The One-Compartment IV Bolus Exponential

From Amount Balance to Concentration

After an intravenous bolus, drug enters the systemic circulation over a very short time. A one-compartment model treats the body as one well-mixed space: at any time, the concentration is taken to be the same throughout that space. The amount of drug then decreases as elimination removes a fraction of the amount per unit time. Together, these assumptions lead to a simple exponential concentration-time curve.

The amount balance for this model is \(dA(t)/dt=-k_e A(t)\), where \(A(t)\) is the amount of drug in the compartment at time \(t\), and \(k_e\) is the first-order elimination rate constant. This tutorial uses that balance to derive the concentration equation and interpret its terms. It does not assume that the curve describes every observed concentration exactly: the equation is a structural prediction under stated assumptions.

Definition: For a one-compartment IV bolus model, let \(D\) be the bolus dose, \(V\) the constant apparent volume of the single well-mixed compartment, and \(k_e\) the nonnegative first-order elimination rate constant. For \(t\geq 0\), the amount follows \(dA(t)/dt=-k_eA(t)\), with \(A(0)=D\). Concentration is \(C(t)=A(t)/V\). Here, \(D\) and \(A\) have units of amount, \(V\) has units of volume, \(C\) has units of amount per volume, and \(k_e\) has units of inverse time.

Deriving the Exponential Equation

The bolus is treated as entering the compartment instantaneously at time zero, so the initial amount is the administered dose: \(A(0)=D\). For \(A(t)>0\), separate the variables in the amount-balance equation and integrate from time zero to time \(t\):

\frac{dA}{A}=-k_e\,dt \qquad\Longrightarrow\qquad \int_D^{A(t)}\frac{dA}{A}=-k_e\int_0^t dt

Evaluating the integrals gives \(\ln(A(t))-\ln(D)=-k_e t\). Exponentiating both sides yields \(A(t)=D\exp(-k_e t)\). If \(k_e=0\), the differential equation instead gives \(A(t)=D\), which is also represented by this expression. With constant volume, dividing the amount by \(V\) gives the concentration-time equation:

C(t)=\frac{D}{V}\exp(-k_e t)=C(0)\exp(-k_e t), \qquad C(0)=\frac{D}{V}.

The exponential argument \(k_e t\) is dimensionless: a rate in inverse hours multiplied by a time in hours has no units. The initial concentration \(C(0)\) has concentration units. For \(k_e>0\), the factor \(\exp(-k_e t)\) starts at 1, remains positive, and decreases toward zero as time increases. Thus the model predicts a positive concentration at each finite time, with a declining curve that approaches zero asymptotically.

The value \(C(0)=D/V\) is the model's concentration immediately after the instantaneous dose has mixed throughout the compartment. It is an initial condition, not necessarily an actually measured value. A sample collected after the bolus and mixing have taken time will be at \(t>0\); treating that sample as \(C(0)\) would generally misstate the starting concentration.

Using the Equation

Worked Example: Predicting Concentration After a Bolus

Suppose, hypothetically, a one-compartment model has an IV bolus dose of \(500\) mg, volume \(25\) L, and \(k_e=0.20\ \mathrm{h}^{-1}\). Find the model-predicted concentration at \(3\) hours.

First calculate the initial concentration: \(C(0)=D/V=500\ \mathrm{mg}/25\ \mathrm{L}=20\ \mathrm{mg/L}\). The product in the exponential is \(k_e t=(0.20\ \mathrm{h}^{-1})(3\ \mathrm{h})=0.60\), so the units cancel.

C(3\ \mathrm{h})= 20\ \mathrm{mg/L}\times \exp(-0.60) \approx 20\times 0.5488\ \mathrm{mg/L} =10.98\ \mathrm{mg/L}.

Under the model assumptions, the predicted concentration is about \(10.98\ \mathrm{mg/L}\) at 3 hours. This is a model prediction, not a guarantee that a measured concentration will equal that value.

Worked Example: Reading Initial Concentration and Decline

Consider a hypothetical dose of \(300\) mg, a volume of \(15\) L, and \(k_e=0.15\ \mathrm{h}^{-1}\). Calculate concentrations at \(0\) and \(4\) hours, and describe what the two values represent.

The initial concentration is \(C(0)=300\ \mathrm{mg}/15\ \mathrm{L}=20\ \mathrm{mg/L}\). At 4 hours, \(k_e t=(0.15\ \mathrm{h}^{-1})(4\ \mathrm{h})=0.60\), so:

C(4\ \mathrm{h})=20\ \mathrm{mg/L}\times\exp(-0.60) \approx 10.98\ \mathrm{mg/L}.

The model starts at \(20\ \mathrm{mg/L}\), immediately after the bolus is mixed in the compartment, and predicts about \(10.98\ \mathrm{mg/L}\) after 4 hours. The decrease does not occur at a constant number of mg/L per hour. The model specifies a constant fractional rate of loss of amount; consequently, the absolute decline depends on how much drug remains.

Worked Example: Inferring the Rate Constant From Two Concentrations

Suppose, hypothetically, the same one-compartment IV bolus model has an initial concentration of \(24\ \mathrm{mg/L}\), and its concentration at 5 hours is \(12\ \mathrm{mg/L}\). Calculate the rate constant implied by these two values.

