Solving a Two-Compartment Repeated-Dose Schedule

Why the Two-Compartment Profile Matters

A dose interval that works for a one-exponential concentration decline need not work when distribution between central and peripheral compartments produces two exponential phases. The more rapid phase can contribute substantially to the early post-dose peak, while the slower phase can persist and shape the pre-dose trough. A schedule should therefore be checked using the full concentration profile, not by treating the terminal half-life as if it described the entire interval.

This tutorial considers repeated instantaneous intravenous boluses in a linear two-compartment model. It shows how to calculate the steady-state concentration after each dose, find the peak and limiting trough, and evaluate candidate intervals against specified concentration limits. The calculations assume a fixed dose and interval; they do not address nonlinear elimination or response-guided titration.

Definition: For a two-compartment linear model after an instantaneous bolus, write the single-dose concentration response as $C_1(t)=a e^{-\alpha t}+b e^{-\beta t}$, for $t\geq 0$, where $a$ and $b$ are concentration contributions, $\alpha$ and $\beta$ are positive rate constants, and typically $\alpha>\beta$. For repeated boluses separated by interval $\tau$, $C_{\mathrm{ss}}(t)$ denotes the steady-state concentration $t$ time units after a dose, on $0\leq t<\tau$. The post-dose peak is $C_{\mathrm{ss}}(0)$; the limiting pre-dose trough is $\lim_{t\to\tau^-}C_{\mathrm{ss}}(t)$.

Assumptions and Repeated-Dose Formula

The concentration coefficients $a$ and $b$ depend on the dose and the model parameters. In this tutorial they are given in concentration units for the specified dose. The rate constants $\alpha$ and $\beta$ have units of inverse time, so that $\alpha t$ and $\beta t$ are dimensionless. For the usual central-compartment response to a bolus, both exponential coefficients are positive. We assume linear, time-invariant pharmacokinetics, identical doses, a constant interval, no missed doses, and no additional concentration input or elimination pathway that changes over time. Under linearity, each dose contributes the same response, shifted in time, and the total concentration is the sum of those contributions.

At steady state, the current dose contributes $a e^{-\alpha t}+b e^{-\beta t}$. Earlier doses were given $\tau,2\tau,3\tau,\ldots$ time units before the current dose. Thus, the $\alpha$-component includes the geometric series $1+e^{-\alpha\tau}+e^{-2\alpha\tau}+\cdots$, with a corresponding series for the $\beta$-component. Since both rate constants and $\tau$ are positive, each series converges.

C_{\mathrm{ss}}(t) = \frac{a e^{-\alpha t}}{1-e^{-\alpha\tau}} + \frac{b e^{-\beta t}}{1-e^{-\beta\tau}}, \qquad 0\leq t<\tau.

The denominator $1-e^{-\alpha\tau}$ accounts for accumulation of the rapid component; $1-e^{-\beta\tau}$ does the same for the slower component. They must not be combined into a single accumulation factor because the two components decay at different rates.

Result: For $a,b,\alpha,\beta,\tau>0$, the steady-state profile is strictly decreasing between doses. It attains its maximum at $t=0$. On the half-open interval $0\leq t<\tau$, it has no attained minimum; its infimum is the limiting trough as $t\to\tau^-$. The peak and trough are $$ C_{\max,\mathrm{ss}} = \frac{a}{1-e^{-\alpha\tau}}+ \frac{b}{1-e^{-\beta\tau}}, \qquad C_{\min,\mathrm{ss}} = \frac{a e^{-\alpha\tau}}{1-e^{-\alpha\tau}}+ \frac{b e^{-\beta\tau}}{1-e^{-\beta\tau}}. $$

To justify the monotonicity, differentiate the profile. Each term in the derivative is negative because its coefficient and rate constant are positive and its exponential is positive:

\frac{dC_{\mathrm{ss}}(t)}{dt} = -\frac{\alpha a e^{-\alpha t}}{1-e^{-\alpha\tau}} -\frac{\beta b e^{-\beta t}}{1-e^{-\beta\tau}} <0.

Therefore the profile decreases throughout $0\leq t<\tau$. It reaches its maximum at the included left endpoint, $t=0$, and approaches the stated trough as $t\to\tau^-$. The trough is the concentration just before the next bolus; at that bolus, concentration jumps upward. It is appropriate to include that pre-dose limit when evaluating a regimen's trough, while keeping clear that the limit is not an attained minimum on the half-open interval.

