Bioanalytical Testing and PK/TK Assays: A Guide to ICH M10

A pharmacokinetic curve looks simple on a slide. Behind every point on it sits a validated bioanalytical method, a chain of custody for a frozen sample, and a set of acceptance criteria that regulators now apply the same way in Washington, Amsterdam and Ottawa. Get the bioanalysis wrong and the curve means nothing. This guide covers how PK and TK assays work, what ICH M10 requires, and what sponsors should check before they hand samples to a lab.

PK and TK: Same Science, Different Purpose

Pharmacokinetics measures absorption, distribution, metabolism and excretion. In practice that means drug concentration in plasma or serum over time, and the parameters derived from it: C_max, T_max, AUC and half-life. PK studies run in animals during discovery and in humans from Phase 1 onward.

Toxicokinetics uses the same tools for a narrower job. FDA’s ICH S3A guidance defines TK as the generation of pharmacokinetic data to assess systemic exposure in nonclinical toxicity studies. Those data help toxicologists interpret findings and judge their relevance to clinical safety. TK answers a basic question: at the dose that caused liver changes in the rat, how much drug was actually in the animal? That exposure margin then shapes the human starting dose.

ICH M10: One Global Standard

Before M10, sponsors juggled separate FDA and EMA bioanalytical guidances with small but real differences. The ICH M10 guideline was endorsed by ICH regulatory members on May 24, 2022, and FDA published it as final guidance in November 2022 (FDA, 2022). Commentators described the FDA adoption as completing international acceptance of a unified approach (BioProcess Online, 2023).

M10 applies to methods that measure drug and metabolite concentrations in studies that support regulatory decisions. That includes pivotal nonclinical TK and PK studies and clinical studies. It does not cover biomarker assays or immunogenicity assays, which need their own fit-for-purpose approaches.

Chromatographic vs Ligand-Binding Assays

M10 splits bioanalysis into two families. Chromatographic methods, almost always LC-MS/MS today, separate the analyte physically and detect it by mass. They are the default for small molecules, peptides and many oligonucleotides. Ligand-binding assays (LBAs) rely on an antibody or other reagent that captures the analyte. Formats include classic plate ELISA, electrochemiluminescence platforms such as MSD, and microfluidic systems such as Gyrolab. LBAs are the workhorse for monoclonal antibodies and other therapeutic proteins.

Because LBAs depend on biological reagents, they are more variable. M10 sets wider limits to match (FDA, 2022).

Parameter	Chromatographic (LC-MS/MS)	Ligand-Binding Assay
Accuracy	±15% (±20% at LLOQ)	±20% (±25% at LLOQ/ULOQ)
Precision (CV)	≤15% (≤20% at LLOQ)	≤20% (≤25% at LLOQ/ULOQ)
Total error	Not specified	≤30% (≤40% at LLOQ/ULOQ)
Selectivity sources	At least 6	At least 10
Calibration levels	At least 6 plus blank and zero	At least 6 plus blank
ISR acceptance	±20% for 2/3 of repeats	±30% for 2/3 of repeats

Validation Parameters That Matter

A full validation under M10 works through a defined set of characteristics. The ones that cause the most trouble in audits are below.

• Selectivity. The method must distinguish the analyte from matrix components. For chromatography, interference in blanks should be no more than 20% of the LLOQ response and 5% of the internal standard response.
• Specificity. The method must tell the analyte apart from related substances such as metabolites, isomers and impurities.
• Accuracy and precision. Chromatographic QCs are tested in at least three runs over at least two days. LBAs need at least six runs over two or more days.
• Calibration curve. At least six concentration levels, including the LLOQ and ULOQ.
• Stability. Freeze-thaw, bench-top, long-term, processed sample, stock solution and, where relevant, whole blood stability (Health Canada, 2023). Mean QC values must stay within ±15% (chromatographic) or ±20% (LBA) of nominal.

Bioanalytical validation is not the same as validating a release test or verifying a compendial method. The matrix is biological, the concentrations are low, and the samples are irreplaceable. Our post on compendial analytical method verification explains how the QC-lab side of the equation differs.

Incurred Sample Reanalysis

Spiked QCs behave well. Real study samples sometimes do not, because of metabolites, protein binding or sample handling. Incurred sample reanalysis (ISR) catches this. Under M10, labs reanalyze 10% of samples up to 1,000, plus 5% of samples beyond that. At least two-thirds of repeats must fall within ±20% for chromatographic assays or ±30% for LBAs. In nonclinical work, ISR should generally be done at least once per species (FDA, 2022). Sponsors should make sure enough sample volume is kept to allow it.

“A validated method proves the assay works on spiked samples. ISR proves it works on yours.”

TK in GLP Toxicology and the Rise of Microsampling

TK samples are usually drawn from the main toxicology animals or from separate satellite groups. In rodents, blood volumes are small, so sponsors have historically added satellite animals just for TK. That raises animal numbers and means exposure is measured in animals that were not assessed for toxicity.

The ICH S3A Q&A on microsampling, endorsed in November 2017, addresses this. It defines microsampling as collecting a very small amount of blood, typically 50 µL or less (FDA, 2018). The benefits include fewer or no satellite animals, especially in mice, and exposure data from the same animals that show the toxicity.

Biologics vs Small Molecules

Small-molecule bioanalysis is mature. LC-MS/MS methods with stable-isotope internal standards are sensitive, specific and fast to transfer. The main risks are matrix effects, unstable metabolites and carryover.

Biologics are harder. LBA reagents such as capture antibodies and reference standards must be characterized and controlled across the life of the program, because a reagent lot change can shift results. Anti-drug antibodies can interfere with PK measurement, so the PK assay and the immunogenicity strategy need to be designed together. Hybrid LBA-LC-MS/MS methods, which use immunocapture followed by mass spectrometry, are an option where LBA selectivity falls short.

Most sponsors outsource this work. Global Market Insights valued the bioanalytical testing services market at USD 4.2 billion in 2024, with small molecules at 56% of that total. It forecasts growth to USD 9.3 billion by 2034, an 8.2% CAGR (GMI, 2025).

What This Means for Sponsors and Quality Leads

Bioanalysis sits at the junction of nonclinical, clinical and quality. When it fails, the cost lands everywhere at once: rejected runs, repeated studies and questions in the IND or marketing application. M10 has made the rules more predictable. It has not made them easier.

• Audit the lab against M10. Check that validation reports cover every required parameter, not just accuracy and precision.
• Control critical reagents. For LBAs, ask how capture antibodies and reference standards are qualified and tracked.
• Protect sample integrity. Confirm stability covers the real storage time and shipping conditions.
• Plan ISR and microsampling early. Build both into the toxicology protocol, not after the fact.

Terraforme Biosciences helps sponsors select and audit bioanalytical CROs, review validation packages against ICH M10, and align TK data with the IND. Learn more about our regulatory guidance services.