Worked Comparison of Kp Methods for Ionizable Drugs

Why Compare Methods for Ionizable Drugs?

A tissue-to-plasma partition coefficient, \(K_p\), summarizes the ratio of total drug concentration in a tissue to total concentration in plasma at equilibrium. For an ionizable drug, that ratio depends not only on how the neutral form partitions into tissue constituents, but also on the pH-dependent balance between neutral and ionized forms. Binding can further separate total concentrations from freely dissolved concentrations.

This tutorial compares two reduced, composition-based calculation routes. One weights aqueous and lipid contributions using tissue composition; the other starts with an unbound pH-partition ratio and applies a fraction-unbound correction. These forms make the assumptions visible and allow transparent hand calculations. They are teaching implementations, not substitutes for every detail of a full software implementation of the Schmitt or Berezhkovskiy approaches. The earlier material on those methods establishes that predictions can differ because methods encode composition and binding differently.

Definition: \(K_p=C_{\text{tissue}}/C_{\text{plasma}}\), where both concentrations are total concentrations expressed on a consistent basis at equilibrium. The unbound partition ratio is \(K_{p,uu}=C_{u,\text{tissue}}/C_{u,\text{plasma}}\), where each unbound concentration refers to freely dissolved drug. When the fraction unbound in tissue and plasma are defined consistently, \(K_p=K_{p,uu}(f_{u,p}/f_{u,t})\).

The calculations below assume passive distribution at equilibrium, a single relevant pKa, fixed tissue and plasma pH, linear nonsaturable binding, and no active transport, metabolism during distribution, or subcellular trapping beyond what is explicitly represented. Fractions unbound and tissue-composition terms are hypothetical inputs. Real tissues can violate these assumptions; the values here illustrate the calculations rather than represent a measured compound or organ.

Two Calculation Routes

For the first route, use a reduced composition-weighted expression. Let \(\phi_w\) and \(\phi_l\) be tissue volume fractions of aqueous and lipid phases, with \(\phi_w+\phi_l=1\). Let \(R_w\) be the ratio of total freely dissolved aqueous concentration in tissue water to the corresponding plasma unbound concentration, and let \(P_l\) be the lipid-phase drug concentration per plasma unbound concentration. Then the reduced composition-weighted total partition estimate is

K_p^{(\mathrm{comp})} = f_{u,p}\left(\phi_w R_w+\phi_l P_l\right).

Here \(R_w\) includes the pH-dependent aqueous ionization ratio, while \(P_l\) represents the lipid contribution under the stated simplified phase model. This expression treats the phase contributions as additive and assumes that their volumes and concentration ratios are defined on compatible bases. It does not separately model specific protein binding or subcellular compartments.

For the second route, first calculate \(K_{p,uu}\) from aqueous pH partitioning, then correct for the fractions unbound:

K_p^{(\mathrm{u\ correction})} = K_{p,uu}\frac{f_{u,p}}{f_{u,t}}.

In this route, the tissue fraction unbound must represent the tissue binding assumption being used. Composition-based methods may estimate that binding from tissue constituents. The correction is valid when the unbound concentrations and fractions unbound are consistently defined and binding is linear over the concentration range considered. It is not an extra pH correction: pH effects are already in \(K_{p,uu}\).

Ionization result: For a monoprotic weak acid \(HA \rightleftharpoons H^+ + A^-\), the total-to-neutral concentration factor in a compartment at pH \(pH\) is \(1+10^{pH-pK_a}\). For a monoprotic weak base \(B+H^+ \rightleftharpoons BH^+\), the corresponding factor is \(1+10^{pK_a-pH}\). Under passive equilibrium of the neutral form, the aqueous unbound ratio is the tissue factor divided by the plasma factor.

This result assumes the neutral form equilibrates between aqueous compartments, the pH values are fixed, and the ionized form does not independently cross the relevant membrane. It is an aqueous equilibrium relation, not a complete tissue model.

Worked Calculations

Worked Example: Weak Acid With an Unbound-Ratio Correction

Consider a hypothetical weak acid with \(pK_a=4.5\), tissue pH \(7.0\), and plasma pH \(7.4\). Suppose \(f_{u,p}=0.050\) and a composition-based tissue binding estimate gives \(f_{u,t}=0.100\). First calculate the aqueous unbound ratio, then the total \(K_p\).

