3.2 Test Article Characterization, Formulation Analysis, Stability & Homogeneity

Why characterization is a design-and-execution topic

DABT Domain I.A.3 asks whether the test article in the animal room is the same chemical, biologic, or mixture the protocol named, at the concentration the dose table assumes. OpenExamPrep independent teaching for this section treats characterization as both a scientific design choice and a GLP execution control. If the high-dose diet is 70% of target because of poor mixing, every no-observed-adverse-effect-level discussion in the report is fiction — even if clinical signs were recorded perfectly.

21 CFR 58.105 requires that identity, strength, purity, and composition, or other characteristics that appropriately define the test or control article, be determined for each batch and documented. Methods of synthesis, fabrication, or derivation are documented by the sponsor or the testing facility. Each storage container is labeled with name (or code), batch, expiration if any, and storage conditions, and is assigned to that article for the duration of the study. For studies lasting more than four weeks, reserve samples from each batch are kept for the archive retention period in 21 CFR 58.195.

Identity answers “is this the intended molecule, serotype, or material?” Purity answers “how much of the batch is that article versus process impurities, degradants, residual solvents, or elemental residues?” Strength (assay/content) answers “how much active is present per unit mass or volume?” A certificate that lists a chemical name without a batch-specific assay is not characterization.

Common analytical methods

Method choice follows the article, not a single “GLP instrument.”

• High-performance liquid chromatography (HPLC), often with ultraviolet detection, is the workhorse for small-molecule assay, chromatographic purity, and formulated concentration in vehicles or diet extracts.
• Mass spectrometry (MS), alone or as liquid chromatography–MS, confirms identity (molecular ion, fragments) and identifies impurities when a peak must be more than an area-percent unknown.
• Enzyme-linked immunosorbent assay (ELISA) and related ligand-binding methods are common for protein, monoclonal antibody, and some peptide test articles, and for concentration of biologics in formulation when a chromatographic assay is not meaningful.

Whatever the platform, the method must be suitable for the matrix you actually dose: corn oil, 0.5% methylcellulose, powdered diet, drinking water, or a lyophilized biologic reconstitution. An HPLC method validated in methanol is not automatically valid in chow extract.

Formulation analysis, homogeneity, and stability

Neat test article on a shelf is not what most animals receive. 21 CFR 58.113 covers mixtures of articles with carriers. For each test or control article mixed with a carrier, appropriate analytical methods must determine uniformity of the mixture and, periodically, the concentration of the article in the mixture, and must determine stability of the article in the mixture as required by the conditions of the study.

Translate that into laboratory practice:

• Formulated concentration verification asks whether the gavage suspension, diet, or water solution matches the protocol milligram-per-kilogram or parts-per-million target within predefined acceptance limits (often something like 90–110% of target, set in the method SOP — do not invent a universal percentage as a regulation).
• Homogeneity asks whether top, middle, and bottom of a mixer, or aliquots from a stirred suspension, agree. Dietary studies fail here when a hydrophobic powder stratifies in meal or when a sticky vehicle is inadequately mixed.
• Stability under storage asks whether the bulk article and the reserved formulation remain within specification in the freezer, refrigerator, or ambient cabinet named on the label.
• Stability under dosing conditions is the trap: room-temperature diet hoppers, amber gavage bottles on a cart for four hours, or drinking-water lines at animal-room temperature. A six-month freezer certificate does not prove 24-hour stability in an open hopper.

Vehicles and excipients are part of the dose. Corn oil, polyethylene glycol, dimethyl sulfoxide (usually in vitro or small fractions in vivo), carboxymethylcellulose, and cyclodextrins each change absorption and local tolerance. The control group receives the same vehicle at the same volume. If the vehicle itself is toxic at the volume used, you have designed a confounder, not a negative control.

Solubility drives the dosage form. A compound that is insoluble in aqueous vehicles becomes a suspension; suspensions demand mixing instructions, resuspension checks, and homogeneity samples. Precipitation in a syringe between the pharmacy and the last cage is a concentration deviation. For inhalation or dermal studies, the “formulation” may be an atmosphere or a applied cream — the same logic applies: measured concentration, spatial uniformity, and stability over the exposure window.

