Why Compartment Labels Matter
A compartment is not automatically an anatomical organ. In a lumped pharmacokinetic model, compartments are defined primarily by how the drug behaves: one may represent a rapidly equilibrating central space and another a more slowly exchanging peripheral space. In an organ-resolved model, state variables are assigned to named organs or tissues, such as liver, kidney, or muscle. The distinction matters because the model structure determines what its states mean and which physiological inputs can be represented explicitly.
The earlier material on model structure as a prediction choice motivates matching model detail to the question being asked. Here the focus is more specific: what changes when a model uses kinetically defined groups rather than organ-specific states? Neither structure is universally superior. A compact model may be adequate for a systemic concentration profile, while a question about concentration in a particular organ requires that the organ be represented or otherwise estimated through a defensible mapping.
In either structure, a model must specify how its states are initialized and how drug enters, moves through, or leaves the represented system. The difference is what the compartments refer to. A lumped model often describes exchange using fitted kinetic quantities, such as an intercompartmental clearance. An organ-resolved model can use inputs tied to physiology, including an organ’s blood flow \(Q_i\), volume \(V_i\), and drug-specific tissue properties. Such inputs do not make a model automatically accurate: their definitions, values, and relevance to the prediction still need to be justified.
States and Inputs at Different Resolutions
Consider a model with a central state and one peripheral state. Its state vector might be written as \(\mathbf{x}(t)=(A_c(t),A_p(t))\), where \(A_c\) and \(A_p\) are drug amounts in the central and peripheral compartments. The peripheral state represents a kinetic group, not necessarily one specific tissue. If the model is fitted to plasma observations, its central concentration may be the directly relevant output, while the peripheral amount helps describe distribution over time.
An organ-resolved representation might instead have \(\mathbf{x}(t)=(A_{\mathrm{liver}}(t),A_{\mathrm{kidney}}(t),A_{\mathrm{muscle}}(t),\ldots)\). The state names then support organ-specific outputs, for example \(C_{\mathrm{liver}}(t)\), if the corresponding volume and concentration definition are specified. Organ-level inputs can vary across physiological scenarios: changing kidney blood flow, for instance, can be represented distinctly from changing liver blood flow. In a purely lumped peripheral state, those organ-specific differences are not separate inputs unless the model is extended to include them.
For a state represented as an amount and a corresponding volume \(V_i\), concentration is calculated as \(C_i=A_i/V_i\). This conversion is meaningful only when the amount and volume refer to compatible boundaries and units. It also does not imply that every organ is uniform at finer scales: an organ-resolved model can still use a single representative concentration for each organ.
Worked Example: Naming the Model States
Suppose two hypothetical models are proposed for a drug after an intravenous dose. Model L has central and peripheral amounts, \(A_c\) and \(A_p\). Model O has amounts in liver, kidney, and skeletal muscle, \(A_{\mathrm{liver}}\), \(A_{\mathrm{kidney}}\), and \(A_{\mathrm{muscle}}\). What can each model report directly from its states?
Model L. Its two states identify kinetic groups. If a central volume of \(3.0\ \mathrm{L}\) is specified and \(A_c=6.0\ \mathrm{mg}\), the corresponding central concentration is \(6.0/3.0=2.0\ \mathrm{mg/L}\). The peripheral state does not reveal how much drug is in liver rather than kidney.
Model O. Its three states identify named organs. If \(A_{\mathrm{liver}}=12.0\ \mathrm{mg}\) and the corresponding liver volume is \(1.5\ \mathrm{L}\), then \(C_{\mathrm{liver}}=12.0/1.5=8.0\ \mathrm{mg/L}\). This calculation is valid under the model’s stated amount and volume definitions.
The example assumes that the listed state variables are the quantities the models actually track and that the example volumes are defined consistently with those states. It does not establish that either model predicts concentrations correctly; that requires suitable structure, inputs, and evaluation against the intended use.
What Combining States Does—and Does Not—Tell You
A basic information-loss argument explains why an organ-specific question can require organ-specific states. If two organ amounts are \(A_L\) and \(A_K\), a combined state \(S=A_L+A_K\) records their total. But the operation is not one-to-one: many organ pairs can produce the same \(S\). Knowing the total alone cannot determine the two amounts.
For example, if \(S=20\ \mathrm{mg}\), both \((A_L,A_K)=(12,8)\ \mathrm{mg}\) and \((18,2)\ \mathrm{mg}\) satisfy the equation. The total is sufficient to answer a question about that total, but not to distinguish the two organ amounts. This is a statement about what the state representation contains; it is not a claim that a particular lumped model can or cannot reproduce every output of a more detailed model.
Worked Example: A Combined Amount Hides an Organ Difference
In a hypothetical organ-resolved snapshot, liver contains \(12.0\ \mathrm{mg}\) of drug and kidney contains \(0.60\ \mathrm{mg}\). Their volumes are \(1.5\ \mathrm{L}\) and \(0.30\ \mathrm{L}\), respectively. Calculate the organ concentrations and the volume-weighted concentration for the combined amount and volume.
The organ concentrations are \(C_{\mathrm{liver}}=12.0/1.5=8.0\ \mathrm{mg/L}\) and \(C_{\mathrm{kidney}}=0.60/0.30=2.0\ \mathrm{mg/L}\). The total amount is \(12.0+0.60=12.6\ \mathrm{mg}\), and the combined volume is \(1.5+0.30=1.8\ \mathrm{L}\). Thus the volume-weighted combined concentration is \(12.6/1.8=7.0\ \mathrm{mg/L}\).
