Why the Exposure Metric Matters
Therapeutic drug monitoring does not begin with the question, “Which concentration is easiest to measure?” It begins with a clinical question: what aspect of exposure is expected to relate to the desired effect or to harm? A single concentration may be informative when it represents that aspect of exposure. It may be a poor guide when the effect instead depends on total exposure, a high concentration, or how long concentrations remain above a relevant threshold.
The exposure metric is the quantity used to summarize a concentration–time profile. Choosing it is not the same as setting a target value. A target must be supported by relevant exposure–response evidence and interpreted for the patient and clinical context. The introductory material on defining drug-specific targets and population reference ranges distinguishes those targets from general laboratory ranges; here the focus is on choosing the quantity to which a target should apply.
What the Candidate Metrics Represent
A concentration measured at one time is a point on a concentration–time profile. Its interpretation therefore depends on when it was measured relative to dosing and on the dosing history. A peak concentration summarizes the high point of a profile; a trough summarizes its low point immediately before a subsequent dose under the specified regimen. These summaries can be useful when the magnitude of a high or low concentration has direct pharmacologic relevance, but neither describes the entire profile.
AUC summarizes concentration accumulated over a stated time interval. The interval must be specified: for example, AUC over one dosing interval or over 24 hours. Average concentration over that interval is AUC divided by interval duration. These quantities describe overall exposure, but can conceal differences in the profile’s shape, including different peaks and troughs.
Other metrics preserve a different feature of the profile. Time above a threshold is the duration, or the proportion of a defined interval, for which concentration exceeds that threshold. A partial AUC restricts the area calculation to a specified time window. A concentration-to-dose ratio describes concentration relative to a dosing amount, rather than exposure alone; it can help characterize an individual’s concentration response to a dose, but it is not a substitute for deciding which exposure feature matters clinically.
No one metric is universally best. The choice should follow the pharmacologic hypothesis and evidence: what feature is expected to drive benefit, what feature is expected to drive toxicity, and over what time scale those effects occur? When efficacy and toxicity have different drivers, a single metric may not adequately represent both.
A Practical Selection Framework
Specify whether monitoring is intended to support efficacy, reduce a particular toxicity, or balance both. Name the outcome and the relevant time horizon.
Use pharmacologic reasoning and exposure–response evidence to decide whether the outcome is associated with overall exposure, a high or low concentration, time above a threshold, or another profile feature.
Specify the time interval, concentration threshold if applicable, and whether the quantity is a measured concentration or a model-based summary. A metric without its time definition is incomplete.
Consider dosing history, sampling times, assay performance, pharmacokinetic variability, and whether the available samples identify the desired summary. A clinically relevant metric is not useful if the data cannot estimate it reliably.
Ask whether the selected metric misses an important safety or efficacy feature, and whether changing the regimen could alter profile shape without much changing the chosen summary.
This framework separates biological relevance from practical convenience. A measurable quantity is not automatically a meaningful target, and a theoretically relevant quantity is not automatically estimable from the planned samples. Both conditions are needed for a useful monitoring strategy.
Why AUC Represents Overall Exposure
A simple mass-balance argument explains why AUC is often considered when the clinical effect is related to total systemic exposure. Suppose drug follows linear pharmacokinetics with clearance \(CL\), and dosing is repeated at a stable regimen. At steady state, the amount of drug in the body at the start and end of a dosing interval is the same. Therefore, the amount entering the systemic circulation during the interval equals the amount eliminated.
For an interval of length \(\tau\), if \(C(t)\) is concentration and clearance is constant, the elimination rate is \(CL\,C(t)\). Integrating that rate over the interval gives the amount eliminated. Equating input and elimination yields:
Here \(F\) is bioavailability, and for intravenous dosing \(F=1\). This result assumes linear, time-invariant clearance and steady state over the interval. It shows that under those conditions, AUC captures total exposure for a given dosing interval. It does not show that AUC is the correct clinical metric for every drug: that conclusion requires pharmacologic and exposure–response justification.
Worked Example: Several Metrics From One Hypothetical Profile
Suppose a hypothetical drug is given as an instantaneous intravenous dose every 3 hours. Use a one-compartment model with first-order elimination, \(k=0.400\ \text{h}^{-1}\), and dose divided by volume \(D/V=6.00\ \text{mg/L}\). Assume linear pharmacokinetics and steady state. Calculate the peak, trough, and AUC over one interval.
For repeated instantaneous doses under this model, the steady-state peak just after a dose is the dose increment divided by the fraction remaining after one interval:
The exponent is \(-1.20\), so \(e^{-1.20}=0.301194\). Thus:
During the interval, concentration declines exponentially. The trough immediately before the next dose is:
Because \(CL/V=k\), \(D/CL=(D/V)/(CL/V)=6.00/0.400=15.0\ \text{mg·h/L}\). For this intravenous example, \(F=1\), so the steady-state AUC over the interval is \(15.0\ \text{mg·h/L}\). The average concentration is \(15.0/3.00=5.00\ \text{mg/L}\). The same profile therefore has a peak of \(8.586\ \text{mg/L}\), a trough of \(2.586\ \text{mg/L}\), and an average concentration of \(5.00\ \text{mg/L}\). Each answers a different question; none is interchangeable merely because all come from the same profile.
Worked Decisions: Matching the Metric to the Question
Worked Example: When Time Above a Threshold Is the Proposed Driver
Consider a hypothetical short-acting drug for which a preclinical and clinical evidence package suggests that benefit is related to the time concentration remains above a specified threshold of \(2.00\ \text{mg/L}\). For a simplified one-interval illustration, suppose concentration immediately after a dose is \(8.00\ \text{mg/L}\), elimination is first-order with \(k=0.500\ \text{h}^{-1}\), and the interval is 6.00 hours. Assume no further absorption in the interval and no contribution from prior doses.
