What an Estimate Above 100% Means
An absolute bioavailability estimate above 100% can look like a physical impossibility. Under the usual definition, the fraction of an administered dose reaching the systemic circulation as unchanged drug cannot exceed 1. Yet an estimate is calculated from observed exposure data, and those data are subject to variability, measurement error, and assumptions. The result can therefore exceed 100% even when the underlying fraction does not.
The first task is to identify what the number represents. An absolute bioavailability estimate compares systemic exposure after an extravascular dose with exposure after an intravenous (IV) dose, adjusting for dose. It is not a direct count of drug molecules that entered circulation. A value over 100% is a signal to check the calculation, data, and comparison conditions—not a reason to automatically discard the result or to claim that more than the administered dose reached the circulation.
The equation is an estimate of the defined quantity only when the comparison is appropriate. In particular, the pharmacokinetics must support dose normalization: exposure must be proportional to dose over the compared conditions, or the doses and conditions must otherwise support a valid comparison. Clearance and other relevant disposition processes must be comparable. IV administration provides the reference in which the drug enters systemic circulation directly. The measured analyte, dose units, and AUC units must also be consistent.
A value above 100% is not automatically problematic for relative bioavailability. That quantity compares exposure from one extravascular product or condition with another; its ratio can exceed 100% because the numerator can be larger than the reference. The physical bound applies to the fraction in an absolute bioavailability estimand, not to every exposure ratio.
Why the Dose-Normalized Ratio Has a Physical Bound
A short derivation clarifies what the bound depends on. In a linear, time-invariant pharmacokinetic model, total exposure after an IV dose is dose divided by clearance. If an extravascular dose contributes fraction \(F\) to systemic circulation, its exposure is that fraction of the corresponding IV exposure:
Dividing each AUC by its dose removes the dose terms. Taking the ratio cancels the common clearance:
Since \(F\) is a fraction of the administered dose, it cannot exceed 1 under this model and definition. The derivation also shows why an observed ratio above 1 points to either ordinary estimation variability or a failure in the data or assumptions used to interpret the ratio as \(F\). The calculation itself does not enforce the physical bound.
Worked Example: A Plausible Estimate Slightly Above 100%
In a hypothetical comparison, an extravascular dose of 75 mg yields an AUC of 63.0 mg·h/L. An IV dose of 30 mg yields an AUC of 22.5 mg·h/L. Calculate the absolute bioavailability estimate and interpret it.
First calculate dose-normalized AUC for each route:
The estimate is
The arithmetic is consistent, and the dose-normalized units cancel. The estimate is 12 percentage points above the physical bound for true absolute bioavailability. This single result does not establish that the true fraction exceeds 1. It may reflect sampling variability, measurement uncertainty, or a problem with the data or assumptions. The next step is to audit those possibilities rather than reinterpret 112% as a literal fraction.
A Structured Check of an Estimate Above 100%
An effective review starts with simple, verifiable issues and then moves to the assumptions behind the comparison. Preserve the original result while checking it; do not alter values merely to force the estimate below 100%.
Confirm which AUC and administered dose belong to each route. Check dose units, dilution factors, and whether the entered dose is the amount actually administered rather than the amount prepared or dispensed.
Verify that both AUCs refer to the same analyte and concentration units. Review calibration, dilution, sample handling, assay performance, and any transcription or sample-identification issues that could selectively raise one exposure or lower the other.
Check the concentration-time records and the chosen exposure interval. Sparse or mistimed samples, a missed terminal phase, and an unstable extrapolated tail can affect estimated exposure. A problem in either route can shift the ratio in either direction.
Ask whether dose proportionality and comparable disposition are credible, and whether the IV and extravascular data are otherwise comparable. When nonlinear pharmacokinetics breaks dose-normalized comparisons, the ratio may not estimate absolute bioavailability.
Look at the individual or replicate exposure data and the design, not just the final ratio. A modest excess may be consistent with ordinary variability; a large or systematic excess warrants a more intensive data and model review.
Sampling problems do not all have the same direction. If exposure after one route is incompletely captured, its AUC may be underestimated and the ratio may move down or up depending on which route is affected. An unreliable terminal extrapolation can also change the ratio if its contribution differs between routes. Thus, “more samples” is not by itself a diagnosis: inspect where the samples fall relative to the concentration-time profiles and how much of each reported AUC depends on estimation beyond observed data.
Worked Example: Finding a Dose-Normalization Error
A hypothetical report lists an extravascular AUC of 48.0 mg·h/L after 100 mg and an IV AUC of 30.0 mg·h/L after 50 mg. A reviewer compares the raw AUCs and reports \(48.0/30.0 \times 100=160\%\). Check the calculation.
The raw AUCs cannot be compared directly because the doses differ. Dose-normalized exposure is
Therefore,
The reported 160% is a calculation error caused by omitting dose normalization. The corrected point estimate is 80.0%, assuming the AUCs and doses are correctly recorded and the pharmacokinetic comparison is valid. This example illustrates why a surprising result should be recalculated from the underlying values before being interpreted.
