HG/T 6248-2023: Reading China's Bilirubin Standard
Share
Bilirubin is the one bile-derived product where a Chinese COA almost always carries an HG/T number, and most buyers never look it up. The number on the paper is HG/T 6248-2023, an industry standard for bilirubin extracted from bile, and the document behind it is shorter and narrower than most buyers assume. Reading it correctly takes ten minutes and changes what you ask for next.
A standard of this kind is not a pharmacopoeial monograph. It is not a substitute for one, and it does not claim to be. It is the reference point that a Chinese plant and its industrial customer actually contract against, which makes it worth decoding before you accept or reject a COA that cites it.
What the standard covers
HG/T 6248-2023 defines bilirubin prepared from bile as the starting material: saponification, extraction, refining and drying. That scope line matters more than it looks. The standard describes a route, and a route determines the impurity profile that its limits were written around.
The identity block is fixed and verifiable: molecular formula C33H34N4O6, relative molecular mass 584.66, and an appearance requirement of an orange to reddish-brown crystalline powder. The normative references pull in Chinese method standards for loss on drying, sampling, water grade, rounding rules and HPLC practice, which is why the methods cited on the COA are GB/T numbers rather than pharmacopoeial ones.
The requirements table is two rows
This is the part buyers find surprising. Clause 4 fixes exactly two measured properties, and each carries its method.
| Item | Limit | Method |
|---|---|---|
| Loss on drying | Not more than 2.0% | GB/T 6284, 0.5 g sample |
| Bilirubin content (w/w) | Not less than 95.0% | HPLC, external standard, UV detection |
Read the third column twice. Both methods are Chinese national method standards, and both differ from how a pharmacopoeial monograph would frame the same material. And notice what the table does not contain: no related-substances limits, no residual solvents, no heavy metals, no microbial limits. A buyer who needs any of those rows supplies them from a pharmacopoeial framework or writes them into a contract specification. The standard will not do it for you.
The light warning is part of the standard
Clause 5 contains a sentence that does real work: the substance is unstable on exposure to light, and standard and sample handling must be protected from it. The HPLC method itself is written around amber or shaded handling conditions.
That single sentence carries commercial weight. A bilirubin COA produced without light protection describes a different sample than the one that arrived at your warehouse. For what that means in purchasing practice, see the companion piece on buying bilirubin around its light instability.
What the standard cannot do for you
Three boundaries follow directly from the document's construction. First, it sets no impurity limits beyond the two rows above, so a 95.0% assay floor is the whole chemical story it tells. Second, it establishes nothing about suitability for any regulated use; that judgement belongs to a pharmacopoeia and a regulator, and bilirubin has no pharmacopoeial monograph to fall back on. Third, its authority is contractual: it binds where the parties cite it, and nowhere else.
The absence of a monograph is worth restating. For UDCA a buyer can appeal to EP or USP limits; for bilirubin the appeal stops at HG/T 6248-2023. That makes the standard simultaneously more important as a shared reference and weaker as a quality floor. The distance between those two facts is exactly the specification work nobody else will do for you. A structured way to close it is described in supplier spec versus in-house spec gap analysis.
The same logic applies to the companion standard for biosynthetic ursodeoxycholic acid, which is decoded in the HG/T 6247-2023 reading guide. Both standards mark the same line: route-specific methods and floors, with everything beyond them left to the contract.
So when a COA cites HG/T 6248-2023, the number tells you the method source, the two rows tell you the floor, and the silence between the rows tells you what to write next into your own specification.
The remaining chapters of the standard are procedural, and they reward a quick read. The inspection rules chapter describes sampling per GB/T 6679, judgment of results with the rounding conventions of GB/T 8170, and how a batch is classified as conforming or re-checked. The marking and packaging chapters close the loop on what the drum itself must say. None of this is glamorous, and all of it explains lines that appear on the COA without obvious origin.
Three reconciliations follow for a buyer working with a Chinese bilirubin COA.
Method versus expectation. The standard's assay is HPLC with UV detection against an external standard, which is methodologically familiar, but the pharmacopoeial framing you may expect elsewhere does not exist here. There is no monograph to cross-cite, which a search of the EDQM CEP database makes visible in seconds: certified-substance coverage stops where monographs stop, and bilirubin is outside both.
Identity versus assumption. The formula, molecular mass and appearance block are the cheapest check in the whole document, and they are verifiable against public chemistry records such as PubChem or the NIST Chemistry WebBook before any sample is shipped. A COA whose identity block disagrees with the public record has answered a question the buyer did not know to ask.
Floor versus need. The two-row table is a floor for industrial trade, not a specification for a use. Where the application needs related substances, residual solvents or metals, the standard is silent and the contract is not.
The reading order that works: identity block first, the two-row table second, the light clause third, and the procedural chapters last, skimmed for the method numbers that will appear on every future COA. Ten minutes spent this way turns the standard from a citation into a working instrument, and it is the same reading order that pays off for the companion UDCA standard in the series.
Frequently asked questions
What is HG/T 6248-2023?
An industry standard issued through China's formal standardisation system for bilirubin extracted from bile by saponification, extraction, refining and drying. It fixes identity, appearance and two measured properties with their methods, plus rules for testing, marking, packaging, transport and storage.
What limits does the standard actually set?
Two: loss on drying not more than 2.0% by GB/T 6284, and bilirubin content not less than 95.0% by HPLC with UV detection against an external standard. There are no related-substances, residual solvent, heavy metal or microbiological limits.
Is HG/T 6248-2023 a pharmacopoeial monograph?
No. It is a chemical-sector standard, contractual in force when cited. Bilirubin has no EP or USP monograph, so the standard is the only public reference a Chinese COA can cite, which makes its two-row floor the entire shared basis.
Why does the standard mention light?
Because the substance is unstable on exposure to light, and the test methods require protected handling. A COA produced without light protection describes a different sample than the one shipped, so the warning has commercial consequences, not just laboratory ones.
Can I rely on the 95.0% assay floor alone?
Not if your use depends on impurities, residual solvents or metals. The standard is silent on all of them. Anything you need beyond the two rows must come from your own specification or a pharmacopoeial framework you impose contractually.
How does it relate to HG/T 6247-2023?
Same series logic, different substance. HG/T 6247 covers biosynthetic ursodeoxycholic acid and has six requirement rows; HG/T 6248 covers extracted bilirubin and has two. Both are route-specific and both leave everything beyond their rows to the contract.