Worked Structural Models For IV Profiles

Why Work Through Intravenous Profiles?

Intravenous (IV) data remove the absorption process from the concentration–time profile. That makes them useful for examining disposition: how drug distributes and is eliminated after entering the systemic circulation. A structural model translates assumptions about those processes into a predicted profile. The central question here is how one- and two-compartment models calculate different IV profiles under bolus and infusion dosing.

The calculations below assume linear, time-invariant disposition: transfer and elimination rates are proportional to the amount present, and their parameters remain constant over the interval considered. These assumptions allow the model equations to be solved with exponential functions. They are not universal properties of drugs; they are modeling assumptions that should be checked against the data and context.

Definition: In a one-compartment model, the body is represented as one kinetically homogeneous space with volume \(V\), and drug is eliminated from that space with clearance \(CL\). In a two-compartment model, drug is distributed between a central compartment with volume \(V_c\) and a peripheral compartment with volume \(V_p\). Elimination occurs from the central compartment, while drug transfers between compartments with intercompartmental clearance \(Q\). In both models, concentration is the amount in the relevant compartment divided by its volume.

One-Compartment IV Bolus

An ideal IV bolus is delivered instantaneously into the compartment. For a one-compartment model, the initial concentration is \(D/V\), where \(D\) is the dose. First-order elimination gives the differential equation \(dA/dt=-kA\), where \(A\) is the amount in the body and \(k=CL/V\) is the elimination rate constant. Since \(C=A/V\), the concentration profile is:

C(t)=\frac{D}{V}e^{-kt} =\frac{D}{V}e^{-(CL/V)t},\qquad t\geq 0.

Here \(C(t)\) is concentration at time \(t\); if \(D\) is in mg, \(V\) in L, and \(CL\) in L/h, then \(C(t)\) is in mg/L and \(k\) is in h\(^{-1}\). The equation predicts an immediate concentration of \(D/V\), followed by a monoexponential decline. It assumes instantaneous mixing throughout the modeled space and no distribution phase that must be represented separately.

Worked Example: One-Compartment Bolus Profile

Suppose a hypothetical 100 mg IV bolus is described by \(V=20\) L and \(CL=4\) L/h. Assume instantaneous mixing, linear elimination, and constant parameters. Then \(k=CL/V=4/20=0.2000\) h\(^{-1}\), and the initial concentration is \(100/20=5.000\) mg/L.

At \(t=2\) h:

C(2)=\frac{100\ \mathrm{mg}}{20\ \mathrm{L}} e^{-(0.2000\ \mathrm{h}^{-1})(2\ \mathrm{h})} =5.000e^{-0.4000} \approx 3.352\ \mathrm{mg/L}.

The units in the exponent cancel, as required. The result is below the initial concentration, consistent with elimination. A different \(V\) changes the initial concentration for a fixed dose; a different \(CL/V\) changes the rate of decline.

Two-Compartment IV Bolus

A two-compartment model allows distribution between a central and peripheral space. Let \(A_c\) and \(A_p\) be the amounts in those spaces. Define the first-order rate constants \(k_{10}=CL/V_c\), \(k_{12}=Q/V_c\), and \(k_{21}=Q/V_p\). For an IV bolus into the central compartment, the model is:

\frac{dA_c}{dt}=-(k_{10}+k_{12})A_c+k_{21}A_p,\qquad \frac{dA_p}{dt}=k_{12}A_c-k_{21}A_p.

These equations assume that each compartment is internally well mixed, transfer is first order in the amount in the compartment from which drug moves, and elimination occurs only from the central compartment. The bolus sets \(A_c(0)=D\) and \(A_p(0)=0\). The resulting central concentration has two exponential terms:

C_c(t)=\frac{D}{V_c} \left[ a e^{-\alpha t}+b e^{-\beta t} \right],\qquad a=\frac{\alpha-k_{21}}{\alpha-\beta},\quad b=\frac{k_{21}-\beta}{\alpha-\beta}.

The positive macro-rate constants \(\alpha\) and \(\beta\), with \(\alpha>\beta\), are the roots of the characteristic equation \(\lambda^2-(k_{10}+k_{12}+k_{21})\lambda+k_{10}k_{21}=0\). The faster component is commonly associated with the distribution phase and the slower component with the terminal phase, though observed phases should not be over-interpreted as direct measurements of separate physiological processes. The coefficients satisfy \(a+b=1\), so the formula gives \(C_c(0)=D/V_c\).

Justification: The coefficient form follows from the two eigenvalues of the amount-balance equations. The central concentration must be a sum \(A e^{-\alpha t}+B e^{-\beta t}\). At \(t=0\), \(A+B=D/V_c\). Also, the initial peripheral amount is zero, so \(dC_c/dt|_{t=0}=-(k_{10}+k_{12})D/V_c\). Solving these two linear equations for \(A\) and \(B\), using the characteristic equation to relate \(\alpha,\beta\) to the rate constants, gives \(A=(D/V_c)(\alpha-k_{21})/(\alpha-\beta)\) and \(B=(D/V_c)(k_{21}-\beta)/(\alpha-\beta)\). Thus the biexponential profile is a consequence of the model’s coupled first-order equations, not an additional elimination mechanism.

