The Two-Compartment IV Bolus Model

Why Use Two Compartments?

A one-compartment model treats the body as a single, well-mixed space: after an IV bolus, drug is represented by one amount and one concentration. That can be too simple when drug leaves the central circulation and distributes to tissues on a timescale that is distinguishable from elimination. A two-compartment model represents this behavior with a central compartment, where the dose enters and elimination occurs, and a peripheral compartment, which exchanges drug with the central compartment.

The compartments are mathematical representations, not necessarily individual organs or anatomical spaces. Their purpose is to describe the observed concentration-time pattern using a specified set of assumptions. The model’s key feature is that the two amounts change together: drug can move between compartments even while drug is being eliminated from the central one.

Model Assumptions and Amount Balances

The standard model assumes fixed positive values of \(V_c\), \(V_p\), \(CL\), and \(Q\); instantaneous and complete entry of the IV bolus into the central compartment; and well-mixed compartments, so each compartment has one concentration at a given time. Elimination is linear and occurs only from the central compartment: its rate is \(CL\,C_c\), with units of amount per time. Exchange is driven by the concentration difference. The net rate from central to peripheral is \(Q(C_c-C_p)\); if \(C_p>C_c\), this expression is negative and the net exchange is back toward the central compartment.

These assumptions give the following amount-balance equations. The exchange term appears with opposite signs because an amount leaving one compartment enters the other.

For an IV bolus at time zero, the initial conditions immediately after dosing are \(A_c(0^+)=D\) and \(A_p(0^+)=0\). Thus \(C_c(0^+)=D/V_c\) and \(C_p(0^+)=0\). The peripheral compartment does not receive a separate dose in this model.

Adding the two amount equations cancels the exchange terms. Exchange redistributes drug within the modeled system; only elimination changes the total amount. This is a useful check on both the equations and their implementation.

From Coupled Amounts to a Biexponential Concentration

To solve the model, write each amount in terms of its compartment concentration and divide by the corresponding volume. Define \(a=(CL+Q)/V_c\), \(b=Q/V_p\), \(c=Q/V_c\), and \(d=Q/V_p\). These quantities have units of inverse time. The system becomes a pair of coupled linear differential equations:

Solutions of a constant-coefficient linear system are combinations of exponential modes. The eigenvalues of the matrix are negative; write them as \(-\alpha\) and \(-\beta\), where \(\alpha>\beta>0\). Substituting an eigenvalue \(-r\) into the characteristic equation gives the following equation for the positive decay rates \(r\):

Since \(ad-bc=CL\,Q/(V_cV_p)\), the two rates are:

where the plus sign gives \(\alpha\) and the minus sign gives \(\beta\). For positive \(CL,Q,V_c,V_p\), the discriminant is positive: equivalently, it is \((a-d)^2+4bc\). The sum and product of the roots are \(S>0\) and \(P>0\), so both rates are positive. The central concentration therefore has the form \(C_c(t)=A e^{-\alpha t}+B e^{-\beta t}\).

The coefficients are determined by the bolus initial condition and the initial rate of change. Immediately after the dose, \(C_c(0^+)=D/V_c=C_0\). Because the peripheral amount is initially zero, the central equation gives \(dC_c/dt|_{0^+}=-aC_0\). Applying these two conditions to the exponential form gives \(A+B=C_0\) and \(\alpha A+\beta B=aC_0\). Solving these two linear equations yields:

This derivation explains why a bolus response can contain two exponential terms: the coupled system has two modes of change. The two rates and their coefficients are determined jointly by the volumes, clearances, and initial dose. A positive \(Q\) allows exchange between compartments; setting \(Q=0\) removes that coupling, leaving the peripheral compartment unconnected and the central response as a one-compartment decay.

Worked Example: Check Exchange and Elimination

Suppose, at a particular instant in a hypothetical model, \(A_c=40\) mg and \(A_p=20\) mg, with \(V_c=10\) L, \(V_p=20\) L, \(CL=2\) L/h, and \(Q=3\) L/h. These amounts need not be initial conditions; they describe a possible state at an instant. The concentrations are \(C_c=40/10=4\) mg/L and \(C_p=20/20=1\) mg/L.

The net central-to-peripheral exchange rate is \(Q(C_c-C_p)=3(4-1)=9\) mg/h. Elimination from the central compartment is \(CL\,C_c=2(4)=8\) mg/h. Therefore the central amount changes at \(-8-9=-17\) mg/h, while the peripheral amount changes at \(+9\) mg/h. The total amount changes at \(-17+9=-8\) mg/h, exactly the negative elimination rate. This check also shows why exchange is not an additional loss from the body.

Worked Example: Total AUC and Mass Balance

For the parameters in the first example, integrate the total-amount balance from immediately after the bolus until all drug has been eliminated. If the amounts approach zero at long times, the total amount falls from \(D\) to zero. Since \(d(A_c+A_p)/dt=-CL\,C_c\), integration gives \(D=CL\int_0^\infty C_c(t)\,dt\). Thus the central concentration AUC is \(D/CL\), provided the model’s linear elimination and complete follow-up to the terminal limit apply.

Numerically, the AUC is \(100\text{ mg}/(2\text{ L/h})=50.0\) mg·h/L. The exponential solution gives the same result: using the unrounded coefficients from the example, \(A=90/11\) mg/L and \(B=20/11\) mg/L, so:

The agreement is a model-based result, not evidence that the two-compartment model is necessarily the best description of any particular data set. It follows here from the IV dose, linear central clearance, and mass balance.

Diagnostics, Alternatives, and Limitations

A two-compartment model should be evaluated against the data and the dosing history, not selected solely because two exponentials can be written down. Check that dose timing and units are correct, that predicted concentrations are positive and consistent with the bolus initial condition, and that the amount-balance equations conserve drug apart from elimination. For observed concentration data, inspect residual patterns over time and across subjects using appropriate diagnostics. Systematic early-time misfit can suggest that the assumed distribution behavior, sampling times, or model structure is inadequate; persistent late-time misfit can signal an inadequate disposition description or other misspecification.

The model may be difficult to estimate reliably if observations are sparse in the period when the compartments exchange drug, if the sampling window does not capture the slower component, or if the available data contain little information to distinguish two modes from one. Parameter estimates can then be imprecise or strongly correlated. A fitted model should not be treated as proof that two anatomically distinct spaces have been identified. If \(Q\) is effectively zero or the data cannot distinguish the two components, the second compartment may add complexity without useful information.

The main simpler alternative is the one-compartment IV bolus model, which has one well-mixed amount, one volume, and central elimination; its concentration declines as a single exponential under linear elimination. The two-compartment model adds an exchanging peripheral amount and can represent a more complex concentration-time pattern. More elaborate structural models may add further compartments or different input processes, but complexity requires evidence and is not justified by mathematical flexibility alone.