What AUC Measures
A concentration–time profile records how drug concentration changes after administration. The area under that profile, or AUC, summarizes concentration accumulated over a specified period. It is an integrated quantity: each small time interval contributes its concentration multiplied by the duration of that interval. AUC therefore reflects both the level of concentration and how long that level persists.
AUC is not a concentration at one instant, nor is it automatically a measure of the total amount of drug in the body. Its interpretation depends on the concentration scale, the time interval, and the pharmacokinetic setting. In bioavailability comparisons, for example, AUC is an exposure metric; it is not itself a fraction absorbed or a bioavailability percentage.
When a concentration curve is available as a mathematical function, the integral may be evaluated directly. In a pharmacokinetic study, however, concentrations are usually measured only at selected times. The observed values are samples from a profile that changes between measurements. A numerical AUC is therefore an estimate of the area beneath the underlying curve, constructed from those observations and an assumption or rule about what happens between them.
From a Continuous Curve to a Numerical Estimate
The basic idea of numerical integration is to divide the interval into smaller pieces and add approximate areas. For a partition \(t_1,t_2,\ldots,t_m\), each piece has a duration and a representative concentration. Adding the products gives an approximation to the total area. More, closely spaced observations can generally describe changes in the profile better, although measurement error and the choice of interpolation still matter.
One simple illustration uses rectangles. On each interval, a rectangle can use the concentration at its left endpoint or its right endpoint. These are called left- and right-endpoint sums. They are deliberately elementary approximations, not a claim that concentration remains constant between samples. When the profile is monotone over an interval, the two sums can also show a range within which the area lies. The usual numerical procedures can use the observations differently; the central point here is that all such estimates depend on the sampled curve and the specified interval.
Worked Example: AUC From a Known Concentration Function
Suppose a hypothetical concentration profile over 0 to 6 hours is \(C(t)=12e^{-0.2t}\) mg/L. Find the AUC over that interval.
Integrating the function gives $$ \operatorname{AUC}_{0-6}=\int_0^6 12e^{-0.2t}\,dt =\frac{12}{0.2}\left(1-e^{-0.2(6)}\right). $$
Since \(e^{-1.2}\approx 0.3012\), the calculation is $$ 60(1-0.3012)=60(0.6988)=41.928\ \text{mg·h/L}. $$
Thus the AUC is approximately 41.93 mg·h/L. The units follow because mg/L is multiplied by hours. This exact-function calculation is useful for understanding the meaning of area; with real sparse observations, the concentration function is not known in this form.
Worked Example: Endpoint Rectangles for Sampled Concentrations
Consider a hypothetical post-peak profile sampled at 0, 2, 4, and 6 hours. The concentrations are 10, 8, 5, and 2 mg/L, respectively. Assume the concentration declines monotonically between each pair of observations. Use endpoint rectangles to illustrate the area approximation over 0 to 6 hours.
Each interval is 2 hours wide. The left-endpoint sum uses the first concentration on each interval: $$ \operatorname{AUC}_{\mathrm{left}} =2(10+8+5)=46\ \text{mg·h/L}. $$
The right-endpoint sum uses the second concentration on each interval: $$ \operatorname{AUC}_{\mathrm{right}} =2(8+5+2)=30\ \text{mg·h/L}. $$
Because the profile is assumed to decline throughout each interval, the left rectangles lie above the curve and the right rectangles lie below it. The area is therefore between 30 and 46 mg·h/L under that monotonicity assumption. This wide range shows the limitation of coarse sampling and crude rectangles; it does not mean that either sum is a precise AUC estimate. If the curve rises or changes direction within a sampling interval, this bracketing argument need not hold.
The Time Window Defines the Metric
An AUC value always belongs to a specified interval. Changing an endpoint changes the area being summarized, even when the concentration profile itself is unchanged. A longer interval includes more of the curve; a shorter interval focuses on exposure during that window. Thus two AUC values with different bounds need not answer the same scientific question and should not be treated as interchangeable.
A window may be chosen to describe exposure over a dosing interval, to focus on a clinically relevant period, or to align with a prespecified comparison. The choice should be driven by the objective, not selected after seeing which interval produces a preferred result. The interval also determines which observations and which between-observation behavior matter to the numerical estimate.