Worked Example: Comparing One- and Two-Compartment Bolus Profiles

Consider a hypothetical 100 mg bolus. Use \(V_c=20\) L, \(V_p=30\) L, \(CL=4\) L/h, and \(Q=6\) L/h for the two-compartment model. These imply \(k_{10}=0.2000\) h\(^{-1}\), \(k_{12}=0.3000\) h\(^{-1}\), and \(k_{21}=0.2000\) h\(^{-1}\). The characteristic equation is \(\lambda^2-0.7000\lambda+0.04000=0\), giving \(\alpha\approx0.6372\) h\(^{-1}\) and \(\beta\approx0.06277\) h\(^{-1}\). The coefficients are \(a\approx0.7611\) and \(b\approx0.2389\).

At 2 h, the central concentration is:

C_c(2)=\frac{100}{20} \left[ 0.7611e^{-(0.6372)(2)} +0.2389e^{-(0.06277)(2)} \right] \approx 5(0.2128+0.2107) \approx 2.1175\ \mathrm{mg/L}.

For comparison, the one-compartment example with \(V=20\) L and \(CL=4\) L/h predicts \(3.352\) mg/L at 2 h. Both models have the same initial concentration, \(5.000\) mg/L, but they do not predict the same decline: the two-compartment model includes redistribution as well as elimination. At sufficiently late times, its slower exponential term dominates; the terminal decline is governed by \(\beta\), not by \(k_{10}\) alone.

Constant-Rate IV Infusion

An infusion adds drug over time rather than placing the entire dose in the body at \(t=0\). Let \(R_0\) be the constant input rate, with units of amount per time, and let \(T\) be the infusion duration. In the one-compartment model, the amount satisfies \(dA/dt=R_0-kA\) during the infusion, with \(A(0)=0\). Solving this equation gives the rising concentration; after the input stops, the concentration declines by elimination alone.

C(t)= \begin{cases} \displaystyle \frac{R_0}{CL}\left(1-e^{-kt}\right), & 0\leq t\leq T,\\[6pt] \displaystyle \frac{R_0}{CL}\left(1-e^{-kT}\right)e^{-k(t-T)}, & t>T. \end{cases}

The amount infused is \(R_0T\). During a sufficiently long infusion, the one-compartment concentration approaches \(R_0/CL\), because input and elimination balance. For a finite infusion, that limiting value need not be reached.

Worked Example: One-Compartment Infusion During and After Input

Suppose the hypothetical one-compartment parameters are \(V=20\) L and \(CL=4\) L/h, and the infusion rate is \(R_0=20\) mg/h for \(T=4\) h. The model assumptions are the same as for the bolus calculation, with a constant input during the stated interval. Here \(k=0.2000\) h\(^{-1}\), the total infused amount is \(20\times4=80\) mg, and the limiting concentration under a continuing infusion is \(R_0/CL=5.000\) mg/L.

At the end of the four-hour infusion:

C(4)=\frac{20}{4}\left(1-e^{-(0.2000)(4)}\right) =5.000(1-e^{-0.8000}) \approx 2.753\ \mathrm{mg/L}.

At \(t=6\) h, two hours after the infusion stops:

C(6)=C(4)e^{-(0.2000)(6-4)} \approx 2.7534e^{-0.4000} \approx 1.846\ \mathrm{mg/L}.

The concentration is lower after the infusion ends because input has ceased while elimination continues. It is not appropriate to substitute \(t=6\) into the during-infusion expression.

For the two-compartment model, the infusion profile can be obtained by integrating the bolus response over the drug entering at each instant. During infusion, each exponential component accumulates separately; after the infusion ends, each accumulated component declines at its own rate:

C_c(t)=\frac{R_0}{V_c} \left[ a\frac{1-e^{-\alpha t}}{\alpha} +b\frac{1-e^{-\beta t}}{\beta} \right],\quad 0\leq t\leq T, $$ $$ C_c(t)=\frac{R_0}{V_c} \left[ a\frac{1-e^{-\alpha T}}{\alpha}e^{-\alpha(t-T)} +b\frac{1-e^{-\beta T}}{\beta}e^{-\beta(t-T)} \right],\quad t>T.

For example, the first term during infusion is the integral of the fast bolus-response component: \(\int_0^t (R_0/V_c)a e^{-\alpha(t-u)}du=(R_0/V_c)a(1-e^{-\alpha t})/\alpha\). The same integration gives the slow term. After \(T\), the input integral ends at \(T\), and the accumulated terms are multiplied by their respective post-infusion exponential declines.

