial Q_H} =\frac{x^2}{(Q_H+x)^2}; \qquad \frac{\partial\ln CL_H}{\partial\ln x} =\frac{Q_H}{Q_H+x}=1-E_H, \qquad \frac{\partial\ln CL_H}{\partial\ln Q_H} =\frac{x}{Q_H+x}=E_H.

For example, the last two expressions show that, locally, a 1% increase in \(x\) corresponds to about a \((1-E_H)\)% increase in clearance, while a 1% increase in flow corresponds to about an \(E_H\)% increase. This is a local sensitivity statement, not a guarantee that a large parameter change produces an exactly proportional clearance change.

Worked Example: Low Extraction and a Change in Unbound Fraction

Suppose \(Q_H=90\) L/h, \(CL_{int}=40\) L/h, and initially \(f_u=0.05\). The effective intrinsic term is \(0.05(40)=2.00\) L/h, much smaller than flow.

The initial clearance is \(CL_H=90(2)/(90+2)=180/92=1.957\) L/h, and extraction is \(E_H=2/92=0.02174\). Now suppose \(f_u\) doubles to 0.10 while flow and intrinsic clearance remain fixed. The new effective term is \(0.10(40)=4.00\) L/h, giving \(CL_H=90(4)/(90+4)=360/94=3.830\) L/h and \(E_H=4/94=0.04255\).

The unbound fraction doubled, and clearance increased by a factor of \(3.830/1.957\approx1.958\), close to but not exactly two. In the low-extraction range, clearance is approximately proportional to \(f_uCL_{int}\); the exact model retains the flow term in the denominator.

Inferring the Intrinsic Term and Recognizing Limits

If hepatic clearance and flow are known, the model can be rearranged to estimate the product \(f_uCL_{int}\). Since \(E_H=CL_H/Q_H\), solving \(E_H=x/(Q_H+x)\) for \(x\) gives \(x=Q_HE_H/(1-E_H)\). This identifies the product, not necessarily its two components separately. Independent information about \(f_u\) is needed to infer \(CL_{int}\) from that product.

Worked Example: Inferring Intrinsic Clearance

Assume a hypothetical hepatic clearance of 60.0 L/h and hepatic blood flow of 100 L/h, both defined on the model’s blood basis. The extraction ratio is \(E_H=60.0/100=0.600\). Therefore,

f_uCL_{int} =\frac{Q_HE_H}{1-E_H} =\frac{100(0.600)}{1-0.600} =\frac{60.0}{0.400} =150\ {\rm L/h}.

If an independently measured unbound fraction were \(f_u=0.25\), then \(CL_{int}=150/0.25=600\) L/h. Without that binding information, the clearance and flow determine only the product \(f_uCL_{int}=150\) L/h; many combinations of unbound fraction and intrinsic clearance could give the same product.

The well-stirred model is not the only physiological representation. A parallel-tube model treats blood as flowing through tubes with concentration declining along each tube rather than mixing uniformly throughout the liver. Under its assumptions, extraction is \(1-\exp(-f_uCL_{int}/Q_H)\). It shares the same low-extraction approximation but can predict different extraction at higher values of \(f_uCL_{int}/Q_H\). Choosing between these structures requires evidence and a suitable model for the system; one should not switch equations merely to obtain a preferred clearance.

The well-stirred model is limited when the liver is not adequately represented as a single mixed unit, when relevant binding or flow is mismeasured, or when the assumed constant linear intrinsic clearance does not describe the conditions. It also estimates a hepatic component: clearance from other organs or pathways is not included in \(CL_H\). In population modeling, a fitted clearance may summarize observed concentration data without uniquely identifying \(Q_H\), \(f_u\), and \(CL_{int}\). The model provides a physiological relationship, but parameter identification still depends on what the data and external measurements can support.

Diagnostics and Common Mistakes

Before interpreting a calculation, check that all clearances and flow use compatible volume and time units, that the fraction \(f_u\) is between zero and one, and that concentrations refer to a consistent blood basis. A predicted \(CL_H\) greater than \(Q_H\), or an extraction ratio outside zero to one for positive model inputs, signals an arithmetic, unit, or model-definition problem. A very high predicted extraction paired with observed evidence of little removal may indicate that one or more assumptions or input values are unsuitable.

Using \(CL_{int}\) alone in the equation. The effective term is \(f_uCL_{int}\), because only unbound drug is available for elimination in this model.
Calling \(CL_H\) the extraction ratio. Clearance has units of volume per time; extraction is dimensionless, with \(CL_H=Q_HE_H\).
Using plasma values without checking the basis. Flow is blood flow in the stated model, so plasma concentrations or clearances must be reconciled with blood concentrations before combination.
Assuming every hepatic drug is flow-limited. That approximation applies when \(f_uCL_{int}\) is large relative to \(Q_H\); at low extraction, clearance is more closely related to \(f_uCL_{int}\).

A common mistake is to infer that any increase in intrinsic clearance must produce the same percentage increase in hepatic clearance. The full equation shows why this is wrong: as \(f_uCL_{int}\) becomes large relative to flow, clearance approaches \(Q_H\), so further increases in intrinsic clearance have progressively less effect. The correction is to calculate or inspect the ratio \(f_uCL_{int}/Q_H\) and use the full well-stirred equation unless a limiting approximation is justified.

Key takeaway: The well-stirred model predicts \(CL_H=Q_Hf_uCL_{int}/(Q_H+f_uCL_{int})\) and \(E_H=f_uCL_{int}/(Q_H+f_uCL_{int})\). The relative size of \(f_uCL_{int}\) and \(Q_H\) determines whether clearance is more sensitive to intrinsic capacity and unbound fraction or to hepatic blood flow.

Check Your Understanding

Answer each question, then compare your reasoning with the solutions.

For \(Q_H=80\) L/h, \(f_u=0.20\), and \(CL_{int}=100\) L/h, calculate \(CL_H\) and \(E_H\).
If \(f_uCL_{int}\) is much smaller than \(Q_H\), which quantity is hepatic clearance approximately proportional to?
A model calculation gives \(CL_H=110\) L/h when \(Q_H=100\) L/h, with positive \(f_u\) and \(CL_{int}\). What should be checked?
If \(CL_H=40\) L/h and \(Q_H=100\) L/h, calculate \(E_H\) and \(f_uCL_{int}\).
Why can hepatic clearance and flow alone fail to identify \(CL_{int}\)?

Solutions. (1) \(f_uCL_{int}=0.20(100)=20\) L/h, so \(CL_H=80(20)/(80+20)=16.0\) L/h and \(E_H=20/100=0.2000\). (2) It is approximately proportional to \(f_uCL_{int}\), because the denominator is then approximately \(Q_H\). (3) Check arithmetic, units, and whether flow and clearance use the same blood basis; the well-stirred model with positive inputs requires \(CL_H<Q_H\). (4) \(E_H=40/100=0.4000\), so \(f_uCL_{int}=100(0.4000)/(1-0.4000)=66.67\) L/h. (5) Clearance and flow identify the product \(f_uCL_{int}\); without independent information on \(f_u\), intrinsic clearance cannot be separated from unbound fraction.