Washout Planning Across Multiple Replicate Periods

Why Washout Planning Is a Timing Problem

In a multi-period crossover, a washout is the time between the last administration in one treatment period and the first administration in the next. It is not simply the number of calendar dates separating periods. The interval must be long enough to limit residual exposure from the prior period, and it must be checked for every transition in every assigned sequence.

Replicate schedules repeat a treatment within a participant, so the design may involve several period-to-period transitions. Each gap can differ because of the calendar, dosing frequency, or sequence. Planning therefore requires two linked checks: estimate a scientifically acceptable minimum interval, then verify that the actual schedule meets it. The earlier material on Carryover Risks in Replicate Schedules motivates the concern; here the focus is on calculating and auditing the time allowance.

Definition: A washout interval is the elapsed time from the last administration of a treatment in one period to the first administration in the next period. A residual-exposure criterion specifies the maximum remaining concentration or amount, often as a fraction \(q\) of a stated reference level, that is considered acceptable at the start of the next period. There is no universal value of \(q\) that is appropriate for every drug or study.

From Half-Life to a Minimum Interval

For a simple terminal elimination phase, assume that concentration declines exponentially after the last relevant input. If \(C_{\text{stop}}\) is the concentration at the chosen reference time and \(t_{1/2}\) is the terminal half-life, the concentration after \(W\) hours is

C(W)=C_{\text{stop}}e^{-kW} =C_{\text{stop}}2^{-W/t_{1/2}}, \qquad k=\frac{\ln 2}{t_{1/2}}.

The reference level must be stated clearly: it might be a measured concentration at the end of dosing, a model-based concentration, or an amount on a defined dose-equivalent scale. If the requirement is \(C(W)/C_{\text{stop}}\le q\), where \(0<q<1\), solving the inequality gives the minimum washout.

Derivation: Under first-order terminal elimination, \(e^{-kW}\le q\). Taking natural logarithms gives \(-kW\le\ln(q)\). Since \(k>0\), rearranging reverses the inequality when dividing by \(-k\): \(W\ge-\ln(q)/k\). Substituting \(k=\ln(2)/t_{1/2}\) yields \(W\ge t_{1/2}\ln(1/q)/\ln(2)\).

This result is a planning calculation, not a universal regulatory rule. For example, a criterion of 5% of the stated reference concentration requires about 4.322 terminal half-lives, because \(\ln(20)/\ln(2)\approx4.322\). A five-half-life rule leaves \(2^{-5}=0.03125\), or 3.125%, under this model. Whether either residual level is acceptable depends on the pharmacology, assay sensitivity, endpoint, and study purpose.

Worked Example: Translating a Residual Limit Into Hours

Suppose a hypothetical compound has a terminal half-life of 18 hours. The protocol team chooses, for this planning illustration, a maximum residual level of 5% of the concentration at the end of the prior treatment period. Calculate the minimum interval under the exponential-elimination model.

The assumed model is first-order terminal decline with a constant half-life and no further input after the last dose. The criterion is \(q=0.05\), so

W_{\min} =18\frac{\ln(1/0.05)}{\ln(2)} =18\frac{\ln(20)}{\ln(2)} =18(4.3219) =77.79\text{ hours}.

Thus, an interval of at least 77.8 hours meets the stated 5% criterion according to this model. An interval of 84 hours also meets it, with about 6.21 hours beyond the calculated minimum. The result is not a guarantee that carryover is impossible: it is conditional on the assumed half-life, concentration reference, and residual threshold.

Repeated Dosing and Accumulation

A repeated regimen can produce a higher concentration at the end of a period than a single dose. This matters because even if the same fraction of concentration remains after a given number of half-lives, that fraction is applied to a larger starting level. For linear one-compartment elimination with instantaneous doses at a fixed dosing interval \(\tau\), the steady-state accumulation ratio relative to the corresponding single-dose concentration immediately after a dose is

R_{\text{acc}}=\frac{1}{1-e^{-k\tau}}.

This expression sums the contributions from the current dose and earlier doses, each reduced by elimination over one or more dosing intervals. It is a useful illustration, not a general model for every dosing regimen. Real absorption, distribution, multiple compartments, or non-steady-state dosing can change the concentration-time profile.

If all residual concentration after stopping a repeated regimen declines with the same terminal rate, then its fraction of the end-of-treatment concentration after \(W\) hours is still \(2^{-W/t_{1/2}}\). But the absolute concentration depends on how high the end-of-treatment level was. A criterion expressed relative to a single-dose reference may therefore require more washout than a criterion expressed as a fraction of the accumulated end-of-treatment concentration. These denominators must not be confused.

Worked Example: Accumulation Changes the Absolute Residual

Consider a hypothetical drug given once every 24 hours, with a 24-hour terminal half-life. Use the instantaneous-dose, linear-elimination model at steady state. Express the single-dose post-dose concentration as 1.00 concentration unit. The study team requires the residual concentration to be no more than 0.05 of that single-dose unit after the repeated regimen ends.

First, \(e^{-k\tau}=2^{-24/24}=0.5\), so the accumulation ratio is

R_{\text{acc}}=\frac{1}{1-0.5}=2.

The end-of-treatment concentration on this scale is therefore 2.00 units. After washout \(W\), the predicted residual is \(2.00(2^{-W/24})\). Apply the specified absolute threshold:

2.00(2^{-W/24})\le0.05 \quad\Longrightarrow\quad W\ge24\frac{\ln(2/0.05)}{\ln(2)} =24\log_2(40) =127.7\text{ hours}.

