Written and maintained by CASRAI Editorial Board
Last updated
Toxicokinetics (TK) is the generation and interpretation of pharmacokinetic data inside a toxicity study — not a separate discipline from pharmacokinetics (PK) so much as PK data collected for a different reason. A researcher running a standard PK study wants to know how a drug behaves in order to design an effective, safe dosing regimen. A toxicologist running a TK component wants to know one specific thing: whether the animals in a GLP toxicity study were actually exposed to the test article at the levels the dosing design intended, and how that exposure related to the toxic effects observed. ICH S3A — Note for Guidance on Toxicokinetics: The Assessment of Systemic Exposure in Toxicity Studies — is the harmonised guideline that defines TK for exactly this purpose and sets the expectations regulators use to judge whether a nonclinical safety package supports moving into or through human trials.
This guide covers what distinguishes TK from PK in intent, the study-design elements a TK component typically includes, how exposure margins get calculated, and where TK data actually lands in a regulatory submission.
Toxicokinetics vs. pharmacokinetics: a difference of purpose, not method
The laboratory techniques — dosing, blood sampling, bioanalysis, calculating Cmax, Tmax, and AUC — are the same whether you call the resulting dataset “PK” or “TK.” What differs is the question being answered:
- Pharmacokinetics characterizes exposure to support an efficacy or dose-finding objective — what dose and regimen will achieve the target therapeutic concentration.
- Toxicokinetics characterizes exposure achieved within a toxicity study to support interpretation of the safety findings — did the animals actually reach the intended systemic exposure, was the dose-exposure relationship linear or did it plateau (saturation), and how does that measured exposure compare with the exposure anticipated or observed in the intended clinical population.
That last comparison is the reason TK exists as a named discipline at all: a toxicology finding at a given dose is only interpretable in the context of the systemic exposure that dose actually produced. Two compounds dosed at the same mg/kg level can produce very different systemic exposures because of differences in absorption, metabolism, or dose-dependent (nonlinear) kinetics — without TK data, a toxicologist reading across species or extrapolating a no-observed-adverse-effect level (NOAEL) to a first-in-human dose has no reliable basis for the comparison. TK is what makes cross-species and cross-study exposure comparison possible; PBPK modeling and formal pharmacometric analysis both depend on having real measured TK data as an input, not a substitute for collecting it.
Typical toxicokinetic study design
TK is rarely a standalone study. In most GLP-compliant nonclinical safety packages, it’s built directly into the repeat-dose toxicity study design (governed procedurally by Good Laboratory Practice under 21 CFR Part 58) using two design features that let exposure data be collected without compromising the toxicity endpoints the main study is measuring.
Satellite animals
Satellite animals are additional animals added to a toxicity study group specifically to support serial blood sampling for TK, separate from the “main study” animals whose survival, clinical pathology, and terminal histopathology are the primary toxicity endpoints. The rationale is direct: repeated blood draws are a stressor and a blood-volume burden, and drawing that volume from the main study animals risks confounding the very toxicity findings the study exists to measure — anemia from sampling can look like (or mask) a genuine hematologic effect of the test article, and handling stress can shift behavior or clinical signs. Satellite animals absorb that burden instead, dosed identically to their main-study cohort but not carried through to the toxicity endpoints. Whether a satellite group is warranted, and how large it needs to be, is a study-design decision that belongs in the protocol, not something to standardize blindly across every study.
Sparse sampling
Sparse (composite) sampling collects only one to a few blood samples per animal, at different timepoints across different animals within the same dose group, and pools the results into a single composite concentration-time profile for that group — rather than drawing a full serial profile from every individual animal. This is the practical mechanism by which TK avoids adding substantially to the animal and blood-volume burden of the toxicity study it’s embedded in. Sparse sampling is explicitly recommended in ICH S3A as the preferred approach for rodent studies in particular, where full serial sampling from a single animal isn’t feasible given blood volume limits.
Timing and dose-response coverage
A TK component typically samples at more than one point in the dosing period — commonly at first dose and again after repeated dosing has reached steady state — because exposure can change over time even at a constant dose, from enzyme induction (autoinduction lowering exposure) or accumulation (raising it). It also needs to characterize the dose-exposure relationship across the full range of doses used in the toxicity study, since a non-dose-proportional (nonlinear) relationship at high doses is itself a finding that affects how the study’s results are interpreted and how a margin gets calculated at each dose level, not just the top one.
Exposure margins: the number a TK study is built to produce
The output that makes TK data actionable for drug development and regulatory review is the exposure margin — the ratio between the systemic exposure (typically AUC, sometimes Cmax) achieved in animals at a toxicologically relevant dose (most often the NOAEL) and the systemic exposure anticipated, or already observed, in humans at the intended clinical dose. A margin comfortably greater than 1 is what supports a conclusion that the nonclinical safety data provide adequate coverage for the human dose being proposed; a narrow or negative margin is a signal that either the clinical dose needs to come down, the nonclinical dose needs to go up, or the program needs additional characterization before it can proceed. ICH S3A frames this exposure comparison — not simple dose (mg/kg) comparison — as the scientifically valid basis for relating animal toxicity findings to human safety, precisely because dose and exposure don’t scale predictably across species.
