Why Predose Concentrations Deserve Attention
In a crossover study, a predose sample is collected before administration of the treatment for that period. It provides a check on what is present at the start of the dosing interval. In later periods, a detectable concentration of the study drug may be consistent with drug remaining from an earlier period. It may also reflect an analytical or sampling problem, an unrecorded exposure, or—in suitable settings—an endogenous source. The value alone does not identify its cause.
This distinction matters because a concentration already present at time zero can contribute to the measured concentration-time profile after the current dose. If that contribution differs according to the treatment given in the preceding period, it may complicate interpretation of a treatment comparison. The earlier material on carryover signals gives the broader crossover context; here the focus is on what predose measurements can show and how to apply a study’s own rules.
Interpret a predose value alongside its collection time, assay LLOQ, analyte, period, sequence, preceding treatment, and the protocol’s instructions. A detectable result is a signal to assess, not proof of residual drug or an automatic reason to remove data. There is no single predose cutoff that applies to every drug, assay, protocol, or regulatory setting.
What Patterns Can and Cannot Establish
A residual-drug explanation is more plausible when concentrations are quantifiable in a later period, are associated with the treatment given in the preceding period, and fit the known persistence of the analyte. For example, if concentrations are repeatedly higher before dosing after treatment T than after treatment R, while first-period samples are below the LLOQ, that pattern is consistent with residual drug following T. It does not, by itself, prove the explanation.
Other patterns may point toward other explanations. A single unexpected value, especially without a matching sequence pattern, warrants checking sample identity, collection and processing records, assay performance, and any relevant exposure history. Similar elevations across sequences might be consistent with a common analytical or procedural issue, but could also have other causes. Predose concentrations from an endogenous analyte require particular care: the earlier material on baseline correction for endogenous compounds addresses that distinct problem.
A sensible review therefore describes the observed values before making a causal claim. It also distinguishes evidence consistent with a mechanism from confirmation of that mechanism. A prespecified rule can govern what happens next even when the cause remains uncertain.
Reviewing Measurements Under Prespecified Rules
Before applying any decision rule, check that the underlying observations are interpretable. The sample must truly have been collected before dosing; the assay must measure the intended analyte; and the LLOQ and units must be correctly reported. If the value is reported as below the LLOQ, do not silently replace it with zero or with the LLOQ. Use the protocol’s specified handling convention for such results.
The protocol may specify a confirmation step, a concentration threshold, a rescheduling action, or a rule for evaluating the affected period. These are study-specific instructions, not universal thresholds. Apply the stated rule to the stated quantity—for instance, a comparison with the LLOQ is not interchangeable with a comparison to a fraction of a postdose concentration. Do not invent a new cutoff after seeing the data or apply a rule intended for a different analyte or study.
Check the sample timing, analyte, units, LLOQ, and available assay or sample-quality information.
Summarize results by period and sequence, and compare later-period concentrations with the preceding treatment and relevant postdose concentrations.
Follow the protocol’s prespecified threshold and action. Record the result and action without treating the threshold as a general scientific boundary.
State whether the finding may affect observed exposure or the interpretation of the comparison, and identify any uncertainty that remains.
Worked Example: Applying a Hypothetical Predose Rule
Hypothetical study data. The protocol states that a predose sample above the LLOQ must be confirmed in a second sample before dosing. If the confirmed concentration exceeds 10% of the planned period’s anticipated \(C_{\max}\), dosing is deferred and the period is rescheduled. This is an example rule only, not a universal standard. The LLOQ is 0.05 mg/L. Before a later period, the first sample is 0.12 mg/L and the confirmation sample is 0.11 mg/L; the anticipated \(C_{\max}\) is 0.80 mg/L.
Both samples are above the LLOQ, and the repeat result confirms that the finding is quantifiable under this hypothetical rule. The comparison with the anticipated peak is \(0.11/0.80 \times 100\% = 13.75\%\), or 13.8% to three significant figures. Since \(13.8\% > 10\%\), the stated rule calls for deferring dosing and rescheduling the period.
The conclusion is limited to that rule: the period meets the hypothetical criterion for rescheduling. The calculation does not establish why the drug is present, and it does not justify extending the 10% cutoff to another study.
How Residual Drug Can Contribute to Exposure
A simple model helps illustrate one possible consequence. Suppose a residual concentration \(C_{\mathrm{pre}}\) is present at dosing time and then declines by first-order elimination, with elimination rate constant \(k\). Assume the residual analyte is the same as the measured analyte, its decline follows the model over the interval, and concentrations from the residual and current doses add linearly. These are modeling assumptions for the calculation, not conclusions available from a predose value alone.
The residual component’s contribution to area under the concentration-time curve over \(0\) to \(T\) follows by integration:
The units are concentration multiplied by time. Under the stated linear-additivity assumption, that component contributes to the observed concentration profile following the new dose. The equation is an illustration, not a general correction formula: actual study data may not satisfy its assumptions, and a predose value does not reveal the entire residual concentration-time curve.
Worked Example: Illustrative Residual AUC
Hypothetical values. Let the predose concentration be 0.12 mg/L, the residual drug’s half-life be 8.0 h, and the interval of interest be \(T=24\) h. Under the first-order model, \(k=\ln(2)/8.0=0.08664\ \mathrm{h}^{-1}\). The half-life, analyte identity, and linear-additivity assumptions are taken as valid for this illustration.
Substitute into the integrated expression: \(0.12\ \mathrm{mg/L}\times(1-e^{-0.08664\ \mathrm{h}^{-1}\times24\ \mathrm{h}})/(0.08664\ \mathrm{h}^{-1})\). Here \(e^{-0.08664\times24}\) is approximately 0.125, so the arithmetic is \(0.12\times(1-0.125)/0.08664=1.21\ \mathrm{mg\cdot h/L}\).
