Why Victim-Dose and Sampling Times Must Be Coordinated
A pharmacokinetic sample has meaning only in relation to the dose that precedes it. A concentration collected two hours after a victim-drug dose represents a different phase of the profile from one collected two hours before that dose, even if both blood draws occur at the same clock time on different days. In a drug-drug interaction (DDI) study, this distinction matters in every treatment condition: the victim dose and the samples must be placed so the intended interaction is present and the relevant parts of the victim profile are observed.
The scheduling task is to coordinate two timelines. One is the perpetrator timeline: when the perpetrator is expected to produce the intended interaction state. The other is the victim timeline: when the victim dose is given and when samples are collected relative to it. Earlier material on scheduling perpetrator dosing and capturing induction and time-dependent inhibition addresses how the interaction state develops. Here, the focus is on placing the victim dose and samples within that state.
This definition assumes that dose and collection times are recorded accurately, that the dose referred to is unambiguous (especially with repeated doses), and that the intended perpetrator condition is established at the victim dose and maintained over the sampling interval as required. If the perpetrator's effect is still changing during the interval, the schedule must also represent that time-varying interaction; simply matching elapsed time after the victim dose does not make the perpetrator state identical.
Build the Schedule From the Victim-Dose Time
PBPK simulations and available pharmacokinetic knowledge can indicate when absorption is expected, the likely period around the peak, and how long the elimination phase needs to be followed. Those predictions guide timing; they do not guarantee the exact profile for every participant. Sampling plans therefore need to cover the expected phases with enough resilience to plausible variation, rather than relying on one predicted peak time.
For each treatment condition, first identify the intended perpetrator state at the victim dose. Then specify the victim dose time and express each proposed collection time as elapsed time after that dose. Check that the schedule includes the early absorption period, a sufficiently informative region around the model-predicted peak, later elimination observations, and the end of the exposure interval relevant to the dosing regimen. A predose sample can establish a baseline concentration for a dose, but it is not a postdose observation and should not be labelled as one.
Where conditions are compared, use comparable elapsed times after the victim dose. If the perpetrator shifts a predicted profile feature, such as the timing of the expected peak, the shared schedule may need additional samples to cover both predictions. Keep clock time and other relevant conditions, such as meals, as consistent as practicable. Otherwise, a treatment difference can become entangled with a difference in time of day or dosing circumstances.
Why Elapsed-Time Alignment Works
Let \(C_j(\tau)\) denote the victim concentration at elapsed time \(\tau\) in treatment condition \(j\). This notation makes the comparison explicit: concentrations at the same \(\tau\) refer to the same nominal time since the victim dose, even if collection clock times differ. If one condition has a victim dose at 08:00 and another at 10:00, a sample at 12:00 is four hours after the first dose but only two hours after the second.
Thus, matching elapsed times requires \(T_{s,1}-T_{v,1}=T_{s,2}-T_{v,2}\), not necessarily \(T_{s,1}=T_{s,2}\). This is the scheduling justification for using dose-relative collection times. It does not assert that the resulting concentrations should be equal: treatment condition may change the profile, and the study is designed to characterize that difference. It only ensures that the compared observations correspond to comparable positions on the postdose time axis.
Worked Example: Converting a Predicted Profile Into Collection Times
Suppose a hypothetical PBPK model predicts that an oral victim drug is absorbed mainly during the first three hours, that its expected peak is between 2.5 and 3.5 hours, and that later elimination needs to be followed through a 24-hour dosing interval. The planned victim dose is at 08:00. Propose a dose-relative schedule and convert it to clock times.
A schedule that covers the stated phases is 0 (predose), 0.5, 1, 2, 3, 4, 6, 8, 12, and 24 hours after the dose. The closely spaced early samples describe absorption and bracket the expected peak period; 6 and 8 hours extend observation beyond that period; 12 and 24 hours provide later observations and the end-of-interval collection.
For a postdose sample at \(\tau=3\) hours, the clock time is \(08{:}00+3\) hours \(=11{:}00\). For \(\tau=12\) hours, it is \(08{:}00+12\) hours \(=20{:}00\). For \(\tau=24\) hours, it is \(08{:}00+24\) hours \(=08{:}00\) the following day. The zero-hour sample is collected immediately before the 08:00 dose, not after it.
The schedule is appropriate to the hypothetical predictions, but the 24-hour collection must be operationally defined. If another victim dose is due at that time, obtain the end-of-interval sample before administering it. Apply the same elapsed-time plan in the comparison condition, with corresponding dose-relative collection times.
Account for Each Treatment Condition
The victim dose should be timed so that the intended interaction is present when the victim drug enters the body and, where relevant, during the subsequent observation interval. That does not mean the victim dose must occur at the same clock time as a perpetrator dose. The relevant question is whether the perpetrator exposure or effect at the victim dose and during follow-up matches the intended condition. For an interaction that changes over time, use the predicted condition-specific time course to select the dosing window and to decide whether observation must extend across changing interaction intensity.
The sample times should then be defined relative to the victim dose in each condition. When PBPK predictions differ by condition, form a schedule that captures the relevant periods in both profiles. A schedule designed only around the control prediction can miss a shifted or broadened feature under the interacting condition. Conversely, sampling at many arbitrary times does not ensure useful coverage if the early or late parts of the interval are omitted.
Worked Example: Covering Different Predicted Peak Windows
In a hypothetical DDI design, the victim dose is planned for 13:00 in both conditions. The perpetrator regimen is arranged so that the intended interaction state is expected to be present at 13:00 and throughout the next 24 hours. A PBPK prediction places the control condition's expected peak period at 1–3 hours after the victim dose and the interacting condition's at 2–5 hours. Early absorption is expected during hours 0–2, and later elimination should be observed through hour 24.
