Rodgers-Rowland Predictions for Neutral Drugs

From Tissue Composition to Neutral-Drug Kp

A neutral drug can distribute from water into both neutral lipids and phospholipids. Its predicted tissue concentration therefore depends not only on how much water and lipid a tissue contains, but also on how strongly the drug partitions into each lipid phase. The Rodgers-Rowland framework combines these contributions and compares the tissue total with the corresponding plasma total.

The result is a mechanistic estimate of the tissue-to-plasma partition coefficient, \(K_p\), for a neutral compound. It is not a measurement of tissue concentration or a substitute for checking the physicochemical inputs. The introductory material on tissue composition and neutral-drug partitioning supplies the relevant context; here the focus is the Rodgers-Rowland calculation and what its terms mean.

Definition: For a specified tissue and plasma reference, \(K_p\) is the ratio of total drug concentration per unit tissue volume to total drug concentration per unit plasma volume at equilibrium. For a neutral drug in the Rodgers-Rowland framework, the predicted value is the sum of tissue water, neutral-lipid, and phospholipid contributions divided by the corresponding plasma sum.

Let \(f_{w,t}\), \(f_{nl,t}\), and \(f_{ph,t}\) be the fractions of tissue volume occupied by water, neutral lipids, and phospholipids. Let \(f_{w,p}\), \(f_{nl,p}\), and \(f_{ph,p}\) be the corresponding fractions of plasma volume. The phase fractions are dimensionless and must be expressed on consistent volume bases. The neutral-lipid and phospholipid partition coefficients are denoted \(K_{nl}\) and \(K_{ph}\).

Model: For a neutral drug, the Rodgers-Rowland prediction is

K_p = \frac{f_{w,t}+f_{nl,t}K_{nl}+f_{ph,t}K_{ph}} {f_{w,p}+f_{nl,p}K_{nl}+f_{ph,p}K_{ph}}

In the standard empirical parameterization, the coefficients are predicted from the drug's octanol-water partition coefficient, \(P\), reported as \(\log P\) on a base-10 logarithmic scale:

K_{nl}=10^{0.30\log P-1.07}, \qquad K_{ph}=10^{0.90\log P-0.81}

These empirical relations distinguish the two lipid phases: the phospholipid coefficient rises more steeply with \(\log P\) than the neutral-lipid coefficient. Neither coefficient is itself the tissue-to-plasma \(K_p\). Each is a phase-to-water partition term that is multiplied by that phase's volume fraction in the tissue or plasma.

Why the Equation Has This Form

The equation follows by adding the amounts held in the three modeled phases. At equilibrium, use a shared aqueous concentration \(C_w\) as the reference. In the model, the concentration in neutral lipids is \(K_{nl}C_w\), and the concentration in phospholipids is \(K_{ph}C_w\). Thus total concentration per unit tissue volume is

C_t=C_w(f_{w,t}+f_{nl,t}K_{nl}+f_{ph,t}K_{ph})

The same reasoning for plasma gives \(C_p=C_w(f_{w,p}+f_{nl,p}K_{nl}+f_{ph,p}K_{ph})\). Dividing \(C_t\) by \(C_p\) cancels the shared aqueous reference concentration and gives the stated expression for \(K_p\). The cancellation depends on the model assumption that the drug is neutral and the phases are at equilibrium with a common aqueous reference.

This derivation also clarifies why the plasma denominator matters. The predicted \(K_p\) is not simply the tissue numerator: it is a ratio of tissue total to plasma total. Even if tissue composition is held fixed, changing the plasma reference fractions changes the prediction.

Worked Predictions

Worked Example: A Neutral Drug With Moderate Lipophilicity

Suppose a hypothetical neutral drug has \(\log P=2.00\). For an illustrative tissue, take \(f_{w,t}=0.60\), \(f_{nl,t}=0.020\), and \(f_{ph,t}=0.040\). Use illustrative plasma fractions \(f_{w,p}=0.930\), \(f_{nl,p}=0.0040\), and \(f_{ph,p}=0.0030\). These hypothetical fractions sum to 0.660 for the tissue and 0.937 for plasma; the unlisted volume is outside the three phases represented in this calculation.

