Protein Binding and Total Concentrations in Hepatic Dysfunction

Why Total Concentrations Can Mislead

Many assays report the total plasma concentration: drug that is both bound to plasma proteins and unbound. The unbound drug is the fraction not bound to proteins; it is often the fraction available to cross membranes and interact with targets. Albumin is an important binding protein for many drugs, and its concentration may be reduced in hepatic dysfunction. A lower albumin concentration can therefore change the relation between total and unbound concentrations, even when the unbound concentration itself changes little.

This distinction matters when interpreting a population PK model. A lower total concentration in a person with low albumin does not, by itself, show that the person has lower pharmacologically available exposure. It may instead reflect a larger unbound fraction. Conversely, a model that predicts total concentration well may still misrepresent unbound exposure if it does not account for binding.

Definition: Let \(C_{\mathrm{tot}}\) be the total plasma concentration and \(C_u\) the unbound plasma concentration. The unbound fraction is \(f_u=C_u/C_{\mathrm{tot}}\), so \(C_u=f_u C_{\mathrm{tot}}\) and \(C_{\mathrm{tot}}=C_u/f_u\). These relations apply when concentrations refer to the same sample and binding is adequately described by a single fraction.

The unbound fraction is dimensionless, while both concentrations have units such as mg/L. If binding is linear and \(f_u\) is constant over the concentration range of interest, the same conversion applies to exposure over time: \(AUC_u=f_u AUC_{\mathrm{tot}}\). This simple conversion is useful, but its assumptions should not be mistaken for a universal description of protein binding.

Albumin, Binding, and Concentration

For a drug that binds reversibly to albumin, lowering albumin can reduce the amount of binding capacity in plasma and increase \(f_u\). The direction and size of the change depend on the drug’s affinity for albumin, the albumin concentration, and the presence of other binding proteins or competing substances. Albumin is therefore a mechanistic clue, not a complete measure of binding or hepatic function.

At a fixed unbound concentration, an increase in \(f_u\) implies a decrease in total concentration because \(C_{\mathrm{tot}}=C_u/f_u\). At a fixed total concentration, an increase in \(f_u\) implies a greater unbound concentration. Neither statement alone predicts the concentration that will occur after a dose: distribution and elimination can also change the concentrations over time.

A key distinction is between a binding effect and a change in intrinsic elimination capacity. A change in albumin can alter the relationship between total and unbound drug. Hepatic dysfunction can also alter metabolic capacity, hepatic blood flow, or other determinants of clearance. These mechanisms may occur together, so an observed albumin association should not automatically be interpreted as an isolated binding effect.

Worked Example: Same Unbound Concentration, Different Total Concentrations

Suppose two hypothetical patients have the same unbound concentration of \(2.00\) mg/L. The first has \(f_u=0.0500\), and the second has \(f_u=0.100\). Calculate their total concentrations.

For the first patient, \(C_{\mathrm{tot}}=C_u/f_u\):

C_{\mathrm{tot}}=\frac{2.00\ \mathrm{mg/L}}{0.0500}=40.0\ \mathrm{mg/L}.

For the second patient:

C_{\mathrm{tot}}=\frac{2.00\ \mathrm{mg/L}}{0.100}=20.0\ \mathrm{mg/L}.

The second patient has half the total concentration but the same unbound concentration. A comparison based only on total concentration would describe a difference that is not present in the stated unbound concentrations. This calculation demonstrates the binding conversion; it does not establish that the two patients would have the same unbound concentration in a clinical setting.

Clearance and the Unbound Fraction

For an intravenously administered drug with linear PK, total clearance is commonly calculated as dose divided by total AUC. Unbound clearance is defined analogously using unbound AUC. If \(f_u\) is constant over the exposure interval, these quantities are related as follows:

CL_{\mathrm{tot}}=\frac{\mathrm{Dose}_{\mathrm{IV}}}{AUC_{\mathrm{tot}}},\qquad CL_u=\frac{\mathrm{Dose}_{\mathrm{IV}}}{AUC_u} =\frac{CL_{\mathrm{tot}}}{f_u}.

Thus, a change in total clearance does not necessarily imply an equal change in unbound clearance. The conversion depends on \(f_u\), and a total-based clearance is defined relative to total concentration. These definitions are for systemic intravenous dose; interpreting oral apparent clearance also requires accounting for bioavailability.

A useful mechanistic illustration is the well-stirred liver model. Let \(Q_h\) denote hepatic blood flow and \(CL_{\mathrm{int}}\) intrinsic hepatic clearance with respect to unbound drug. Under the model’s assumptions, hepatic clearance referenced to total plasma concentration is:

CL_h=\frac{Q_h f_u CL_{\mathrm{int}}}{Q_h+f_u CL_{\mathrm{int}}}.