Worked Example: Calculating the Peak and Trough at Steady State

Consider a hypothetical fixed bolus whose single-dose response is $C_1(t)=(36/7)e^{-t}+(6/7)e^{-t/8}$ mg/L, with $t$ measured in hours. Thus $a=36/7$ mg/L, $b=6/7$ mg/L, $\alpha=1\ \mathrm{h}^{-1}$, and $\beta=0.125\ \mathrm{h}^{-1}$. Calculate the steady-state peak and limiting trough for $\tau=4$ h.

The peak is evaluated just after a dose, at $t=0$:

C_{\max,\mathrm{ss}} = \frac{36/7}{1-e^{-4}} + \frac{6/7}{1-e^{-0.5}} \approx 5.23881+2.17842 =7.41723\ \mathrm{mg/L}.

The first exponential has accumulated by a factor $1/(1-e^{-4})$; the second has accumulated by $1/(1-e^{-0.5})$. The different exponents reflect the different rates and must be retained in each term.

The limiting trough is calculated by evaluating the steady-state profile as $t\to4^-$:

C_{\min,\mathrm{ss}} = \frac{(36/7)e^{-4}}{1-e^{-4}} + \frac{(6/7)e^{-0.5}}{1-e^{-0.5}} \approx 0.09595+1.32128 =1.41723\ \mathrm{mg/L}.

At steady state the profile decreases from about $7.417$ mg/L just after a dose toward a pre-dose concentration of about $1.41723$ mg/L. The slower component contributes most of this trough, even though it is only one part of the single-dose response.

Evaluating Candidate Intervals Against a Concentration Window

Suppose the design criteria are a peak no greater than $8.00$ mg/L and a limiting trough no lower than $1.95$ mg/L. These are hypothetical targets for illustration, not clinical recommendations. Each candidate interval must satisfy both inequalities using the same two-compartment model and dose-specific coefficients. A regimen that meets only one bound is not feasible under these criteria.

Worked Example: Comparing Three Candidate Intervals

Use the same hypothetical model and dose as above. Compare $\tau=3.0$, $3.2$, and $4.0$ h with the constraints $C_{\max,\mathrm{ss}}\leq8.00$ mg/L and $C_{\min,\mathrm{ss}}\geq1.95$ mg/L.

For $\tau=3.0$ h, substitution gives a peak of approximately $8.153$ mg/L and a trough of approximately $2.153$ mg/L. The trough meets its lower bound, but the peak exceeds its upper bound, so this interval fails.

For $\tau=3.2$ h, the peak components are:

\frac{36/7}{1-e^{-3.2}}\approx5.3614, \qquad \frac{6/7}{1-e^{-0.4}}\approx2.5999.

Their sum is $C_{\max,\mathrm{ss}}\approx7.9613$ mg/L. At the trough, multiply each accumulated component by its decay across the interval:

C_{\min,\mathrm{ss}} = \frac{(36/7)e^{-3.2}}{1-e^{-3.2}} + \frac{(6/7)e^{-0.4}}{1-e^{-0.4}} \approx0.2185+1.7428 =1.9613\ \mathrm{mg/L}.

Both conditions are met: $7.9613<8.00$ mg/L and $1.9613>1.95$ mg/L. The interval passes, but the small margins make rounding and parameter uncertainty relevant in a real design assessment.

For $\tau=4.0$ h, the peak is approximately $7.41723$ mg/L, but the trough is approximately $1.41723$ mg/L, as calculated above. The peak constraint is met and the trough constraint is not. Among these three candidates, $3.2$ h is the only interval satisfying both stated limits.

The comparison also illustrates the trade-off: shortening the interval tends to raise the trough, but increases accumulation and can raise the peak. The calculation should not be reduced to checking one concentration or a single half-life.

Finite Dosing Versus Steady State

The steady-state expression represents the limiting profile after many repeated doses. A patient who has received only a few doses will generally not yet have that profile. After $N$ identical doses at times $0,\tau,\ldots,(N-1)\tau$, the concentration $t$ time units after the most recent dose is the sum of $N$ shifted single-dose responses. For either rate constant $k$, the finite sum is $1+e^{-k\tau}+\cdots+e^{-(N-1)k\tau}=(1-e^{-Nk\tau})/(1-e^{-k\tau})$.

Worked Example: Peak After Three Doses

For the hypothetical model above, suppose three doses have been administered at 0, 4, and 8 h. Calculate the concentration immediately after the third dose, taking that dose time as $t=0$. This is not the steady-state peak.

The contributions from the current dose and the two earlier doses yield:

C_3(0) = \frac{36}{7}(1+e^{-4}+e^{-8}) + \frac{6}{7}(1+e^{-0.5}+e^{-1}).