First calculate the partition coefficients:

K_{nl}=10^{0.30(2.00)-1.07}=10^{-0.47}=0.3388

K_{ph}=10^{0.90(2.00)-0.81}=10^{0.99}=9.772

The tissue numerator is \(0.60+(0.020)(0.3388)+(0.040)(9.772)\). The three terms are \(0.6000\) from water, \(0.006777\) from neutral lipids, and \(0.3909\) from phospholipids, giving a total of \(0.9977\).

The plasma denominator is \(0.930+(0.0040)(0.3388)+(0.0030)(9.772)\), or \(0.9300+0.001355+0.02932=0.9607\). Therefore,

K_p=\frac{0.9977}{0.9607}=1.039

The tissue numerator shares are approximately 60.14% water, 0.6793% neutral lipids, and 39.18% phospholipids. These percentages describe the breakdown of the numerator, not the proportion of the drug in the whole body. In this example, the large water fraction contributes most, while phospholipids contribute substantially despite their smaller volume fraction.

Worked Example: A Less Lipophilic Neutral Drug

For a second hypothetical neutral drug, let \(\log P=1.00\), and use tissue fractions \(f_{w,t}=0.70\), \(f_{nl,t}=0.020\), and \(f_{ph,t}=0.030\). Keep the same illustrative plasma fractions as before. The tissue fractions sum to 0.750; the remaining volume is not represented as one of the three modeled phases.

The predicted phase coefficients are

K_{nl}=10^{0.30(1.00)-1.07}=10^{-0.77}=0.1698, \qquad K_{ph}=10^{0.90(1.00)-0.81}=10^{0.09}=1.230

The tissue contributions are \(0.70\) from water, \((0.020)(0.1698)=0.003396\) from neutral lipids, and \((0.030)(1.230)=0.03691\) from phospholipids. The tissue numerator is \(0.7403\). In plasma, the denominator is \(0.930+(0.0040)(0.1698)+(0.0030)(1.230)=0.9344\). Thus,

K_p=\frac{0.7403}{0.9344}=0.7923

Here the water term dominates the tissue numerator. The lipid terms are smaller because both the predicted phase coefficients and the phase fractions limit their contributions. The result, below 1, means the predicted total concentration per unit tissue volume is lower than that per unit plasma volume under these model inputs.

Worked Example: How Lipid Contributions Change With logP

Return to the tissue and plasma fractions in the first example, but consider a hypothetical neutral drug with \(\log P=3.00\). Its coefficients are

K_{nl}=10^{0.30(3.00)-1.07}=10^{-0.17}=0.6761, \qquad K_{ph}=10^{0.90(3.00)-0.81}=10^{1.89}=77.62

The tissue terms are \(0.6000\) for water, \((0.020)(0.6761)=0.01352\) for neutral lipids, and \((0.040)(77.62)=3.105\) for phospholipids. The tissue numerator is \(3.719\). The plasma denominator is \(0.930+(0.0040)(0.6761)+(0.0030)(77.62)=1.166\). Therefore,

K_p=\frac{3.719}{1.166}=3.190

In the tissue numerator, phospholipids contribute approximately \(3.105/3.719\times100=83.50\%\), water contributes approximately 16.14%, and neutral lipids contribute approximately 0.364%. In this hypothetical calculation, the phospholipid term dominates because its coefficient has risen substantially with \(\log P\). The example illustrates the equation's arithmetic, not a claim that every tissue or neutral drug will follow this pattern in measured data.

Assumptions and Input Checks

A calculation is interpretable only if its assumptions and inputs match the intended use. The neutral-drug Rodgers-Rowland expression assumes that the compound is represented as neutral, that the tissue and plasma phases are at equilibrium, and that distribution can be represented by water, neutral lipids, and phospholipids with the stated coefficients. It also assumes that the composition fractions and partition coefficients use compatible volume and concentration bases.