Exam trap: a 90-day dietary study arrives with a supplier certificate of analysis (CoA) from receipt of the neat powder. No mixed-diet samples are collected. That CoA may support identity and bulk purity at one time point. It does not satisfy 21 CFR 58.113 uniformity or periodic concentration, and it does not show stability in chow under animal-room conditions. Dose-analysis samples from the mixed diet are not optional decoration.

ICH impurity thinking without dumping the tables

Toxicologists are asked to connect formulation quality to impurity qualification, especially for pharmaceuticals. Teach the classification idea and that numeric thresholds depend on dose and duration; do not memorize every row of every ICH table as if it were a GLP clause.

ICH Q3A(R2) (organic impurities in new drug substances) sets three escalating expectations as an impurity’s level rises: reporting (list it), identification (know the structure), and qualification (justify biological safety). Those thresholds are expressed as a percentage of the drug substance, sometimes with a milligram-per-day cap, and they tighten when the maximum daily dose is high because the patient then swallows more impurity mass. Qualification of an ordinary, non-mutagenic impurity may use literature, in silico reasoning, or toxicology studies — including the impurity as it occurred in batches used in nonclinical or clinical work. Q3A is not a license to skip test-article assay in a rodent diet study.

ICH Q3C addresses residual solvents. Class 1 solvents (benzene is the teaching example) have unacceptable toxicities (carcinogenicity, environmental harm, or similar) and should be avoided unless their use is strongly justified and tightly limited. Class 2 solvents are limited using a permitted daily exposure (PDE) because they are less dire but still not unlimited (methanol and acetonitrile are typical teaching examples). Class 3 solvents have low toxic potential; a higher default such as 50 mg/day or a 5000 ppm option is the usual teaching frame, still not “unlimited.” Duration of treatment and daily dose still matter because PDE is an intake, not a magic percentage.

ICH Q3D addresses elemental impurities (catalysts, contaminants from equipment or water). Elements are grouped by toxicity and likelihood of occurrence. Class 1 elements — arsenic, cadmium, mercury, and lead — are human toxicants that require evaluation across products. Other classes (for example 2A, 2B, and 3 in Q3D) modulate how routinely you assay versus how you can justify omission. Limits are PDEs that differ by route (oral, parenteral, inhalation). A metal that is acceptable orally may be unacceptable inhaled. Q3D is not a residual-solvent table and is not an excuse to skip formulated-diet analysis.

ICH M7 addresses DNA-reactive (mutagenic) impurities. Ordinary Q3A qualification percentages do not apply once mutagenicity is the concern. Impurities are binned conceptually as known mutagenic carcinogens (compound-specific limits), known mutagens without carcinogenicity data, alerting structures that need data or control to the threshold of toxicological concern, alerting structures shared with a non-mutagenic drug substance, and structures without alerts. For many Class 2/3 mutagenic impurities, the default lifetime acceptable intake is 1.5 µg/day (theoretical excess cancer risk on the order of 1 in 100,000). Less-than-lifetime treatment allows higher daily intakes that keep cumulative exposure comparable: 120 µg/day for less than 1 month, 20 µg/day for more than 1 to 12 months, 10 µg/day for more than 1 to 10 years, and 1.5 µg/day beyond 10 years to lifetime. Cohort-of-concern structures (certain high-potency mutagenic carcinogens, including many N-nitrosamines) need compound-specific assessment, not the default 1.5 µg/day TTC. The teaching point for DABT: qualification and control limits depend on dose and duration, and mutagenic impurities are a different decision tree from garden-variety related substances.

Putting the pieces on a 90-day study

Imagine a hydrophobic small molecule mixed weekly into rodent meal at three concentrations. Before day 1 you need batch identity (HPLC/MS), purity (including a rational look at residual solvents and, if relevant, elemental residues), and strength. Each weekly mix needs homogeneity samples. Across 13 weeks you need periodic concentration results on diet pulled from hoppers, not only from the mixer immediately after blending. You need stability covering the interval the diet sits in the hopper. If week 8 assay is 62% of target, the Study Director documents a deviation, stops using that mix, and the report must discuss actual consumed dose — not the protocol’s intended parts per million. That is test-article quality as toxicology, not as paperwork.