Worked Example: Two-Compartment Infusion Profile

Use the two-compartment parameters from the bolus example: \(V_c=20\) L, \(CL=4\) L/h, \(Q=6\) L/h, \(\alpha\approx0.6372\) h\(^{-1}\), \(\beta\approx0.06277\) h\(^{-1}\), \(a\approx0.7611\), and \(b\approx0.2389\). Give a hypothetical constant infusion of \(20\) mg/h for 4 h. The model assumes constant parameters, first-order transfer and elimination, and input into the central compartment.

At 4 h, \(R_0/V_c=1.000\) mg/L/h. The fast and slow contributions are approximately:

0.7611\frac{1-e^{-(0.6372)(4)}}{0.6372} \approx 1.101\ \mathrm{mg/L}, \qquad 0.2389\frac{1-e^{-(0.06277)(4)}}{0.06277} \approx 0.8450\ \mathrm{mg/L}.

Therefore, \(C_c(4)\approx1.101+0.8450=1.946\) mg/L. The total input is 80 mg, but the central concentration is not calculated as 80 mg divided by \(V_c\): drug has been eliminated and some has distributed to the peripheral compartment during the infusion. The model’s long-infusion central concentration approaches \(R_0/CL=5.000\) mg/L, but four hours is not enough to reach that steady state.

Comparing Predictions and Checking Model Adequacy

The one-compartment model gives a single exponential after a bolus and a single accumulation process during infusion. The two-compartment model gives a sum of two exponentials, allowing an early distribution component and a slower terminal component. Both are linear models in these examples, but they make different predictions because they represent different disposition structures.

Feature	One compartment	Two compartments
Bolus profile	One exponential decline	Sum of fast and slow exponentials
Modeled spaces	One homogeneous space	Central and peripheral spaces with exchange
Parameters shaping profile	\(V\) and \(CL\)	\(V_c\), \(V_p\), \(CL\), and \(Q\)
Infusion response	Single accumulation and post-infusion decline	Two accumulating components and two post-infusion declines

When examining observed IV profiles, systematic curvature in a plot of log concentration against time can be a clue that a single exponential is inadequate. A two-compartment model can represent a steeper early decline followed by a shallower terminal decline. This is not proof that the two-compartment model is correct: sampling times, assay limits, dose records, and other sources of model misspecification also matter. Early samples help characterize rapid distribution, while sufficiently late samples are needed to characterize the terminal phase. Sparse sampling in either phase can leave the corresponding rate and coefficient poorly informed.

The equations also have limits. A true bolus may take time to administer; a changing infusion rate requires a time-varying input rather than the constant-rate formula. Time-varying clearance, nonlinear elimination, or changing transfer processes violate the assumptions used to obtain these exponential solutions. Moreover, a compartment is a useful mathematical representation, not necessarily a literal anatomical space. Structural choices should be judged by whether their assumptions and predictions are defensible for the data and purpose.

Common Mistakes

Using the wrong infusion expression. The during-infusion equation applies only up to \(T\). After \(T\), use the concentration at the end of infusion multiplied by the appropriate decline term.
Treating the infused amount as an instantaneous bolus. An infusion delivers \(R_0T\) over time; elimination and distribution occur while drug enters.
Confusing \(CL\) with \(k\). Clearance has units of volume per time, whereas \(k=CL/V\) has units of inverse time.
Reading a two-exponential fit as direct proof of two anatomical spaces. The biexponential curve follows from the chosen model equations; the mathematical components do not, by themselves, establish distinct physiological structures.
Rounding too early. Keep unrounded parameter values through intermediate calculations, then round the final concentration to a sensible number of significant figures.

Key takeaway: An IV bolus or infusion profile follows from the input history and the disposition structure. State the assumptions, use the matching time-specific equation, and compare the predicted curve with the features the data can actually support.

Check Your Understanding

Try each problem, then compare your reasoning with the solutions below.

A hypothetical one-compartment model has \(D=60\) mg, \(V=12\) L, and \(CL=3\) L/h. Calculate \(k\) and the predicted concentration at 2 h after an IV bolus.
For an infusion at \(R_0=12\) mg/h into a one-compartment model with \(CL=3\) L/h, what is the limiting concentration if the infusion continues indefinitely? State the assumption behind this result.
Why does a two-compartment bolus profile contain two exponentials even though elimination occurs only from the central compartment?
A four-hour infusion is stopped. Which rate constant governs the one-compartment decline after stopping, and which rates govern the two-compartment decline?

Solutions. (1) \(k=CL/V=3/12=0.2500\) h\(^{-1}\). Thus \(C(2)=(60/12)e^{-(0.2500)(2)}=5e^{-0.5}\approx3.033\) mg/L. (2) The limiting concentration is \(R_0/CL=12/3=4.000\) mg/L, assuming constant input and linear, time-invariant one-compartment disposition until steady state is approached. (3) The coupled central and peripheral amount equations have two eigenvalues; their central-concentration response is a sum of two exponential terms. The second exponential reflects the coupled distribution–elimination dynamics, not a second elimination site. (4) The one-compartment decline is governed by \(k=CL/V\). In the two-compartment model, both accumulated terms decline after stopping, one with \(\alpha\) and the other with \(\beta\).