Before calculating, check the following:

Drug form. The method here is for neutral drugs. Do not apply the neutral equation to an ionizable drug while ignoring its ionized fraction or treating \(\log D\) as though it were automatically \(\log P\).
Property assignment. Use \(\log P\) appropriate to the neutral compound. Confirm that it is on the base-10 scale and that the value has not been confused with a solubility, molecular-weight, or distribution measurement.
Composition basis. Tissue fractions must refer to tissue volume and plasma fractions to plasma volume. Fractions must be entered consistently as decimals, not mixed with percentages. For example, 4% is entered as 0.040.
Denominator and units. The plasma denominator must be calculated using the same phase coefficients as the tissue numerator. The resulting \(K_p\) is a ratio of concentrations and is dimensionless.
Scope of the model. The equation does not explicitly represent every possible binding environment or distribution process. A numerical prediction should not be interpreted as proof that the three phases fully explain a particular measured tissue concentration.

Reading Contributions and Comparing Approaches

The three tissue numerator terms are useful for explaining a prediction: \(f_{w,t}\) is the water contribution, \(f_{nl,t}K_{nl}\) is the neutral-lipid contribution, and \(f_{ph,t}K_{ph}\) is the phospholipid contribution. A phase's volume fraction alone does not determine its importance. A smaller phospholipid fraction can contribute more than a larger water fraction when its partition coefficient is sufficiently high. Conversely, a large lipid coefficient may have a modest effect if the relevant tissue fraction is very small.

Compared with a simpler neutral-drug composition prediction such as Poulin-Theil, Rodgers-Rowland explicitly separates neutral lipids and phospholipids and uses distinct empirical relationships to \(\log P\) for those phases. This makes the phase contributions more visible, while also making the prediction dependent on the phase-specific coefficients and composition inputs. The methods should not be treated as interchangeable equations with identical assumptions.

A practical diagnostic is to inspect the numerator and denominator before interpreting their ratio. A very large predicted \(K_p\) may result from a large phospholipid term, a small denominator, or both. A negative result is inconsistent with the positive fractions and positive partition coefficients in this model and points to an input or arithmetic error. A near-zero positive result is mathematically possible, though its plausibility depends on the intended tissue and inputs. Also check whether the entered fractions are plausible for the intended tissue and whether a reported “contribution” is being confused with a fraction of total tissue volume.

Common Mistake and Limitations

Common mistake: “The tissue contains only 4% phospholipid, so phospholipids can contribute at most 4% of the predicted partitioning.” This is incorrect because the phospholipid volume fraction is multiplied by \(K_{ph}\). The relevant numerator term is \(f_{ph,t}K_{ph}\), and it can exceed the water or neutral-lipid term when \(K_{ph}\) is large enough. Compare the calculated terms, rather than comparing volume fractions alone.

The model is a composition-based prediction, not a complete description of every determinant of tissue exposure. Its output depends on the empirical partition-coefficient relationships and the quality of the composition and \(\log P\) inputs. It does not, by itself, establish that equilibrium has been reached in a particular experiment or explain a discrepancy between predicted and measured values. For that reason, report the inputs and, when useful, the separate phase terms alongside the final \(K_p\).

Check Your Understanding

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

For a neutral drug, what does the Rodgers-Rowland \(K_p\) compare?
Why can a small phospholipid fraction make a large contribution to the tissue numerator?
If the tissue fractions are entered as percentages while the plasma fractions are entered as decimals, what is wrong with the calculation?
Using the first worked example, which tissue phase contributes the least to the numerator, and approximately what percentage does it contribute?

Solutions. (1) It compares total concentration per unit tissue volume with total concentration per unit plasma volume at equilibrium. (2) The phospholipid fraction is multiplied by \(K_{ph}\), so a sufficiently large coefficient can make \(f_{ph,t}K_{ph}\) large despite a small fraction. (3) The composition inputs are on inconsistent scales; convert all fractions to decimals before applying the equation. (4) Neutral lipids contribute the least: their term is approximately \(0.006777\) of a numerator of \(0.9977\), or about 0.6793%.