Diagnosing Route-Dependent Absorption From Plasma Profiles

Reading a Plasma Profile as Input Plus Disposition

A plasma concentration-time profile is the result of both drug entering the systemic circulation and drug leaving it. A delayed peak after an injection, a slow rise during dermal administration, or a long terminal tail after inhalation may suggest route-dependent input. None of these features, by itself, identifies a unique absorption process. Distribution and elimination also shape the measured profile, and different input and disposition combinations can produce similar curves.

The useful diagnostic question is therefore not simply “What does this curve look like?” but “Which features are consistent with the proposed input, what independent information constrains disposition, and what remains ambiguous?” The earlier material on route input functions and depot mass balance provides the model basis for describing input. This tutorial uses plasma patterns to assess those assumptions, rather than to infer a unique route mechanism from curve shape alone.

For a simple diagnostic benchmark, consider a linear, one-compartment model with constant volume \(V\), first-order elimination rate constant \(k\), and systemic input \(R_{\mathrm{sys}}(t)\):

This benchmark assumes that concentration is proportional to amount in one well-mixed compartment, elimination is linear over the concentration range, and \(V\) and \(k\) do not change during observation. It is a diagnostic simplification, not a claim that a PBPK model has only one compartment. In a distributed model, tissue exchange and other processes may contribute additional phases to plasma concentration.

What Onset, Peak, and Tail Can Suggest

The onset is the early appearance of measurable drug in plasma. An early rise is consistent with rapid systemic entry, but observed onset also depends on the sampling schedule, assay sensitivity, and any delay before drug is available for absorption. A late first detectable concentration does not establish a physiological lag if early samples are sparse or concentrations are below quantification.

The peak is described by both its height and its timing. A rapid input often produces an earlier, sharper peak than a slower input, all else being equal. However, peak timing and height also depend on dose, systemic availability, distribution, and elimination. A lower peak alone is not evidence of lower total absorption: the same systemic amount spread over a longer time can yield a lower peak with similar exposure.

The terminal slope describes the late decline in concentration, often summarized by the slope of log concentration against time over a selected terminal interval. It is an observed profile feature. In a one-compartment IV bolus model with no continuing input, it equals the elimination rate constant \(k\). Following an extravascular dose, ongoing absorption can instead control the late decline. In multicompartment settings, distribution can also contribute to apparent phases, so a terminal slope should not automatically be labelled an elimination rate.

Route-specific expectations help frame the assessment, but are not definitive signatures. Injection-site precipitation or dissolution-limited release can delay or extend systemic input after an injection. Dermal permeation and a skin reservoir can produce delayed or prolonged input, with finite-dose depletion changing input over time. After inhalation, regional deposition and subsequent clearance or absorption affect the available input; deposited dose and systemic appearance are not interchangeable. These processes were addressed in the earlier material on injection, dermal, and pulmonary inputs. Here the diagnostic task is to check whether the plasma pattern is compatible with a proposed input and to identify what the plasma profile cannot establish.

A Mass-Balance Check on Exposure

Integrating the benchmark model gives a useful exposure result. Define \(A_{\mathrm{sys}}\) as the total systemic amount entering during the integration interval, so \(A_{\mathrm{sys}}=\int_0^\infty R_{\mathrm{sys}}(t)\,dt\), when all input is included. Define \(\mathrm{AUC}_{0-\infty}=\int_0^\infty C(t)\,dt\). Integration of the differential equation yields:

If concentration returns to zero at infinity, then \(C(\infty)=0\). If the observation interval starts with no drug in the compartment, \(C(0)=0\), and all systemic input is included in \(A_{\mathrm{sys}}\), rearrangement gives:

The initial condition matters. If \(C(0)\ne 0\), then the same integration gives \(\mathrm{AUC}_{0-\infty}=(A_{\mathrm{sys}}/V+C(0))/k\), provided \(C(\infty)=0\). For an IV bolus, one may either represent the dose as an instantaneous input starting from pre-dose \(C(0)=0\), or start the post-bolus interval at \(C(0)=D/V\) with no further input. The conventions must not be mixed.