Starting with \(C(t)=C(0)\exp(-k_e t)\), divide by \(C(0)\) and take natural logarithms: \(\ln(C(t)/C(0))=-k_e t\). Therefore \(k_e=\ln(C(0)/C(t))/t\). Substituting the values gives:

k_e= \frac{\ln(24\ \mathrm{mg/L}\,/\,12\ \mathrm{mg/L})}{5\ \mathrm{h}} = \frac{\ln(2)}{5\ \mathrm{h}} \approx 0.1386\ \mathrm{h}^{-1}.

The concentration units cancel in the ratio, leaving a dimensionless logarithm; dividing by hours gives inverse-hour units. This calculation is valid as a model-based calculation only if the concentrations represent the same exponential phase, with the stated time origin and one-compartment assumptions.

Interpreting Parameters and Checking the Model

The equation separates two roles. The ratio \(D/V\) sets the curve's initial height, while \(k_e\) determines how rapidly the exponential factor decreases over time. Holding \(V\) and \(k_e\) fixed, doubling \(D\) doubles every predicted concentration. Holding \(D\) and \(k_e\) fixed, a larger \(V\) lowers the initial concentration and scales the entire curve downward. Holding \(D/V\) fixed, a larger positive \(k_e\) gives a faster decline after time zero.

These statements describe the equation, not causal effects established from observational data. In practice, estimated or observed curves can differ for many reasons, including measurement error, uncertainty in recorded dose or time, and misspecification of the structural model. In a mixed-effects analysis, subjects may have different parameter values; the equation then describes an individual trajectory conditional on that subject's parameters. It should not be assumed that a pooled concentration curve is itself the trajectory of a typical individual.

The model is most appropriate when the bolus can reasonably be treated as instantaneous, mixing is sufficiently rapid for a single well-mixed compartment to be useful, volume is constant, and elimination is first-order over the range being modeled. These are modeling assumptions, not facts guaranteed by the equation. The time origin must correspond to the bolus, and dose, volume, concentration, and time units must be compatible.

One practical diagnostic is to compare observed concentrations with the fitted exponential curve over time. A persistent pattern—such as early observations systematically above or below the curve, or later observations bending away from it—can signal that a single exponential is not capturing the concentration-time shape. Before changing the model, check units, dose timing, sample times, and whether the first sample was taken after a meaningful delay. Random scatter around the curve is different from a systematic time pattern.

A straight-line decline in concentration is not the same model. Under the exponential equation, the concentration's instantaneous slope is \(dC(t)/dt=-k_eC(t)\); it depends on the current concentration. A constant-slope linear model instead subtracts the same concentration amount per unit time, and can eventually predict negative concentrations. The exponential form is appropriate for the stated first-order amount balance; a linear form would require a different assumption about elimination.

Likewise, a good-looking exponential curve does not establish that the body truly behaves as one homogeneous space. If observed data show a reproducible shape that a single exponential cannot represent, adding complexity may be justified, but it must be supported by the data and scientific context. The one-compartment equation remains useful as a clear baseline: it makes explicit how the dose, volume, and first-order decline combine in the simplest IV bolus setting.

Common Mistakes

Calling \(D/V\) the concentration at every time. It is the model's initial concentration. For \(t>0\), multiply it by \(\exp(-k_e t)\).
Using inconsistent units. If \(k_e\) is in \(\mathrm{h}^{-1}\), time must be in hours for \(k_e t\) to be dimensionless.
Interpreting \(k_e\) as a constant absolute concentration loss. It is a first-order rate constant; the instantaneous rate of concentration decline is proportional to the current concentration.
Ignoring the time origin. The equation's \(t=0\) is the modeled bolus time. A delayed sample is not automatically an observation of \(C(0)\).
Treating a model prediction as an exact observation. The equation gives a structural prediction under assumptions; measured values can differ because of noise or model mismatch.

Key takeaway: Under an instantaneous IV bolus, a constant-volume, one-compartment model with first-order elimination predicts \(C(t)=(D/V)\exp(-k_e t)\). The initial value is \(D/V\); the exponential factor describes the subsequent decline.

Check Your Understanding

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

A hypothetical dose is \(400\) mg, \(V=20\) L, and \(k_e=0.10\ \mathrm{h}^{-1}\). What concentration does the model predict at 2 hours?
In the equation, what does \(D/V\) represent, and why does it have concentration units?
If \(k_e\) is in \(\mathrm{h}^{-1}\), what time units should be used in the exponential argument?
Why does a systematic curved pattern in residual differences from a fitted exponential raise concern about the one-compartment assumption?

Solutions. (1) \(C(0)=400/20=20\ \mathrm{mg/L}\), and \(k_e t=(0.10)(2)=0.20\). Thus \(C(2)=20\exp(-0.20)\approx16.37\ \mathrm{mg/L}\). (2) \(D/V\) is the modeled initial concentration immediately after mixing; amount divided by volume has units of amount per volume. (3) Use hours, so the product of \(\mathrm{h}^{-1}\) and hours is dimensionless. (4) A systematic pattern suggests the exponential curve is missing a time-dependent feature rather than merely differing from observations by unsystematic scatter; assumptions, timing, and the structural form should be checked.