A shared dose-relative schedule might include 0, 0.5, 1, 2, 3, 4, 5, 6, 8, 12, and 24 hours. The added 4- and 5-hour samples ensure coverage of the later predicted peak period in the interacting condition, while the earlier samples still cover absorption and the control prediction. This choice is based on both condition-specific model predictions, not on the assumption that the profiles have the same shape.
Convert the key times to clock times: 13:00 plus 2 hours is 15:00; plus 5 hours is 18:00; and plus 24 hours is 13:00 the following day. These conversions are identical in both conditions because the victim dose is at the same clock time. If the victim dose time had differed, each condition's collection clock times would need to shift to preserve the same elapsed times.
The assumption that the interaction is present throughout the 24-hour interval must be checked against the hypothetical perpetrator time course. If it is not maintained, the schedule may still characterize the victim profile, but it would not represent one constant interaction state. The study plan would then need to recognize the changing perpetrator condition rather than describe the interval as uniformly interacting.
Operational Checks and Failure Modes
Before finalizing a schedule, make a condition-by-condition time table that states the perpetrator regimen, victim dose clock time, each sample clock time, and elapsed time since the relevant victim dose. This simple audit can reveal mislabeled predose samples, collections assigned to the wrong dose, and times that look aligned on a calendar but are not aligned postdose. Specify whether a nominal time is a collection target or a permitted window, and define how deviations will be recorded.
Diagnostics include checking actual dose and collection timestamps, comparing actual elapsed times with the planned times, and confirming whether meals or other protocol conditions were held consistent. If a draw is delayed, calculate its actual elapsed time rather than treating the nominal time as exact. A missed sample near a predicted profile feature or at the end of the interval should be documented as a potential limitation, not silently replaced by a nearby clock-time draw.
Worked Example: Detecting a Misaligned Collection
A hypothetical control-period victim dose is scheduled for 08:00, and the interacting-period dose for 10:00. In both periods, the protocol lists a 12:00 sample and calls it the “2-hour sample.” Check the label and correct the schedule.
In the control period, elapsed time is \(12{:}00-08{:}00=4\) hours. In the interacting period, elapsed time is \(12{:}00-10{:}00=2\) hours. The two samples are not aligned: one is four hours postdose and the other is two hours postdose. To collect both at two hours, schedule the control draw at \(08{:}00+2=10{:}00\) and the interacting draw at \(10{:}00+2=12{:}00\). Alternatively, to compare four-hour samples, use 12:00 in control and 14:00 in the interacting period.
This correction aligns elapsed time, but the shifted clock times may introduce other differences, such as meal timing or time of day. Those conditions should be standardized or explicitly considered in the design. A correct postdose label alone does not resolve every source of schedule-related variability.
Alternatives, Limitations, and a Common Pitfall
A dose-relative schedule is usually the clearest basis for interpreting samples because it preserves time since victim administration. A fixed clock-time schedule can be useful when clock time itself is important or when tightly synchronized visits are operationally necessary, but it is suitable for comparing postdose phases only if victim doses also occur at matching clock times. Otherwise, convert each draw to elapsed time and interpret the mismatch explicitly. When dose times cannot be matched, a planned and documented shift in collection times is generally preferable to collecting at identical clock times and assuming the observations are equivalent.
| Approach | Strength | Risk or limitation |
|---|---|---|
| Dose-relative collection times | Directly identifies the postdose phase in each condition | Clock times may differ; meal and time-of-day conditions need attention |
| Fixed clock-time collections | Can simplify coordination across participants or visits | Can compare different elapsed times if victim doses shift |
| Condition-specific schedule informed by PBPK | Can cover predicted differences in profile timing | Predictions may be wrong; a schedule tailored too narrowly can miss individual variation |
No schedule can guarantee that every participant's actual peak falls exactly on a collection time, and a protocol based on a predicted profile is limited by the model and available prior information. The later material on sampling schedules for AUC and Cmax considers endpoint-oriented sampling choices in more detail; the present scheduling principle is to coordinate the dose and observations so the relevant phases are observable in each condition.
Common pitfall: “The blood draws were at the same clock times, so the profiles are aligned.” This is wrong whenever the victim doses occurred at different times. The correction is to calculate \(T_s-T_v\) for every draw and compare samples at the intended elapsed times, while also checking whether clock-time differences affect other protocol conditions.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- A victim dose is given at 07:30 and a sample is collected at 10:00. What is the elapsed time after the dose?
- In one condition, the victim dose is at 09:00; in another, it is at 11:00. A sample is collected at 13:00 in both. Are the samples aligned by elapsed time? If not, give clock times for matched two-hour samples.
- A model predicts an expected peak period at 2–4 hours in control and 3–5 hours in an interacting condition. What scheduling principle should guide the shared collection plan?
- Why can dose-relative alignment be insufficient if the perpetrator effect changes substantially during the observation interval?
Solutions. (1) \(10{:}00-07{:}30=2.5\) hours after the dose. (2) No. The elapsed times are four and two hours, respectively. Matched two-hour samples would be at 11:00 in the first condition and 13:00 in the second. (3) Choose times that cover both predicted periods, including the later 3–5-hour window, while also covering absorption and the rest of the relevant interval. (4) Matching time since victim dose does not make the perpetrator state the same; the interaction may differ across the interval, so its predicted time course must also be considered.