Dose-Dependent Oral Bioavailability

Why Bioavailability May Change With Dose

A structural model may predict dose-proportional exposure when the fraction of an oral dose reaching systemic circulation is constant and elimination is linear. But the fraction absorbed and escaping presystemic loss need not remain constant as dose increases. A limited capacity for absorption can make the fraction reaching circulation smaller at higher doses; saturation of first-pass loss can make it larger. Either pattern changes exposure even if systemic clearance is unchanged.

Here, dose-dependent bioavailability is a way to represent how the systemic fraction varies with the administered oral dose. The equations below describe the total fraction reaching circulation; they do not, by themselves, specify the time course of absorption. The absorption structures introduced in earlier material can still describe that time course.

The main exposure result used here requires linear, time-invariant systemic elimination with clearance \(CL\), and a complete concentration-time profile after the dose. Under those assumptions, the total area under the concentration-time curve is \(AUC_{\infty}=F(D)D/CL\). Thus, \(AUC_{\infty}/D=F(D)/CL\): dose-normalized AUC is constant only when both bioavailability and clearance are constant. AUC alone cannot distinguish a change in \(F(D)\) from a change in \(CL\).

Two Bounded Ways to Represent Dose Dependence

For saturating first-pass loss, a convenient empirical model lets bioavailability rise from a lower value toward an upper value:

Here \(F_{\mathrm{low}}\) and \(F_{\mathrm{high}}\) are fractions, with \(0 \leq F_{\mathrm{low}} \leq F_{\mathrm{high}} \leq 1\), and \(K_{\mathrm{FP}}>0\) has dose units. As \(D\) approaches zero, \(F_{\mathrm{FP}}(D)\) approaches \(F_{\mathrm{low}}\); as \(D\) becomes large, it approaches \(F_{\mathrm{high}}\). This is a bounded description of an increasing fraction escaping presystemic loss. Its parameters should not automatically be interpreted as direct measurements of a particular enzyme or organ process.

For a fixed absorption capacity, a useful simple representation is a declining fraction with a finite systemic-amount limit:

Here \(0 \leq F_0 \leq 1\) is the low-dose limiting fraction, and \(K_{\mathrm{A}}>0\) has dose units. The systemic amount is \(F_0K_{\mathrm{A}}D/(K_{\mathrm{A}}+D)\), which approaches the finite limit \(F_0K_{\mathrm{A}}\) as dose increases. This is a compact representation of a capacity limit, not a complete time-course model of the absorption process. A mechanistic saturable-absorption structure is a separate modeling choice.

Both equations stay within plausible fraction bounds for nonnegative doses and the parameter constraints stated above. They encode opposite trends: the first predicts that the systemic fraction increases with dose, while the second predicts that it decreases. The choice should follow the scientific hypothesis and the observed dose-exposure pattern, not just which curve fits a small number of observations.

How the Model Changes Exposure

With linear elimination and \(CL\) constant across doses, dose proportionality means that multiplying \(D\) by a factor multiplies \(AUC_{\infty}\) by the same factor. A changing \(F(D)\) changes that relationship. The following derivative makes the role of bioavailability explicit. It is the elasticity of AUC with respect to dose: the local percentage change in AUC for a one-percent change in dose.

To obtain the second equality, take logarithms of the first: \(\log AUC_{\infty}=\log D+\log F(D)-\log CL\). Differentiating with respect to \(\log D\) gives 1 from \(\log D\), \(d\log F(D)/d\log D\) from the dose-dependent fraction, and zero from the constant \(CL\). Therefore, increasing bioavailability makes the local AUC response more than dose proportional; declining bioavailability makes it less than dose proportional.

For the capacity model, substituting \(F_{\mathrm{cap}}(D)\) gives \(AUC_{\infty}=F_0K_{\mathrm{A}}D/[CL(K_{\mathrm{A}}+D)]\). Its elasticity is \(K_{\mathrm{A}}/(K_{\mathrm{A}}+D)\), which is positive but below 1 for \(D>0\). At high doses, exposure approaches a plateau under this model. For the first-pass model, the fraction increases with dose, so the elasticity exceeds 1 wherever the fraction is increasing. These conclusions concern AUC under the stated assumptions; a change in total bioavailability does not alone determine a change in peak concentration or the shape of the concentration-time profile.

Using Dose-Normalized AUC to Assess the Pattern

A dose-normalized AUC comparison is useful when complete AUC profiles are available at more than one dose. Under linear, dose-invariant clearance, the ratio of dose-normalized AUCs equals the ratio of the corresponding bioavailability fractions. This can reveal a trend without identifying the mechanism that caused it. If clearance is known independently, the relation \(F(D)=AUC_{\infty}CL/D\) also gives an estimate of the fraction at each dose.

Assumptions, Diagnostics, and Model Choice

Before interpreting dose-related exposure as changing oral bioavailability, check the conditions that make that interpretation meaningful. In particular, the simple AUC relationship assumes linear, time-invariant systemic elimination; comparable formulation, food conditions, and dose administration; and AUC measured over a sufficiently complete interval. When elimination is saturable, AUC need not equal systemic amount divided by a constant clearance. A dose-related AUC change can then reflect elimination nonlinearity, so \(F(D)\) cannot be inferred from AUC using a single \(CL\).

A useful diagnostic sequence is to examine dose-normalized AUC, check the shape and completeness of the profiles, and compare plausible explanations. If dose-normalized AUC is approximately constant, constant bioavailability may be adequate over the tested range. If it rises or falls systematically, a dose-dependent fraction is one candidate. Consider whether independent information supports altered presystemic loss or limited absorption, and whether the elimination model is appropriate. Concentration-time profiles can also show whether only total exposure changes or whether the timing and shape of absorption change too.

Interpretation has important limits. Oral data alone generally identify the exposure relationship, not its unique physiological cause. Without an independent clearance estimate or suitable reference data, \(F(D)\) and \(CL\) are confounded in the AUC equation. A simple dose-dependent fraction may be a useful empirical structural component, but it does not identify the site of loss, explain variability among people, or describe how the input rate changes over time. Estimates near the bounds of zero or one, poorly determined asymptotes, and strong dependence on a few dose levels are reasons to scrutinize the model rather than to infer a precise mechanism.

The main alternatives answer different questions. A constant-\(F\) model is simpler and should be preferred when the observed dose range provides no clear evidence of a trend. The increasing first-pass model and the decreasing capacity model encode different directional hypotheses. A model of saturable absorption instead represents a change in the absorption process itself and may be needed when the time course of input is central. These structures should not be treated as interchangeable: dose-normalized AUC can signal nonproportional exposure, but additional evidence is needed to decide which process explains it.

Common mistake: “AUC rose more than the dose, so absorption must have increased.” That conclusion is too strong. AUC depends on both systemic amount and elimination, and a rising dose-normalized AUC is not specific to absorption or first-pass loss. The correction is to state the result first—exposure was more than dose proportional—and then assess whether clearance is constant and whether the data support a dose-dependent bioavailability explanation.