The model assumes a well-mixed liver, linear elimination, and that unbound drug is available for hepatic elimination; it also treats \(Q_h\), \(f_u\), and \(CL_{\mathrm{int}}\) as defined quantities for the condition being modeled. Define the extraction ratio as \(E_h=CL_h/Q_h\). When \(f_u CL_{\mathrm{int}}\) is small relative to \(Q_h\), the drug has low extraction and \(CL_h\) is approximately \(f_u CL_{\mathrm{int}}\). When \(f_u CL_{\mathrm{int}}\) is large relative to \(Q_h\), clearance approaches \(Q_h\), a high-extraction limit.

This gives a reason that albumin changes can have different total-concentration consequences for different drugs. For a low-extraction drug, a higher \(f_u\) can raise total hepatic clearance if \(CL_{\mathrm{int}}\) is unchanged. For a high-extraction drug, clearance is more constrained by hepatic blood flow in this model. These are conditional model implications, not a claim that hepatic dysfunction changes albumin alone or that every drug follows the same limiting case.

Worked Example: Total and Unbound AUC

Consider hypothetical intravenous data for a drug with constant \(f_u\) during each exposure interval. In a reference condition, \(AUC_{\mathrm{tot}}=1.00\times10^3\) mg·h/L and \(f_u=0.0500\). In a second condition, \(AUC_{\mathrm{tot}}=501\) mg·h/L and \(f_u=0.100\). Calculate and compare unbound AUC.

Using \(AUC_u=f_u AUC_{\mathrm{tot}}\), the reference-condition calculation is:

AUC_u=0.0500(1.00\times10^3)=50.0\ \mathrm{mg\cdot h/L}.

For the second condition:

AUC_u=0.100(501)=50.1\ \mathrm{mg\cdot h/L}.

The displayed values give unbound AUCs of 50.0 and 50.1 mg·h/L, respectively; they are nearly equal at the precision shown. Total AUC is about half as large in the second condition, while its unbound fraction is twice as large. This example illustrates why the total AUC alone can give a different impression from unbound AUC. It assumes constant \(f_u\) in each condition and does not show that the conditions differ only in albumin.

Modeling Total and Unbound Concentrations

A population PK analysis can use total concentrations as the dependent observations, or it can model unbound concentrations. The choice determines what the model’s parameters and covariate effects directly describe. In either case, the concentration type, units, assay, and any conversion between total and unbound values must be explicit.

Modeling choice	Directly modeled observation	Interpretation and main caution
Total-concentration model	\(C_{\mathrm{tot}}\)	Predicts measured total concentration; an albumin effect may combine binding and disposition effects.
Unbound-concentration model	\(C_u\), measured or derived	Targets unbound exposure more directly; derived values depend on a valid estimate of \(f_u\).

In a total-concentration model, albumin may be included as a covariate on a PK parameter such as clearance or volume. The resulting coefficient describes how the modeled total-concentration profile changes with albumin, conditional on the model and its other covariates. It is not automatically a direct estimate of the effect of albumin on binding, nor does it necessarily indicate a change in intrinsic metabolic capacity.

An unbound-concentration model can represent concentrations directly when unbound measurements are available. Alternatively, a model may predict total concentration and convert it using a binding model, for example by making \(f_u\) depend on albumin. That strategy separates the binding relationship from disposition only to the extent that the data and model can identify the two components. If unbound measurements are sparse or absent, different combinations of binding and clearance effects may explain similar total-concentration data.

An important comparison is therefore not simply which model fits better. A total model can be appropriate when the clinical decision concerns measured total concentrations and the model is validated for that purpose. An unbound model is useful when unbound exposure is the intended target and the measurements or conversion assumptions support it. If binding is stable and the clinical target is total concentration, modeling unbound values may add complexity without answering the main question. If binding varies substantially, relying only on total concentrations can obscure relevant changes in unbound exposure.

Worked Example: A Low-Extraction Infusion Model

For illustration, assume a hypothetical drug follows the well-stirred model, has \(Q_h=100\) L/h and \(CL_{\mathrm{int}}=5.00\) L/h in both conditions, and is given by IV infusion at \(R_0=5.00\) mg/h. Let \(f_u\) change from 0.0500 to 0.100. Calculate hepatic clearance and steady-state total and unbound concentrations. Assume hepatic clearance is the only clearance and binding is linear and constant in each condition.

At steady state, \(C_{\mathrm{ss,tot}}=R_0/CL_h\), and \(C_{\mathrm{ss},u}=f_u C_{\mathrm{ss,tot}}\). For \(f_u=0.0500\):

CL_h=\frac{100(0.0500)(5.00)}{100+(0.0500)(5.00)} =\frac{25.0}{100.25}=0.2494\ \mathrm{L/h}.

