Why Single-Dose Predictions Can Fail
Single-dose pharmacokinetics can provide useful information about a drug’s disposition, but it does not always predict exposure during repeated administration. Extrapolation is most straightforward when pharmacokinetics are linear and time-invariant: the fraction reaching the systemic circulation is stable, and clearance does not change with concentration or treatment duration. Under those conditions, contributions from successive doses can be combined using linear superposition.
Three departures matter in multiple-dose studies. Saturable clearance means that elimination capacity does not rise in proportion to concentration. Time-dependent clearance means that the drug’s elimination capacity changes over the course of treatment, for example through enzyme induction or inhibition. Dose-dependent bioavailability means that the fraction of an administered dose reaching systemic circulation changes with dose. Each can make exposure depart from the prediction based on a single dose, but the mechanisms and evidence needed to distinguish them differ.
These properties must be assessed over the doses and treatment durations relevant to the study. A drug may behave approximately linearly at a low concentration but become nonlinear at a higher one. It may also have stable clearance after a treatment-related change has settled, even though clearance was not stable during the transition.
What Steady-State Mass Balance Does—and Does Not—Tell Us
At periodic steady state, the amount entering the body over one interval must equal the amount eliminated over that interval, provided there is no net change in body amount between the interval’s start and end. For an extravascular dose \(D\) with a stable bioavailability \(F\), and complete systemic input of that dose over the interval, the systemic input is \(FD\). Thus, the total elimination over the interval is also \(FD\).
If elimination is linear with a constant clearance \(CL\), its rate at concentration \(C(t)\) is \(CL\,C(t)\). Integrating that rate over the dosing interval \(\tau\) gives the amount eliminated. This leads to the familiar relationship between steady-state interval AUC and dose:
This result requires periodic steady state, constant clearance over the interval, and stable bioavailability and systemic input. It is not, in general, the AUC over the first finite interval after a dose. For example, drug remaining in the body at the interval’s end can contribute to later concentrations, while the first interval begins without the residual drug present at periodic steady state.
With concentration-dependent elimination, the corresponding mass balance still applies, but the linear relationship between elimination rate and concentration does not. For saturable elimination following Michaelis–Menten kinetics, the elimination rate is \(V_{\max}C/(K_m+C)\). At periodic steady state, the input balances the integral of that rate—not a constant clearance multiplied by AUC.
The concentration-time profile is needed to evaluate this integral. In general, one cannot replace it with \(CL\times AUC_\tau\) using a single constant \(CL\).
Worked Example: Recognizing Saturable Clearance
Consider a hypothetical drug whose elimination rate follows Michaelis–Menten kinetics with \(V_{\max}=12\) mg/h and \(K_m=3\) mg/L. Compare the instantaneous elimination rate and the apparent clearance, defined here as elimination rate divided by concentration, at concentrations of 3 mg/L and 12 mg/L.
At \(C=3\) mg/L:
At \(C=12\) mg/L:
Concentration increased fourfold, but the instantaneous elimination rate increased only from 6 to 9.6 mg/h, a 1.6-fold increase. The apparent clearance decreased from 2 to 0.8 L/h. This calculation illustrates why a constant-clearance extrapolation may fail; these instantaneous values alone do not determine the multiple-dose AUC. At periodic steady state, the interval input must instead balance the integral of the concentration-dependent elimination rate.
Three Ways Multiple-Dose Exposure Can Depart from Prediction
Saturable clearance is concentration-dependent. When concentrations rise into a range where elimination approaches its capacity, the elimination rate increases less than proportionally with concentration. Consequently, increasing the dose can produce a more-than-proportional increase in exposure. The size of that departure depends on the concentration-time profile, the dosing interval, and the saturation parameters; it cannot be inferred from a single concentration alone.
Time-dependent clearance changes during treatment, even at a given concentration. Enzyme induction can increase clearance over time, whereas inhibition or other treatment-related changes can decrease it. During the transition, successive dosing intervals need not have repeating profiles, so the regimen has not yet reached periodic steady state. A single-dose estimate of clearance may therefore describe neither the early repeated-dose phase nor the eventual stabilized regimen. After the change settles, a periodic steady state may be established, but its clearance can differ from the single-dose value.
Dose-dependent bioavailability changes the systemic input rather than necessarily changing elimination. For example, absorption may become incomplete at a larger dose. If \(F\) falls as \(D\) rises, the systemic dose \(FD\) may increase less than proportionally—or, in some cases, not increase. A dose-proportional prediction that assumes the same \(F\) at both doses will then be misleading even when clearance remains constant.
Worked Example: A Clearance Change After Treatment Begins
Suppose a hypothetical drug is given orally at 120 mg every 12 hours, with stable \(F=0.75\). Assume linear elimination within each stabilized treatment condition. A hypothetical study estimates clearance of 9 L/h before a treatment-related induction and 15 L/h after the induction has settled. Calculate and compare the interval AUC values at periodic steady state under these two conditions.
The assumptions permit \(AUC_\tau=FD/CL\) for each condition separately: the regimen has reached periodic steady state in each, the bioavailability is stable, and clearance is constant across that condition’s interval. The systemic dose per interval is \((0.75)(120\ \text{mg})=90\ \text{mg}\).