That combined value is a snapshot summary, not either organ’s concentration: it is \(1.0\ \mathrm{mg/L}\) below the liver concentration and \(5.0\ \mathrm{mg/L}\) above the kidney concentration. If the target is liver concentration, reporting \(7.0\ \mathrm{mg/L}\) as though it were the liver value would be incorrect. The calculation assumes the two amounts and volumes refer to the same time and to the stated combined region.
Physiological Inputs Versus Kinetic Description
The inputs available to a model depend on its structure. A lumped model may use a clearance or exchange parameter that summarizes drug movement between kinetic compartments. An organ-resolved model can instead distinguish organ-specific quantities. For example, blood flow \(Q_i\) can be specified for a named organ, and a scenario can alter one organ’s flow without requiring the same change in every other organ. This distinction is useful when the prediction concerns a physiological perturbation, but it also increases the number of inputs that need to be supported.
Worked Example: Translating Organ Flows into Model Inputs
Consider a hypothetical resting scenario with cardiac output \(Q=5.0\ \mathrm{L/min}\). A simplified organ-flow table assigns \(1.4\ \mathrm{L/min}\) to liver, \(1.0\ \mathrm{L/min}\) to kidney, \(1.2\ \mathrm{L/min}\) to skeletal muscle, and \(1.4\ \mathrm{L/min}\) to other tissues. Check the total and compare two ways to represent the flows.
The organ flows sum to \(1.4+1.0+1.2+1.4=5.0\ \mathrm{L/min}\), matching the stated cardiac output. Their fractions of that output are \(1.4/5.0=0.280\) for liver, \(1.0/5.0=0.200\) for kidney, \(1.2/5.0=0.240\) for muscle, and \(1.4/5.0=0.280\) for other tissues. The fractions sum to \(1.000\).
An organ-resolved model can assign the four flow inputs separately. A model that groups these tissues into one peripheral compartment could, for this particular grouping, supply the combined flow \(1.0+1.2+1.4=3.6\ \mathrm{L/min}\), but would not retain the separate kidney, muscle, and other-tissue values in that one input. The arithmetic assumes that the categories are disjoint and that the supplied flows use a consistent basis. Whether the compact grouping is adequate depends on the prediction target, not just on the fact that the totals agree.
Choosing and Checking the Level of Resolution
A practical choice starts with the quantity to be predicted. A systemic plasma profile may not require a separate state for every organ if the model is intended only to summarize the observed concentration-time behavior. A prediction of liver exposure, kidney exposure, or responses to organ-specific changes in physiology calls for a structure that can represent those distinctions, or for a separately justified method of obtaining them. Adding organ labels without reliable parameter values does not by itself improve a prediction.
Check the model against its intended interpretation. Confirm that each state has a clear meaning, units, and boundary; that a concentration is not confused with an amount; and that physiological inputs correspond to the same scenario and flow basis. If organ-specific observations or predictions are important, ask whether the model actually contains the relevant states. If only a total or plasma output is needed, ask whether the extra organ-level inputs are supported and useful rather than assuming greater detail is always better.
A common mistake is to treat the word “peripheral” as though it meant a specific organ, or to assign an organ-specific interpretation to a fitted exchange parameter without additional evidence. The correction is to distinguish the model’s mathematical state from an anatomical location: a kinetic compartment describes the model’s chosen grouping, while an organ-resolved state explicitly names the represented tissue. Another pitfall is to infer that a more detailed structure must be more accurate. Detail can make a target representable, but it also introduces inputs and assumptions whose uncertainty or mismatch can undermine the result.
| Feature | Lumped-compartment model | Organ-resolved model |
|---|---|---|
| Meaning of compartments | Kinetic groups, often defined by exchange behavior | Named organs or tissues at the model’s chosen scale |
| Typical states | Central and peripheral amounts or concentrations | Amounts or concentrations for individual organs |
| Potential inputs | Aggregate kinetic quantities such as exchange parameters | Organ-specific physiological quantities, such as flow and volume |
| Useful when | The target is adequately summarized by a few kinetic states | The target or scenario requires organ-specific distinctions |
| Key limitation | Organ-level differences are not directly identified by a pooled state | Requires more organ-specific inputs and assumptions |
Check Your Understanding
Answer each question, then compare your reasoning with the solutions below.
- A model has central and peripheral states. Does the name “peripheral” establish that its state is the amount in skeletal muscle? Explain.
- A hypothetical organ-resolved model has \(A_{\mathrm{kidney}}=4.0\ \mathrm{mg}\) in a corresponding volume of \(0.80\ \mathrm{L}\). What is the kidney concentration?
- Two organs contain amounts of \(5\ \mathrm{mg}\) and \(15\ \mathrm{mg}\). What information is available from their combined amount, and what organ-specific information is missing?
- A prediction concerns how a change in kidney blood flow affects kidney concentration. Which structure can represent this distinction more directly, and what additional inputs does that choice require?
Solutions. 1. No. “Peripheral” identifies a kinetic grouping unless the model definition explicitly assigns it to skeletal muscle; the name alone does not locate the state anatomically. 2. \(C_{\mathrm{kidney}}=A_{\mathrm{kidney}}/V_{\mathrm{kidney}}=4.0/0.80=5.0\ \mathrm{mg/L}\). 3. The combined amount is \(5+15=20\ \mathrm{mg}\). It does not establish which organ contains \(5\ \mathrm{mg}\) and which contains \(15\ \mathrm{mg}\), since the organ assignments could also be reversed. 4. An organ-resolved structure represents the kidney separately, so it can accept a kidney-specific flow input. It requires supported kidney-specific physiological and drug-related inputs, including a consistent flow value and the organ quantities needed to define the state and concentration.