This is the model-based residual component’s AUC over 24 h. If a hypothetical observed total AUC were 40.0 mg·h/L, the calculated component would be about \(1.21/40.0\times100\%=3.03\%\) of that value. That comparison is illustrative, not a general correction or a basis for removing a subject. The observed total, the estimated residual component, and the protocol’s analysis rule must not be conflated.
Sequence Patterns, Diagnostics, and Alternatives
For a standard two-treatment, two-period crossover, the treatment received in period 1 is also the preceding treatment for a subject’s period 2 predose sample. Comparing period 2 predose results by preceding treatment can therefore be informative. First-period predose results provide a useful reference for identifying whether measurable concentrations were present before any study treatment. Describe the values and their distribution; with small samples, a formal statistical test may be imprecise and should not replace inspection of the data and protocol.
Worked Example: Comparing Predose Results by Preceding Treatment
Hypothetical data. The assay LLOQ is 0.02 mg/L. First-period predose samples are below the LLOQ in both sequences. In period 2, subjects who received T in period 1 have predose concentrations 0.14, 0.11, and 0.16 mg/L; subjects who received R in period 1 have concentrations 0.02, 0.03, and 0.01 mg/L. The 0.01 mg/L result is below the LLOQ.
Among subjects preceded by T, the mean of the measured period 2 values is \((0.14+0.11+0.16)/3=0.137\) mg/L. Among those preceded by R, the two quantifiable values average \((0.02+0.03)/2=0.025\) mg/L; the third result is below quantification and is not treated as zero for this comparison. The higher quantifiable values after T, together with first-period results below the LLOQ, are consistent with a residual-drug pattern following T.
This small hypothetical dataset cannot establish the source of the concentrations. The near-LLOQ results and the unquantified result also limit a numerical comparison. The appropriate conclusion is descriptive: the pattern warrants review under the protocol’s rule, not that residual drug has been proven or that the period should automatically be excluded.
| Approach | What it can contribute | Important limitation |
|---|---|---|
| Apply the protocol’s predose threshold or action rule | A consistent, prespecified operational decision | A threshold is study-specific and does not identify the cause of a value |
| Describe values by period and preceding treatment | A direct view of whether concentrations align with a residual-drug pattern | Small samples, values near the LLOQ, and other imbalances can make patterns uncertain |
| Check sample and assay records | May identify a handling, identity, or analytical explanation | A clean audit does not prove that a value represents residual drug |
| Model a residual concentration-time component | Quantifies a possible contribution if the model assumptions are reasonable | Requires assumptions about kinetics and additivity that a predose measurement alone cannot verify |
A diagnostic review should also consider whether the sampling time was recorded correctly, whether repeated or confirmatory measurements agree, whether values cluster near the LLOQ, and whether apparent elevations are concentrated in one sequence or period. If measurements are inconsistent with the expected assay range or show an isolated extreme, investigate data and sample integrity rather than relying on a smooth pharmacokinetic explanation. These checks help assess plausibility; they do not replace the prespecified rule.
Consequences and Common Pitfalls
A measurable predose concentration may affect the observed exposure after dosing, including AUC and potentially the observed maximum concentration. If its size or frequency differs by preceding treatment, the resulting treatment-period profiles may not reflect only the current dose. The direction and size of any effect on a treatment comparison depend on the concentrations, timing, and study design; a detectable predose value does not automatically imply a material bias or a failed bioequivalence conclusion.
Do not subtract the predose concentration from every postdose measurement or “correct” AUC by the illustrative equation unless the method is justified and prespecified. Such actions can impose an unverified kinetic model and change the defined endpoint. Likewise, do not infer that all quantifiable period 2 concentrations are carryover, or that a below-LLOQ result establishes absence of drug. State what the data show, apply the protocol, and make uncertainty explicit. Any decisions about analysis sets belong to the study’s prespecified procedures rather than an improvised response to a surprising value.
Check Your Understanding
For each question, consider the interpretation and the arithmetic carefully. Solutions follow the questions.
- A predose result is reported as below an LLOQ of 0.04 mg/L. Can it be entered as a measured concentration of zero? Explain.
- A protocol says to confirm a predose value above 0.03 mg/L and defer dosing if the confirmed value exceeds 5% of anticipated \(C_{\max}\). The first result is 0.08 mg/L, the confirmed result is 0.07 mg/L, and anticipated \(C_{\max}\) is 1.0 mg/L. What action follows?
- Predose values are quantifiable in period 2 mainly after treatment T, while first-period values are below the LLOQ. What does this pattern support, and what does it not prove?
- Under first-order elimination with \(C_{\mathrm{pre}}=0.05\) mg/L, half-life 10 h, and \(T=10\) h, calculate the residual component’s AUC over the interval.
Solutions. (1) No. “Below the LLOQ” means the assay did not quantify the concentration reliably below 0.04 mg/L; it is not a measured zero. Follow the protocol’s handling convention. (2) The first result exceeds 0.03 mg/L, so confirmation is required. The confirmed value is also above 0.03 mg/L. Its percentage of anticipated \(C_{\max}\) is \(0.07/1.0\times100\%=7.0\%\). Since \(7.0\%>5\%\), defer dosing under this hypothetical rule. (3) The association with the preceding treatment is consistent with residual drug after T. It does not prove the cause, exclude sample or assay problems, or by itself dictate an action outside the protocol. (4) The rate constant is \(k=\ln(2)/10=0.06931\ \mathrm{h}^{-1}\). Thus \(\mathrm{AUC}_{\mathrm{res},0-10}=0.05(1-e^{-0.06931\times10})/0.06931=0.05(1-0.5)/0.06931=0.361\ \mathrm{mg\cdot h/L}\), to three significant figures.