With \(CL=Vk\) in this benchmark, exposure is \(A_{\mathrm{sys}}/CL\). Thus, if two routes have the same linear disposition and comparable conditions, different dose-normalized AUCs can indicate different systemic availability. This exposure comparison does not specify the shape or mechanism of the input. It also depends on the assumed dose, the completeness of AUC measurement or extrapolation, and the assumption that clearance is shared.

Why a Route Terminal Slope May Not Be Elimination

A compact example shows the source of ambiguity. Suppose absorption into the systemic circulation follows first-order input from a depot, with absorption rate constant \(k_a\), fraction reaching the systemic circulation \(F\), and administered amount \(D\). Under the one-compartment assumptions above, with \(C(0)=0\), the resulting concentration for \(k_a\ne k\) is:

This equation contains both an elimination exponential and an absorption exponential. At late times, the slower exponential decays less quickly and dominates. Therefore the apparent terminal rate is the smaller of \(k\) and \(k_a\), under this specific model. When \(k_a<k\), the late slope reflects absorption-limited input, a pattern often called flip-flop kinetics. The profile alone may then fail to identify the elimination rate.

This calculation depends on the assumed one-compartment, first-order model. A delayed peak or a long terminal tail may be compatible with slow absorption, but the model does not prove that the actual input is first-order or that a particular depot mechanism is present. An IV reference, independent disposition information, dense early samples, and route-specific measurements can constrain interpretation. If those data are absent, several parameter combinations may fit the same plasma observations.

Comparing Routes and Diagnosing Model Mismatch

A practical comparison uses the same subject or a justified shared disposition model, matched or appropriately accounted-for doses, and profiles sampled over the periods needed to characterize onset, peak, and tail. An IV profile is especially informative because it removes an extravascular absorption phase from the input, although distribution and elimination still shape the profile. Route comparisons should distinguish administered dose from the amount that actually enters systemic circulation.

Patterns that warrant checking include a predicted onset that precedes every observed concentration despite dense, sensitive early sampling; a predicted sharp peak when repeated observations show a broad rise; a terminal slope that changes substantially with the chosen fitting interval; or an AUC that cannot be reconciled with the assumed systemic input and disposition. These are warnings, not automatic proof that the route model is wrong. Sampling gaps, assay limits, distribution phases, and uncertainty in dose delivery can produce apparent discrepancies.

For inhaled profiles, for example, a small or delayed plasma signal could be consistent with limited systemic input, but the plasma curve alone does not tell whether the cause was regional deposition, airway clearance, slow absorption, or measurement timing. For injection and dermal profiles, prolonged input may similarly resemble slow elimination. A useful diagnostic workflow is to plot concentration on both linear and log scales, inspect early and late sampling coverage, compare against an IV disposition reference when available, and test whether the proposed input preserves mass balance and reproduces the observed profile without implausible parameters.

Limits, Alternatives, and a Common Pitfall

The main alternative to interpreting shape alone is to use independent evidence: an IV profile or other justified estimate of disposition, route-specific measurements of deposition or depot behavior, and adequately timed concentration samples. These sources can improve separation of input from disposition. A joint model of multiple routes may also be useful when its shared parameters and route-specific assumptions are justified. By contrast, fitting a flexible absorption function to one extravascular plasma profile can describe the curve without uniquely identifying the biological input process.

The diagnostic benchmark has important limitations. Real PBPK systems may have multiple tissue compartments, changing blood flows, route-specific barriers, nonlinearity, or time-varying input. Plasma is one observation site and may not capture local tissue exposure. Sparse sampling can miss a short-lived peak or obscure a lag; terminal AUC extrapolation can be unstable when the tail is poorly measured. In those settings, a one-compartment slope or exposure calculation is a check on assumptions, not a substitute for the full model.

Common mistake: “The extravascular profile has a longer terminal half-life than the IV profile, so elimination must have become slower.” This conclusion confuses the observed terminal slope with elimination. If absorption continues slowly, the input can govern the tail even when elimination is unchanged. Check the model’s assumptions and compare with independent IV disposition information before interpreting the slope as an elimination parameter.