Therefore:

C_{\mathrm{ss,tot}}=\frac{5.00\ \mathrm{mg/h}}{0.2494\ \mathrm{L/h}} =20.05\ \mathrm{mg/L}, \qquad C_{\mathrm{ss},u}=0.0500(20.05)=1.003\ \mathrm{mg/L}.

For \(f_u=0.100\):

CL_h=\frac{100(0.100)(5.00)}{100+(0.100)(5.00)} =\frac{50.0}{100.5}=0.4975\ \mathrm{L/h},

C_{\mathrm{ss,tot}}=\frac{5.00}{0.4975}=10.05\ \mathrm{mg/L}, \qquad C_{\mathrm{ss},u}=0.100(10.05)=1.005\ \mathrm{mg/L}.

Under these assumptions, doubling \(f_u\) approximately doubles total hepatic clearance and halves total steady-state concentration, while unbound steady-state concentration is nearly unchanged. The unbound values are not exactly identical because the model is not at the limiting approximation \(f_u CL_{\mathrm{int}}/Q_h=0\). This result is conditional on unchanged intrinsic clearance and flow; actual hepatic dysfunction could alter either.

Assumptions, Diagnostics, and Limitations

A binding-based interpretation should be checked against the data and the model’s assumptions. Linear conversion using a single \(f_u\) is most defensible when the unbound fraction is approximately constant over the observed concentration range and time interval. If binding is saturable, \(f_u\) can vary with concentration, and \(AUC_u=f_u AUC_{\mathrm{tot}}\) with one fixed fraction is not generally valid. If albumin is only a proxy for binding capacity, other binding proteins or competing substances may also matter.

Useful checks include comparing measured total and unbound concentrations from the same samples, examining whether the ratio \(C_u/C_{\mathrm{tot}}\) changes across concentration levels or albumin values, and checking whether derived \(f_u\) values are plausible and consistent over time. Inspect model residuals against albumin, concentration, time, and relevant patient groups. A systematic trend in residuals across albumin values can indicate a missing covariate relationship, misspecified binding, or another correlated difference. Residual patterns are diagnostic clues, not proof of a particular mechanism.

The estimates also depend on assay quality. Total and unbound concentrations may have different measurement error, and unbound measurements require a separation method that preserves the relevant binding equilibrium. Samples collected or processed under conditions that alter binding can yield a misleading \(f_u\). Sparse unbound sampling can limit the ability to distinguish a binding effect from changes in clearance or distribution.

The simple relationship \(C_u=f_u C_{\mathrm{tot}}\) concerns plasma concentrations. It does not establish that unbound plasma concentration equals concentration at the site of action or that it fully determines response. Distribution into tissues, transport, active metabolites, and pharmacodynamic sensitivity may also matter. Likewise, a change in total concentration may be clinically important if the dosing target or toxicity monitoring is defined using total concentration; interpretation must match the actual measurement and decision.

Common Pitfall

Wrong reasoning: “Total concentration is lower when albumin is low, so the patient has lower drug exposure and may need a higher dose.” This conclusion assumes that total concentration changes in the same direction and proportion as unbound exposure.

Correction: First establish whether the observation is total or unbound, and determine whether \(f_u\) changes. A lower total concentration can coexist with similar unbound concentration when \(f_u\) increases. Then assess whether clearance, distribution, or other aspects of hepatic dysfunction also changed. Do not infer a dose adjustment from albumin and total concentration alone.

Key takeaway: Total and unbound concentrations answer related but distinct questions. Albumin can change their relationship; the interpretation of a population PK covariate effect depends on which concentration was modeled and on whether binding, clearance, and their assumptions are adequately characterized.

Check Your Understanding

Answer each question, then compare with the brief solutions below.

If \(C_{\mathrm{tot}}=30.0\) mg/L and \(f_u=0.0800\), what is \(C_u\)?
With constant \(f_u=0.0400\), what is \(AUC_u\) when \(AUC_{\mathrm{tot}}=750\) mg·h/L?
Why can a total-concentration model’s albumin coefficient not automatically be called an effect on intrinsic clearance?
What assumption is needed to use one \(f_u\) to convert total AUC into unbound AUC?
In the well-stirred model, what clearance limit is expected when \(f_u CL_{\mathrm{int}}\) is much greater than \(Q_h\)?

Solutions. (1) \(C_u=f_u C_{\mathrm{tot}}=0.0800(30.0)=2.40\) mg/L. (2) \(AUC_u=0.0400(750)=30.0\) mg·h/L. (3) The coefficient describes a conditional association with modeled total concentrations and can reflect binding, disposition, or correlated changes; it does not isolate intrinsic clearance. (4) Binding must be adequately represented by a constant unbound fraction over the exposure interval, so \(AUC_u=f_u AUC_{\mathrm{tot}}\) holds. (5) \(CL_h\) approaches \(Q_h\), the hepatic blood-flow limit in the well-stirred model.