For an acid, the relevant factor is \(1+10^{pH-pK_a}\). Therefore,

K_{p,uu} = \frac{1+10^{7.0-4.5}}{1+10^{7.4-4.5}} = \frac{1+316.2278}{1+794.3282} = \frac{317.2278}{795.3282} \approx 0.398864.

The total partition estimate is

K_p = 0.398864\left(\frac{0.050}{0.100}\right) = 0.199432 \approx 0.1994.

Under these assumptions, the predicted total tissue concentration is about 0.1994 times the total plasma concentration at equilibrium. The pH gradient gives a tissue-to-plasma unbound ratio below one; the assumed weaker tissue binding, represented by \(f_{u,t}>f_{u,p}\), reduces the total-ratio estimate further. This result is not a probability or a measured value; it is conditional on the stated model and inputs.

Worked Example: Weak Base With an Unbound-Ratio Correction

For a hypothetical weak base, let \(pK_a=8.0\), tissue pH \(5.5\), and plasma pH \(7.4\). Let \(f_{u,p}=0.080\) and \(f_{u,t}=0.020\). Assume neutral-form equilibrium and fixed pH. Calculate \(K_{p,uu}\) and \(K_p\).

For a base, the total-to-neutral factor is \(1+10^{pK_a-pH}\), so

K_{p,uu} = \frac{1+10^{8.0-5.5}}{1+10^{8.0-7.4}} = \frac{1+316.227766}{1+3.981072} = \frac{317.227766}{4.981072} \approx 63.68665.

Apply the binding correction:

K_p = 63.68665\left(\frac{0.080}{0.020}\right) = 63.68665(4) = 254.7466 \approx 254.7.

The low tissue pH favors the protonated form of this weak base, yielding a large predicted aqueous unbound ratio. The assumed tissue binding is also stronger than plasma binding, so \(f_{u,p}/f_{u,t}=4\) increases the total \(K_p\) relative to \(K_{p,uu}\). These two effects multiply in this calculation; neither should be confused with the other.

Worked Example: Weak Acid by a Composition-Weighted Route

Use the hypothetical acid from the first example. Suppose a simplified tissue has \(\phi_w=0.75\) and \(\phi_l=0.25\). Use its aqueous ratio \(R_w=0.398864\), a hypothetical lipid-phase ratio \(P_l=2.00\), and \(f_{u,p}=0.050\). Calculate the reduced composition-weighted estimate.

The phase fractions sum to one, as required by this two-phase illustration. Substitute into the composition expression:

K_p^{(\mathrm{comp})} = 0.050\left[0.75(0.398864)+0.25(2.00)\right].

The weighted terms are \(0.75(0.398864)=0.299148\) and \(0.25(2.00)=0.500000\). Thus,

K_p^{(\mathrm{comp})} = 0.050(0.799148) = 0.0399574 \approx 0.03996.

In this reduced model the lipid term dominates the bracketed phase contribution, despite its smaller volume fraction. This \(K_p\) differs from the unbound-correction result of 0.1994 because the calculation routes do not use the same tissue binding representation: the composition-weighted example uses a specified lipid contribution, whereas the correction example uses a specified \(f_{u,t}\). The difference is meaningful as a diagnostic of assumptions, not proof that either estimate is correct.

Why the Results Can Differ

A short derivation explains the fraction-unbound correction. By definition, \(C_{u,p}=f_{u,p}C_p\) and \(C_{u,t}=f_{u,t}C_t\). Hence \(C_t=C_{u,t}/f_{u,t}\) and \(C_p=C_{u,p}/f_{u,p}\). Taking their ratio gives

K_p = \frac{C_t}{C_p} = \frac{C_{u,t}/f_{u,t}}{C_{u,p}/f_{u,p}} = \frac{C_{u,t}}{C_{u,p}}\frac{f_{u,p}}{f_{u,t}} = K_{p,uu}\frac{f_{u,p}}{f_{u,t}}.

The derivation requires fractions unbound that relate the same total and unbound concentration definitions used in \(K_p\) and \(K_{p,uu}\). If a tissue assay, homogenate correction, or model uses a different basis, direct substitution can be invalid.