Variability, Assay Issues, and Model Assumptions
Even with correct arithmetic, an estimate is based on measurements from a finite set of observations. Biological differences between participants, within-person variability, and measurement error can make the observed ratio depart from the underlying value. A point estimate just above 100% may therefore be compatible with a true fraction below 1. It is not, by itself, proof of an assay failure or proof that the study’s assumptions hold.
A useful diagnostic is to inspect the underlying exposure pairs rather than only the final summary. A consistent elevation across participants could suggest a systematic difference in dose recording, assay performance, sampling, or model conditions. A few high ratios among otherwise mixed results may instead reflect variability or influential observations. Neither pattern is conclusive on its own: the individual data must be checked against records, design, and assay quality.
Assay review should be targeted to errors capable of distorting the comparison. Examples include a unit conversion applied to only one route, an incorrect dilution factor, a calibration issue, or a sample assigned to the wrong route or participant. If a result is corrected, the correction should be supported by records or valid reanalysis. A plausible-looking value after an undocumented adjustment is not a valid resolution.
Dose normalization has a specific limitation. The derivation assumes a linear relationship between dose and exposure under comparable conditions. If clearance, absorption, or another relevant process changes with dose or concentration, then dividing AUC by dose need not produce a common value. The earlier material on nonlinear pharmacokinetics explains why such a ratio can fail as a comparison. In that setting, an estimate above 100% may indicate that the ratio is not estimating the intended \(F\), rather than that a physical fraction exceeded its bound.
Worked Example: An Assay Audit Changes the Estimate
In a hypothetical study, the extravascular dose is 60 mg with AUC 52.8 mg·h/L; the IV dose is 30 mg with AUC 24.0 mg·h/L. Calculate the initial estimate. Then suppose a documented assay review finds that the IV AUC was underreported and verifies a corrected value of 26.4 mg·h/L.
Using the original results, the normalized exposures are \(52.8/60=0.880\ \mathrm{h/L}\) and \(24.0/30=0.800\ \mathrm{h/L}\). Thus,
With the verified IV AUC correction, its normalized exposure is \(26.4/30=0.880\ \mathrm{h/L}\). The revised calculation is
The arithmetic shows how an error in the reference exposure can inflate the estimate. The revised result is justified here only because the correction is stipulated to be documented and verified. Without evidence for that correction, replacing the original value with 100% would be unjustified.
How to Report and Compare the Result
A report should distinguish the calculated estimate from the scientific interpretation. State the point estimate, the route-specific doses and exposure measures used, and any uncertainty information available from the design and analysis. Describe important data checks and limitations. Do not claim that the true bioavailability is above 100% merely because the point estimate is; equally, do not silently cap the estimate at 100%, because that hides what the data and calculation produced.
The main alternatives are not competing formulas so much as different diagnostic views. Recalculating the dose-normalized ratio checks arithmetic. Reviewing individual exposures helps reveal variability or influential observations. Auditing concentration data and assay records investigates measurement problems. Reassessing the pharmacokinetic assumptions determines whether the ratio has the intended interpretation at all. Each addresses a different possible cause, so one should not substitute for the others.
A limitation of any diagnosis based only on the point estimate is that it cannot separate random variability from systematic error. A value of 101% and a value of 160% are both above the bound, but their size alone does not establish the cause. Interpretation requires the study’s uncertainty, data quality, and pharmacokinetic context. The correct conclusion may be that the estimate is a noisy but interpretable result, that it is unreliable pending investigation, or that the comparison does not estimate absolute bioavailability under the conditions used.
Common mistake: “Absolute bioavailability cannot exceed 100%, so a result of 112% must be a calculation error.” The bound applies to the underlying fraction under the definition’s conditions, not necessarily to a finite-sample estimate. Recalculate and audit the data, then assess variability and assumptions. Conversely, variability is not a blanket excuse: a demonstrable dose, assay, or sampling error must be addressed rather than described away as noise.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- For an extravascular dose of 40 mg with AUC 36.0 mg·h/L and an IV dose of 20 mg with AUC 20.0 mg·h/L, calculate the absolute bioavailability estimate. What does a result above 100% establish?
- Why is it incorrect to compare the raw AUCs when the extravascular and IV doses are unequal?
- Name two types of data or assay problems that could inflate an absolute bioavailability estimate.
- Can a relative bioavailability estimate legitimately exceed 100%? Explain why this differs from absolute bioavailability.
Solutions. (1) The normalized exposures are \(36.0/40=0.900\ \mathrm{h/L}\) and \(20.0/20=1.00\ \mathrm{h/L}\). The estimate is \(0.900/1.00\times100=90.0\%\), so it is not above 100%. In general, a point estimate above 100% does not establish that the true absolute fraction exceeds 1. (2) Raw AUC reflects both dose and systemic availability; divide each AUC by its administered dose before taking the ratio. (3) Examples include an incorrect dose or dilution factor, inconsistent concentration units, a sample assigned to the wrong route, or an assay calibration problem affecting one exposure. (4) Yes. A relative estimate is a ratio between extravascular products or conditions, and the test product can produce more exposure than its reference. It is not a fraction of an administered dose reaching systemic circulation, so the absolute-fraction bound does not apply.