A 120-hour washout leaves \(2.00(2^{-5})=0.0625\) units, which exceeds 0.05 and does not meet the criterion. A 144-hour washout leaves \(2.00(2^{-6})=0.03125\) units, which is below 0.05 and does meet it. If instead the criterion were 5% of the accumulated end-of-treatment concentration, the requirement would be \(24\log_2(20)=103.7\) hours. The difference comes from the reference level, not a change in the elimination half-life.

Auditing Actual Multi-Period Schedules

Once a minimum interval has been justified, calculate each scheduled gap from actual administration times. For every sequence, record the last administration in period \(j\), the first administration in period \(j+1\), and their elapsed hours. Include the clock time; counting intervening dates can misstate the true interval. If there are multiple doses in a period, the relevant start of washout is the last dose, not the first or the nominal period date.

Worked Example: Checking Every Transition in a Replicate Schedule

A hypothetical three-period schedule gives each period's treatment at 08:00 on its listed day. The planned minimum interval is 80 hours. Sequence A is scheduled on days 1, 5, and 9; sequence B is scheduled on days 1, 5, and 10. Assume one administration per period and that the stated criterion applies to every transition.

For sequence A, the first gap is from day 1 at 08:00 to day 5 at 08:00: \(4\times24=96\) hours. The second gap is also \(4\times24=96\) hours. Both exceed 80 hours by 16 hours.

For sequence B, the first gap is \(4\times24=96\) hours. The second is \(5\times24=120\) hours. Both meet the minimum. The shortest gap across the schedule is 96 hours, so the schedule satisfies the 80-hour requirement, provided these are the actual administration times and no additional dose occurs later in a period.

A separate clock-time check illustrates why the interval should be calculated rather than inferred from day labels. Tuesday 08:00 to Friday 20:00 is \(3\times24+12=84\) hours. That actual schedule interval meets an 80-hour minimum by 4 hours. A 72-hour interval, considered separately, does not meet the 80-hour minimum; it falls short by \(80-72=8\) hours.

Assumptions, Diagnostics, and Alternatives

A half-life calculation is most defensible when the relevant terminal phase is known, elimination is approximately first order over the concentration range of interest, and the washout clock starts after the final dose. Check whether the selected half-life describes the analyte that matters for the next period. A pharmacologically active metabolite may persist longer than the parent drug. Delayed absorption, modified-release formulations, enterohepatic recirculation, or flip-flop kinetics can also make the apparent terminal decline different from the expected elimination phase. Nonlinear kinetics invalidate simple proportional scaling when exposure or elimination changes with concentration.

Diagnostics are practical as well as model-based. Audit the dosing and visit calendar for every sequence and transition; verify that administration timestamps, not just nominal study days, support the intended gap. Compare observed predose concentrations with the planned residual expectation when such measurements are available. Unexpectedly high or persistent predose values can indicate an incorrect half-life assumption, a missed late dose, accumulation beyond the model, or another source of residual exposure. A fixed margin beyond the calculated minimum can address schedule uncertainty, but it does not repair a wrong kinetic model.

The main alternatives answer related but different planning needs. A fixed washout in days is simple to implement, but should be justified against the relevant pharmacokinetics and checked in hours. A direct residual-concentration target is more closely tied to a biological or analytical threshold, but requires appropriate measurements and a defensible threshold. Pharmacokinetic simulation can represent repeated dosing, variable absorption, multiple compartments, and uncertainty in half-life; it is preferable when the simple exponential calculation is inadequate, but depends on the quality of the model and inputs. None of these approaches alone proves that carryover cannot affect an endpoint.

A common mistake is to say that five half-lives always remove the drug, or to count calendar days instead of elapsed time. Five half-lives leave 3.125% under a single exponential model; they do not mean zero concentration, and the remaining amount may still matter. Likewise, a three-date separation is not necessarily 72 hours: Tuesday 08:00 to Friday 20:00 is 84 hours. State the reference concentration, residual criterion, kinetic assumptions, and actual gap so the conclusion can be checked.

Key takeaway: Set a residual-exposure criterion, translate it into a minimum interval using an appropriate kinetic model, and verify the actual elapsed time for every transition. For repeated dosing, specify whether the criterion is relative to the accumulated end-of-treatment level or to another reference, because that choice changes the required washout.

Check Your Understanding

Use the stated assumptions in each question. Brief solutions follow each problem.

A hypothetical drug has a 12-hour terminal half-life. How long is required for the concentration to fall to 10% of its reference level? Solution: \(W=12\ln(10)/\ln(2)=39.86\) hours, or about 39.9 hours under the exponential model.
A schedule requires at least 80 hours between periods. Does a 72-hour gap satisfy the requirement, and by how much? Solution: No. Since \(72<80\), it falls short by \(80-72=8\) hours.
A last dose is given Monday at 18:00 and the next period begins Friday at 06:00. Find the elapsed interval. Solution: Monday 18:00 to Thursday 18:00 is 72 hours; Thursday 18:00 to Friday 06:00 adds 12 hours, for 84 hours total.
Why can a washout calculated relative to a single-dose level be longer after repeated dosing? Solution: Repeated dosing can raise the end-of-treatment concentration through accumulation. The remaining fraction may decline at the same terminal rate, but it is a fraction of a larger starting concentration, so more time may be needed to reach the same absolute threshold.