The concentration declines as \(C(t)=8.00e^{-0.500t}\). It reaches the threshold when:
Concentration is above the threshold for 2.773 hours of the 6.00-hour interval, or \(2.773/6.00=0.462\), approximately 46.2% of the interval. Under the stated simplified assumptions, time above threshold is directly aligned with the proposed driver. A peak alone would tell us the initial concentration, but not how long it remains above the threshold. The calculation is an illustration, not a validated target: the threshold and its relevance must come from suitable pharmacologic and clinical evidence.
Worked Example: Same AUC, Different Peak
Suppose a hypothetical drug has an efficacy outcome associated with overall exposure, while evidence also suggests that acute toxicity is more closely associated with peak concentration. Two modeled 12-hour regimens have the following summaries: Regimen A has AUC \(60.0\ \text{mg·h/L}\) and peak \(10.0\ \text{mg/L}\); Regimen B has AUC \(60.0\ \text{mg·h/L}\) and peak \(7.50\ \text{mg/L}\). Treat these as hypothetical profile summaries from a model, not as measurements from a real study.
Both regimens have the same average concentration over the interval:
However, their peak-to-average ratios differ: \(10.0/5.00=2.00\) for A and \(7.50/5.00=1.50\) for B. If the stated evidence supports peak concentration as the relevant toxicity driver, A would be less attractive on that safety dimension despite having the same AUC as B. If efficacy is driven by AUC, the AUC summaries do not distinguish the regimens for that outcome. The appropriate comparison therefore depends on the outcome: one metric may address efficacy while another addresses toxicity.
Comparing Approaches and Checking for Failure
| Metric | Best suited to a question about | What it can miss or require |
|---|---|---|
| Concentration at a defined time | Exposure at a clinically meaningful sampling time | Other parts of the profile; correct timing and dosing history |
| Peak concentration | Effects related to a high concentration | Total exposure and duration at high concentrations; adequate sampling or model support |
| Trough concentration | Effects related to the low concentration before the next dose | Peak and cumulative exposure; an interpretable dose and sampling schedule |
| AUC or average concentration | Effects related to overall exposure over a stated interval | Profile shape and time above or below a threshold; sufficient information to estimate the area |
| Time above a threshold | Effects related to duration above a justified concentration threshold | Threshold validity and enough profile information to determine crossing times |
Several warning signs suggest the metric may not be working as intended. A concentration changes substantially when the sample time shifts slightly, but the monitoring interpretation treats it as stable. Two regimens have similar values for the chosen summary but markedly different profiles in a way that plausibly matters. The chosen metric has weak or inconsistent association with the clinical outcome, or a model estimates it imprecisely from the available observations. A result outside the anticipated range can also reflect an inaccurate dose history, missed doses, time-varying pharmacokinetics, or an incorrectly specified model rather than a true change in the underlying exposure–response relationship.
A point concentration is often simpler to obtain than an AUC or time-above-threshold estimate. That convenience can make a fixed-time concentration a practical alternative when it reliably tracks the clinically relevant quantity. AUC can be more informative when total exposure is the proposed driver, but may require multiple observations or a validated model. Peak and trough summaries emphasize profile extremes, but can ignore cumulative exposure. Time above threshold can be conceptually direct when a threshold is supported, yet it may be difficult to estimate from sparse observations. The choice is a trade-off between relevance and reliable estimation, not a contest in which one summary always wins.
Common Pitfall and Limitations
Wrong reasoning: “The laboratory can report a trough, so trough must be the best monitoring target.” This confuses ease of measurement with relevance to the clinical outcome. Correction: first identify the exposure feature supported by pharmacology and exposure–response evidence, then determine whether a trough is a reliable proxy for it under the dosing and sampling conditions. If it is not, a different metric or additional information may be needed.
Metric selection is limited by the evidence available. An observed association does not by itself establish that changing the metric will change the outcome. Relationships may differ across patients, disease severity, concomitant treatments, or time. Model-based metrics also inherit model assumptions and uncertainty. When efficacy and toxicity depend on distinct exposure features, reporting only one metric can obscure a clinically important trade-off. Monitoring plans should therefore specify the metric and interval clearly, state why it is relevant, and recognize what it cannot represent.
Check Your Understanding
Answer each question, then compare with the solutions below.
- A hypothetical drug’s effect is supported by evidence to depend on total exposure over each 24-hour interval. Which candidate metric is most directly aligned, and what essential interval detail must be stated?
- Under the first-order elimination assumptions in the time-above-threshold example, how long does concentration remain above \(4.00\ \text{mg/L}\), starting from \(8.00\ \text{mg/L}\) with \(k=0.500\ \text{h}^{-1}\)?
- Two regimens have equal AUC but substantially different peaks. Give one reason that comparing AUC alone could be insufficient.
- Why does a precisely measured concentration at an unspecified time not necessarily provide an interpretable monitoring result?
Solutions. (1) AUC over the 24-hour interval is most directly aligned with total exposure. The interval must be explicitly defined as 24 hours, along with the dosing context for which it applies. (2) Solve \(4.00=8.00e^{-0.500t}\): \(t=\ln(8.00/4.00)/0.500=\ln(2)/0.500=1.386\ \text{h}\). (3) If a clinically relevant toxicity is associated with high concentrations, equal AUC can conceal a meaningful difference in peak-related risk. (4) Concentration changes over time, so without the time relative to dosing and a defined sampling context, the result cannot be interpreted as a specific point on the profile or compared appropriately with a target.