Composition-weighted methods make the tissue constituents explicit: a larger lipid fraction or a larger lipid affinity can increase the predicted contribution, while the aqueous contribution reflects ionization and pH. The unbound-correction route instead expresses the total ratio through \(K_{p,uu}\) and a tissue-binding term. When both routes use internally consistent versions of the same physical assumptions, they should be compatible. In practical applications, however, composition estimates, pH assumptions, and binding inputs may differ, so numerical agreement is not guaranteed.

Feature	Composition-weighted route	Unbound-ratio correction
Main inputs	Tissue phase fractions, phase partition terms, plasma fraction unbound	Aqueous \(K_{p,uu}\), plasma and tissue fractions unbound
Ionization	Enters the aqueous contribution through pH-dependent ratios	Enters directly through the pH-partition calculation
Binding representation	Implicit or explicit in the chosen constituent terms	Explicit through \(f_{u,p}/f_{u,t}\)
Useful diagnostic	Check phase fractions and the magnitude of each contribution	Check the definitions and plausibility of both fractions unbound

Diagnostics, Limits, and a Common Pitfall

Before trusting either calculation, check that pH and pKa are on the same scale, all fractions are dimensionless and between zero and one, and concentration ratios use compatible bases. Verify that tissue phase fractions sum to one if the model defines a complete two-phase composition. Recalculate the acid and base exponents with the correct sign: acids use \(pH-pK_a\), while bases use \(pK_a-pH\). A reversed sign can change a large accumulation prediction into a small one.

Also inspect the assumptions that are most influential. A predicted \(K_{p,uu}\) near one may be reasonable when tissue and plasma pH are similar or ionization is limited, while a large pH-driven ratio deserves scrutiny of the assumed pH gradient and passive-equilibrium premise. Large lipid terms should be checked against the phase definitions and input partition coefficient. Large binding corrections can arise when \(f_{u,t}\) is very small; uncertainty in that input may then dominate the total \(K_p\).

Wrong: “The tissue is acidic, so a weak base must have a low tissue concentration.” Correction: Under the passive neutral-equilibrium model, an acidic tissue can retain more total weak base because protonation increases the total-to-neutral factor.
Wrong: “The two methods disagree, so one must contain an arithmetic error.” Correction: First compare their composition, binding, and pH assumptions. Different representations can produce different estimates even when each calculation is done correctly.
Wrong: “A large \(K_{p,uu}\) means the same large total \(K_p\).” Correction: The total ratio also depends on \(f_{u,p}/f_{u,t}\).

These reduced calculations do not represent active transport, nonlinear or concentration-dependent binding, intracellular organelle pH gradients, or time-dependent disequilibrium. A measured concentration collected before distribution equilibrium may not be comparable to an equilibrium prediction. The earlier material on equilibrium data and sampling time emphasizes that sampling conditions must match the interpretation of \(K_p\).

Key takeaway: For ionizable drugs, calculate the pH-dependent unbound ratio with the correct acid or base expression, then apply binding and tissue-composition assumptions explicitly. A method comparison is informative only when the inputs, concentration bases, and equilibrium assumptions are stated.

Check Your Understanding

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

For a weak acid, why does the tissue-to-plasma aqueous ratio use the tissue total-to-neutral factor divided by the plasma factor?
A hypothetical weak base has \(K_{p,uu}=12.0\), \(f_{u,p}=0.060\), and \(f_{u,t}=0.030\). Calculate \(K_p\).
In a two-phase composition calculation, \(\phi_w=0.80\), \(\phi_l=0.20\), \(R_w=0.50\), \(P_l=3.0\), and \(f_{u,p}=0.040\). Calculate \(K_p^{(\mathrm{comp})}\).
Name two assumptions that could make the pH-partition calculation unreliable for a real tissue.

Solutions. (1) At neutral-form equilibrium, the neutral concentrations are equal across the aqueous compartments; each total concentration is its neutral concentration multiplied by its local total-to-neutral factor. Taking tissue over plasma therefore gives the tissue factor divided by the plasma factor. (2) \(K_p=12.0(0.060/0.030)=12.0(2)=24.0\), dimensionless. (3) \(K_p^{(\mathrm{comp})}=0.040[0.80(0.50)+0.20(3.0)]=0.040(0.40+0.60)=0.0400\), dimensionless. (4) Examples include failure of passive neutral-form equilibrium, uncertain or nonuniform pH, active transport, and non-linear binding; any two are sufficient.