Worked Example: Comparing AUCs Over Different Windows
Suppose a hypothetical profile is \(C(t)=10e^{-0.5t}\) mg/L. Compare the AUC from 0 to 2 hours with the AUC from 0 to 6 hours, and identify the exposure in the later part of the interval.
For 0 to 2 hours, $$ \operatorname{AUC}_{0-2} =\int_0^2 10e^{-0.5t}\,dt =20(1-e^{-1}) \approx 20(1-0.3679) =12.64\ \text{mg·h/L}. $$
For 0 to 6 hours, $$ \operatorname{AUC}_{0-6} =20(1-e^{-3}) \approx 20(1-0.04979) =19.00\ \text{mg·h/L}. $$
The additional area from 2 to 6 hours is the difference: $$ \operatorname{AUC}_{2-6} =19.00-12.64=6.36\ \text{mg·h/L}. $$
The 0-to-2-hour value summarizes early exposure, whereas the 0-to-6-hour value includes both early and later exposure. The latter is larger because the profile is positive over the added interval. Neither value is universally the “correct” AUC: each answers a question tied to its bounds.
When a concentration profile is nonnegative, extending an interval while keeping its starting time fixed cannot decrease the true AUC: the added area is nonnegative. This provides a useful check on calculations. It does not imply that every wider-window estimate is more scientifically relevant, or that a numerical estimate will always increase if different measurements or estimation procedures are used.
Assumptions, Diagnostics, and Alternatives
The integral definition presumes a concentration profile that is integrable over the stated interval. For numerical estimation, one also needs observed concentrations that adequately characterize the profile and a rule for representing unobserved times. The numerical calculation does not recover concentrations that were never measured without assumptions. Smooth changes between samples may be plausible in many settings, but a rapid peak, a changing slope, or sparse sampling can make the area less certain.
Several checks help reveal problems. Confirm that time points are correctly ordered, that each concentration is paired with the right time, and that elapsed-time units are consistent. Inspect the concentration–time plot: an unusually wide gap near a rapid rise or decline is more concerning than a similar gap along a slowly changing segment. Check whether the estimate changes substantially when the sampling grid or interpolation choice changes. A calculation that is inconsistent with the observed profile—for example, an area with the wrong units or an AUC that decreases after adding a positive time interval—signals an error or a mismatch in definitions.
Numerical integration from observed concentrations is often practical because it relies on measured data and does not require choosing a full pharmacokinetic model. A model-based approach instead specifies a mathematical concentration curve and integrates that curve. It can provide a smooth profile between observations, but its result depends on whether the model form and its assumptions describe the data adequately. Neither approach removes uncertainty: numerical methods depend on sampling and interpolation, while model-based integration depends on model specification and parameter estimates.
AUC also has limitations as a summary. It compresses a time-varying profile into one number, so profiles with different timing can have the same area. It cannot, by itself, show when the concentration was highest or how quickly it rose. Moreover, AUC is not inherently adjusted for dose; dose normalization is a separate operation whose appropriateness depends on the comparison and pharmacokinetic conditions. These distinctions matter when interpreting exposure or comparing formulations.
The interval is just as important as the numerical method. Reporting only “AUC” without bounds leaves the estimand unclear. State the start and end times, the units, and whether the value is based on direct integration of a specified function or numerical estimation from sampled concentrations. These details allow a reader to understand what exposure the number represents and to judge whether the sampling schedule supports it.
Check Your Understanding
Answer each question, then compare with the solutions.
- If concentration is measured in mg/L and time in minutes, what are the units of AUC?
- For a nonnegative concentration profile, how does extending the end time while keeping the start time fixed affect the true AUC?
- Why can left- and right-endpoint sums bracket the area for a decreasing profile but fail to do so when the profile changes direction?
- A numerical AUC is reported without start or end times. What interpretive information is missing?
- Why does summing observed concentrations alone not produce an AUC?
Solutions. (1) The units are mg·min/L, because concentration is multiplied by time. (2) The AUC cannot decrease: the extended interval adds the integral of a nonnegative concentration. (3) For a monotone decreasing curve, left-endpoint rectangles are above the curve and right-endpoint rectangles are below it on each interval; a direction change can reverse those relationships within an interval. (4) The time window, and therefore the exposure being summarized, is unknown. (5) A sum of concentrations omits the elapsed time between observations and does not specify how to represent the curve between them.