Under the stated assumptions, the stabilized interval AUC is 4 mg·h/L lower after induction, a 40% decrease relative to the pre-induction value. This calculation does not mean that an interval AUC measured during the transition can be interpreted using either clearance value. During the transition, clearance is changing and consecutive profiles may not repeat.
Worked Example: Dose-Dependent Bioavailability
Suppose a hypothetical drug has linear, constant clearance of 10 L/h. At a 100 mg dose its bioavailability is 0.80, while at a 200 mg dose it is 0.50. Assume each dose regimen has reached periodic steady state, bioavailability is stable for that regimen, and clearance is the same at both doses. Calculate the interval AUC under each regimen.
At 100 mg, the systemic dose per interval is \((0.80)(100\ \text{mg})=80\) mg. Therefore:
At 200 mg, the systemic dose is \((0.50)(200\ \text{mg})=100\) mg, giving:
The administered dose doubled, but interval AUC rose by a factor of \(10/8=1.25\), or 25%. The difference follows from the lower bioavailability at the higher dose. This comparison applies only under the stated steady-state and constant-clearance assumptions; it does not show that bioavailability is the only possible explanation for a nonproportional exposure pattern.
Evidence, Diagnostics, and Alternative Explanations
A practical assessment compares exposure across doses or treatment durations under controlled, interpretable conditions. Dose-normalized AUC can be informative: under linear pharmacokinetics with dose-independent \(F\) and constant \(CL\), \(AUC_\tau/D\) is constant at periodic steady state. A systematic change in this quantity is a signal of departure from those assumptions, not a diagnosis of a particular mechanism.
Other useful clues include whether concentrations continue to rise or fall over successive intervals, whether trough concentrations change over time on an unchanged regimen, and whether peak-to-trough patterns differ across doses. These observations should be interpreted alongside the sampling schedule and the evidence that periodic steady state was reached. A change in trough alone does not identify the cause: changes in absorption, adherence, dosing time, or the interval can also alter the measured concentration.
The main alternatives are to use a single-dose linear prediction, to compare multiple-dose exposure directly under the regimens of interest, or to fit a mechanistic model that represents concentration-dependent elimination, changing clearance, or dose-dependent input. Single-dose prediction is simplest but relies most heavily on linearity and time-invariance. Direct multiple-dose comparison answers the exposure question for the studied conditions but may not explain the mechanism or predict unstudied regimens. A mechanistic model can represent a proposed process, but its conclusions depend on whether the data support its parameters and assumptions.
No single diagnostic proves nonlinearity. Before attributing departures to saturation, induction, inhibition, or bioavailability, check dose preparation, adherence, meal conditions, sampling times, assay performance, and whether the compared intervals represent comparable treatment states. Limited dose ranges, sparse sampling, and substantial between-subject variability can make distinct mechanisms difficult to separate. Model complexity does not compensate for weak information in the data.
Common Pitfall and Key Takeaway
Wrong reasoning: “The administered dose and bioavailability are known, so \(AUC_\tau=FD/CL\) gives the AUC after the first dose.” This uses a periodic steady-state interval relationship as though it applied to any finite interval. A first-interval AUC depends on the concentration profile and the amount remaining at the interval boundary; the steady-state formula does not supply that information by itself.
Correction: Use \(AUC_\tau=FD/CL\) only when periodic steady state has been reached and the stated constant-clearance and stable-input assumptions hold. If clearance is saturable, use the concentration-dependent elimination model and the full interval mass balance. If clearance changes over time, distinguish transition intervals from a stabilized periodic state. If bioavailability changes with dose, calculate systemic input using the bioavailability applicable to each dose.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- A drug has a stable oral bioavailability of 0.60, a dose of 150 mg every 24 hours, and constant clearance of 9 L/h. What is \(AUC_\tau\) at periodic steady state?
- Why is it generally incorrect to use \(AUC_\tau=FD/CL\) for the first finite dosing interval after a single dose?
- At periodic steady state, a dose increase is followed by a more-than-proportional rise in AUC. Name two different mechanisms that could contribute, and explain why the observation alone does not identify which occurred.
- Clearance increases during the first week of an unchanged regimen. Should a profile measured during that week automatically be treated as a periodic steady-state profile? Explain.
Solutions. (1) The systemic dose per interval is \((0.60)(150\ \text{mg})=90\) mg. Under the stated constant-clearance and periodic-steady-state assumptions, \(AUC_\tau=90\ \text{mg}/(9\ \text{L/h})=10\ \text{mg}\cdot\text{h/L}\). (2) The formula follows from equal body amounts at the start and end of a repeating interval, together with constant clearance. A first finite interval need not start with the residual drug present at periodic steady state, so the relationship does not generally apply. (3) Saturable clearance could cause exposure to rise more than proportionally as concentration increases; dose-dependent bioavailability could also change systemic input, although a falling \(F\) alone would tend to blunt the increase. The observation does not isolate a mechanism because exposure also depends on the concentration-time profile, treatment state, and other study conditions. (4) No. If clearance is still changing, consecutive interval profiles may not repeat. The profile should not be called periodic steady state unless the repeating-profile condition is supported.