Estimating Fraction Escaping Gut Metabolism

From Enterocyte Exposure to Intestinal Availability

After a drug is absorbed from the intestinal lumen, some of it may be metabolized in enterocytes before reaching portal blood. The fraction escaping gut metabolism is therefore not determined by enzyme activity alone. It also depends on how much absorbed drug is available to the metabolically accessible intracellular pool, and on how metabolism competes with exit from that pool.

This tutorial derives a compact model for those processes. Its result is intestinal availability \(F_g\): the fraction of absorbed drug that reaches portal blood unchanged with respect to gut metabolism. It is conditional on absorption. It does not include the fraction of an oral dose that fails to dissolve or cross the intestinal wall, nor does it include hepatic first-pass extraction.

The model treats drug outside the accessible pool as bypassing metabolism in that pool. Within the pool, metabolism and exit are the only two competing fates. The effective metabolic clearance is \(f_{u,g}CL_{\mathrm{int},g}\), because the model assumes only unbound drug is available for metabolism. Both this quantity and \(CL_{\mathrm{exit},g}\) have units of volume per time.

Deriving the Accessible-Pool Extraction Fraction

Write \(M_g=f_{u,g}CL_{\mathrm{int},g}\) for effective metabolic clearance and \(X_g=CL_{\mathrm{exit},g}\) for exit clearance. Assume both clearances are nonnegative, linear in the relevant drug concentration, and constant over the interval being represented. Suppose drug enters the accessible pool at a constant rate and the pool is at steady state. If its concentration is \(C_g\), the rates of metabolism and exit are \(M_gC_g\) and \(X_gC_g\). With no other loss pathway, the input rate equals their sum.

When \(M_g+X_g>0\), the fraction of input leaving by metabolism is its metabolic outflow divided by total outflow. The corresponding fraction leaving unchanged by exit is the exit outflow divided by total outflow. Thus the fraction extracted within the accessible pool is

This rate-ratio result also follows from competing first-order removal: the instantaneous metabolic hazard is \(M_g/V_g\) and the exit hazard is \(X_g/V_g\) for pool volume \(V_g\), so the probability that metabolism occurs first is the metabolic hazard divided by the sum. The steady-state derivation makes explicit that the same ratio applies to the fraction of constant input assigned to each pathway.

Only the fraction \(a_g\) enters this pool. Consequently, the fraction of all absorbed drug metabolized in it is \(a_gE_g\); the remainder escapes metabolism. For \(X_g+M_g>0\), the central result is

The parameters have distinct roles. \(E_g\) describes extraction conditional on entering the accessible pool, whereas \(a_g\) determines how much absorbed drug is exposed to that extraction. For fixed positive exit clearance, increasing \(f_{u,g}CL_{\mathrm{int},g}\) increases \(E_g\) and lowers \(F_g\); increasing \(a_g\) also lowers \(F_g\). A large intrinsic clearance need not imply low intestinal availability if little drug is accessible to the enzyme.

Worked Calculations

Boundary Cases and a Short Sensitivity Result

Boundary cases matter because the extraction ratio is defined only when the accessible pool has at least one removal pathway. If \(CL_{\mathrm{int},g}=0\), \(f_{u,g}=0\), or both, then \(M_g=0\). When \(CL_{\mathrm{exit},g}>0\), \(E_g=0\), so \(F_g=1\) for every \(a_g\): there is no modeled metabolism. If \(M_g>0\) and \(CL_{\mathrm{exit},g}=0\), then \(E_g=1\), so \(F_g=1-a_g\). When both \(M_g=0\) and \(CL_{\mathrm{exit},g}=0\), the denominator is zero. The model does not determine the fate of drug in the accessible pool; it is not valid to assign that pool an extraction fraction of zero or one.

If \(a_g=0\), none of the absorbed drug enters the accessible pool. The overall model then gives \(F_g=1\), regardless of the pool's hypothetical fate. If both pool clearances are zero, that fate remains undefined, but no drug is assigned to the pool when \(a_g=0\).

For positive total clearance, let \(M_g=f_{u,g}CL_{\mathrm{int},g}\) and \(X_g=CL_{\mathrm{exit},g}\). The model gives \(F_g=1-a_gM_g/(X_g+M_g)\). Differentiating with respect to \(M_g\), with \(a_g\) and \(X_g\) held fixed, gives

For \(X_g>0\) and \(a_g>0\), this derivative is negative: increasing effective metabolic clearance lowers availability. If \(X_g=0\) and \(M_g>0\), all drug entering the pool is metabolized, and changing the magnitude of \(M_g\) does not change the limiting fraction \(E_g=1\). This distinction prevents a general claim that increasing enzyme clearance always lowers \(F_g\) in every boundary case.

Estimation, Diagnostics, and Limits

If \(F_g\) and both clearances are known, the equation can be rearranged to estimate intracellular availability only when the accessible-pool extraction fraction is positive. Specifically, if \(X_g+M_g>0\) and \(E_g>0\),

If \(E_g=0\), then \(F_g=1\) for any \(a_g\), provided \(CL_{\mathrm{exit},g}>0\). That observation cannot identify \(a_g\). Even for positive extraction, the estimate must satisfy \(0\leq a_g\leq1\). Equivalently, the model requires \(0\leq1-F_g\leq E_g\). A violation signals incompatible inputs, measurement error, or failure of the model assumptions; it should not be repaired by reporting an impossible availability.

Useful checks include confirming that both clearances use consistent units, that \(f_{u,g}\) and \(a_g\) are fractions between zero and one, and that predicted \(F_g\) lies between zero and one. A result below zero or above one usually indicates an algebra or input error. If enzyme activity changes with concentration, if pathways saturate, or if drug returns from the accessible pool after exit, constant competing clearances may not adequately represent the fate of drug. The model also compresses spatial variation along the intestine into one effective set of parameters.

This formulation is more informative than simply assigning an empirical \(F_g\), because it distinguishes enzyme extraction from intracellular access and makes their assumptions visible. A direct empirical \(F_g\) can be useful when data support only a net fraction, but it does not explain how enzyme activity or availability changes that fraction. A spatially resolved intestinal PBPK model can represent regional differences, but requires more inputs and parameters. This reduced model is useful when a compact mechanistic estimate is appropriate, not when its assumptions are known to fail.

Common Mistakes

• Equating pool extraction with whole absorbed-drug extraction. \(E_g\) applies only to drug entering the accessible pool. Overall metabolism is \(a_gE_g\), not \(E_g\), unless \(a_g=1\).
• Treating \(F_g\) as total oral bioavailability. It excludes incomplete absorption and hepatic first-pass loss.
• Setting \(F_g=1\) whenever metabolic clearance is zero without checking exit clearance. This conclusion holds when \(CL_{\mathrm{exit},g}>0\). If both pool clearances are zero, the accessible pool has no modeled removal route and its fate is undetermined.
• Solving for \(a_g\) by dividing by zero extraction. When \(E_g=0\), \(F_g=1\) regardless of \(a_g\), so that observation does not identify intracellular availability.