Almost every line on a peptide certificate of analysis measures molecules. Purity, water content, counterion, residual solvent — all chemistry, all static. The microbial line is different. It reports on something that was alive and could, given the right conditions, multiply. Research-grade material of this kind is supplied for research use only, and its microbial specification describes the conditions the powder passed through rather than the molecule inside it.
That makes bioburden one of the more informative entries on a certificate, and one of the most frequently misread. What follows covers what the number counts, how USP <61> produces it, the method-suitability step that decides whether the result means anything, where the pass/fail limits come from, and the three related tests that get confused with one another.
Why a Count of Living Things Belongs on a Chemistry Certificate
Because the rest of the analytical panel can't see it — and because the regulation requires it.
What bioburden actually is
Bioburden is the population of viable microorganisms living on or in a material that has not been sterilized, reported in colony forming units — CFU — per gram or per millilitre.
The unit is worth pausing on, because it isn't a count of cells. A colony forming unit is a colony that appeared on a plate, and a single colony can grow from one isolated cell or from a clump of fifty that travelled together. So CFU is a deliberately conservative floor on the real population. Not a census of it.
The industry used to call the whole exercise Microbial Limits Testing, or MLT. The modern chapters split the work in two: USP <61> counts how much is there, USP <62> asks whether specific named organisms are among it.
Why it is not optional
In the United States this sits in the Code of Federal Regulations, not a guidance document: 21 CFR 211.110(a)(6) requires bioburden measurement during drug manufacturing, and ISO 11737 governs the equivalent measurement on medical devices before sterilization.
It's also a structurally different kind of contaminant from the others the certificate tracks. Elemental impurity testing under ICH Q3D measures a fixed quantity — the lead in a lot does not become more lead in storage. A microbial population can grow, if water content and formulation permit, which is why FDA's guidance on non-sterile manufacturing counts water activity and preservative system as part of the control strategy rather than background detail.
Inside USP <61>: Two Numbers, Three Methods, One Week
In short: a bacterial count, a fungal count, three permitted ways of arriving at them, and an incubation clock that puts the answer most of a week out.
TAMC and TYMC
A <61> result is two numbers, not one. The total aerobic microbial count (TAMC) covers bacteria that grow in air, cultured on soybean-casein digest agar — the medium also sold as tryptic soy agar. The total combined yeasts and moulds count (TYMC) covers fungi, cultured on Sabouraud dextrose agar.
They're separated for a practical reason: bacteria and fungi want different nutrients, temperatures, and amounts of time, so a single plate asked to count both would under-report whichever it suited less well. The split is also diagnostic. A dry powder picking up environmental mould spores tells a different story from one picking up a water-borne bacterium.
The three enumeration approaches
The count can be produced three ways, and the laboratory has to justify which it used.
- Membrane filtration. The sample passes through a 0.45 micrometre filter, which is then laid onto the agar and incubated. The method of choice for liquids and low-count materials, because it concentrates organisms out of a larger volume and rinses interfering matrix away first.
- Pour plate, or plate count. The sample is mixed into molten agar in the dish and incubated there. Fewer steps — but the matrix stays in contact with the growing colonies throughout.
- Most probable number. A statistical estimate from sets of inoculated tubes, reserved for products with very low expected bioburden where the other two are impractical.
Sample matrix drives the choice, not operator preference — a constraint that turns out to be the hinge of the whole test.