Using $e^{-4}\approx0.0183156$, $e^{-8}\approx0.00033546$, $e^{-0.5}\approx0.606531$, and $e^{-1}\approx0.367879$:

C_3(0) \approx 5.142857(1.018651) + 0.857143(1.974410) \approx5.23878+1.69235 =6.93113\ \mathrm{mg/L}.

This finite-dose peak is below the steady-state peak of $7.41723$ mg/L because only two prior doses contribute to accumulation, rather than the infinite sequence represented at steady state.

Practical Checks, Alternatives, and Limitations

Before using a calculated interval, check that the model input and arithmetic describe the intended regimen. The dose-specific coefficients must be for the dose being evaluated, and every time value must use the same units as the rate constants. A rate constant in $\mathrm{h}^{-1}$ cannot be combined with an interval entered in days without conversion. Verify that both accumulation denominators are positive and that the trough is computed from the entire profile, not from just the terminal exponential.

Check direction. For positive coefficients, a shorter interval increases accumulation and usually raises the trough, while also raising the peak. If a calculation shows the profile increasing between boluses, recheck the signs and timing.
Check the trough boundary. Use the left-hand limit before the next bolus. Do not describe it as an attained minimum over $0\leq t<\tau$, even though it is the clinically relevant pre-dose value.
Check convergence and scale. When $\alpha\tau$ or $\beta\tau$ is small, $1-e^{-k\tau}$ is small and accumulation is substantial. Numerical software should evaluate such expressions carefully to avoid loss of precision; the function $\operatorname{expm1}$ is useful for computing $e^x-1$ accurately near zero.
Check the desired regimen. A periodic steady-state comparison is not a loading-dose calculation and does not describe the first dosing interval. A finite-dose calculation or a separate loading regimen is needed when the early profile is the design target.

The formula here is an analytic alternative to simulating a concentration-time curve numerically. For this linear, two-exponential case, the closed form is transparent and makes peak and trough calculations direct. Numerical simulation can be more flexible when doses vary over time, infusion inputs are present, or the model has additional compartments, but its predictions should still be checked against the model and dose history. A one-compartment exponential approximation is simpler, but it can misrepresent both the early peak and the slow-component trough when distribution is important.

The method is limited by its assumptions. It presumes linear pharmacokinetics, fixed model parameters, instantaneous boluses, and a regular repeated schedule. If clearance depends on concentration, the superposition and geometric-series calculation do not apply; if absorption or infusion takes time, the single-dose response must represent that input instead. Further, meeting deterministic peak and trough thresholds does not by itself account for between-person parameter variability or uncertainty in the estimated model. Those considerations require uncertainty-aware evaluation rather than simply treating the coefficients and rate constants as known without error.

Common Mistake

A common error is to calculate the steady-state accumulation using the slow rate constant for both components, on the reasoning that the terminal phase determines the trough. That changes the rapid component's accumulation and gives an incorrect peak and trough. The correction is to preserve each exponential's own rate constant in its own denominator, and to evaluate both components at the same interval endpoint. Another error is to round component values too early: add unrounded contributions first, then round the final concentration to a sensible number of significant figures.

Key takeaway: Under linear two-compartment kinetics, the repeated-bolus steady-state profile is the sum of two separately accumulated exponential components. Calculate both the immediate post-dose peak and the limiting pre-dose trough, then check each candidate interval against every specified concentration constraint.

Check Your Understanding

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

For a positive two-exponential response with $\alpha>\beta>0$, why is the steady-state profile decreasing between doses?
In the hypothetical model used above, what is the expression for the steady-state peak at interval $\tau=2$ h? Is it valid to use $1-e^{-0.25}$ as the denominator for both components?
Why is the trough in this tutorial a limit as $t\to\tau^-$, rather than a minimum attained on $0\leq t<\tau$?
What key assumption fails if the amount eliminated per unit time depends on concentration, and what consequence does that have for the repeated-dose formula?

Solutions. (1) Differentiation gives $-\alpha a e^{-\alpha t}/(1-e^{-\alpha\tau})-\beta b e^{-\beta t}/(1-e^{-\beta\tau})$, which is strictly negative for positive coefficients, rates, and interval. (2) The peak is $(36/7)/(1-e^{-2})+(6/7)/(1-e^{-0.25})$ mg/L. The denominator $1-e^{-0.25}$ applies only to the slower component; the rapid component uses $1-e^{-2}$. (3) The interval excludes $t=\tau$, so the decreasing profile approaches its pre-dose value but does not attain it within the interval. The next dose then causes a concentration jump. (4) Concentration superposition and the geometric-series accumulation rely on linear time-invariant kinetics. With concentration-dependent elimination, each dose's contribution is not an independent shifted copy of the same response, so this formula cannot be used as written.