This is also where TK data intersects most directly with the broader nonclinical package described in toxicology testing more generally: the margin calculation only means something if the underlying exposure numbers are real, measured values from the same study generating the toxicity findings, not assumed or interpolated ones.
How TK data supports dose selection and the regulatory submission
TK data does two distinct jobs across a nonclinical-to-clinical program:
- Selecting doses for subsequent studies. Once a repeat-dose toxicity study’s TK component establishes the dose-exposure relationship (including where it becomes nonlinear), that relationship informs dose selection for the next study in the sequence — a longer-duration toxicity study, a carcinogenicity study, or the pivotal GLP studies intended to support an IND. The goal is choosing doses that will produce a measurable, interpretable exposure margin over the anticipated clinical exposure without simply escalating dose past the point where absorption saturates and exposure stops rising.
- Informing the safety margin presented in the submission. The exposure margins derived from TK data feed directly into the nonclinical overview and written summaries of the submission (eCTD Module 2.4/2.6), where they’re presented alongside the toxicity findings to justify that the proposed human starting dose and dose-escalation plan sit at an adequate distance below exposures associated with adverse findings in animals. This is the same underlying logic ICH M3(R2) — the guideline governing what nonclinical safety studies are needed to support human clinical trials and marketing authorization — expects a sponsor to demonstrate before human dosing begins, and it’s what a reviewer is checking for when reading the Investigator’s Brochure‘s nonclinical summary against the proposed clinical dosing.
TK for tissue distribution specifically — as opposed to plasma/systemic exposure — is addressed by the companion guideline ICH S3B, which covers repeated-dose tissue distribution studies; it’s a related but distinct dataset from the systemic-exposure TK work ICH S3A governs.
IACUC protocol and 3Rs considerations for TK study components
Because a TK component adds live-animal procedures — additional satellite animals, repeated blood draws — to a toxicity study, it belongs in the study’s IACUC protocol as its own justified element, not assumed as an automatic add-on to the base toxicity design. A protocol incorporating a TK component should specify: the scientific justification for the satellite group size and the sampling schedule (why this many animals, why these timepoints); why sparse rather than serial sampling was or wasn’t chosen for this species and study duration; the blood volume limits being observed relative to the animal’s total blood volume and body weight; and the humane endpoints and personnel qualifications that apply to the sampling procedure itself, consistent with the site’s IACUC protocol requirements. Sparse sampling and the satellite-animal design both exist principally as Refinement under the 3Rs framework — reducing the procedural burden and stress placed on any individual animal relative to full serial sampling from every study animal — with a secondary Reduction benefit where composite sampling avoids the need for a separate, larger TK-only cohort. Where TK data is reported alongside toxicity findings in a publication or regulatory submission, ARRIVE 2.0’s reporting items on blinding and randomization of dose-group allocation and sample analysis remain relevant, the same as for the toxicity study itself — TK results are not exempt from the bias controls that apply to the rest of the study just because they’re a supporting dataset rather than the primary endpoint.
FAQ
What’s the difference between toxicokinetics and pharmacokinetics?
The techniques are the same; the purpose differs. Pharmacokinetics characterizes exposure to support dose-finding for efficacy. Toxicokinetics characterizes the exposure actually achieved within a toxicity study, to interpret the toxicity findings against real systemic exposure rather than nominal dose, and to calculate an exposure margin relative to intended human exposure.
What are satellite animals?
Additional animals added to a toxicity study group specifically for TK blood sampling, kept separate from the main-study animals whose survival and pathology are the toxicity study’s primary endpoints — so that the blood-volume burden and handling stress of TK sampling don’t confound the toxicity results.
What is sparse sampling and why is it used?
A design where each animal contributes only one or a few blood samples at different timepoints, pooled across animals in a dose group into a single composite exposure profile, rather than drawing a full serial profile from every animal. ICH S3A recommends it, particularly for rodents, as a way to characterize exposure without exceeding blood-volume limits or adding meaningfully to the animal burden of the study.
What is an exposure margin?
The ratio between the systemic exposure (usually AUC) achieved in animals at a toxicologically relevant dose, often the NOAEL, and the systemic exposure anticipated or observed in humans at the intended clinical dose. It’s the basis ICH S3A specifies for relating animal safety findings to human risk, in place of comparing dose (mg/kg) directly, since exposure doesn’t scale predictably with dose across species.
Is TK data required to support an IND?
TK is expected as part of the nonclinical safety package for programs proceeding to human trials — it’s how a sponsor demonstrates the animals in the pivotal toxicity studies were exposed at levels that produce an interpretable safety margin over the intended clinical exposure. ICH M3(R2) describes the nonclinical study package a program needs before first-in-human dosing, and TK data is what makes the exposure comparisons in that package possible.
