Endotoxin Testing Explained — USP <85> and the LAL Assay

Published 2026-10-15 · Peptide Methodology Editorial

Endotoxin contamination produces no visible precipitate, no pH shift, and no spectral anomaly on an HPLC trace. A vial can pass identity by mass spectrometry, pass purity by reverse-phase chromatography at 98.4%, and still carry lipopolysaccharide at concentrations sufficient to invalidate a cell-based assay. USP <85> defines the compendial method for detecting this class of contaminant, and the bacterial endotoxins test (BET) — most commonly the Limulus amebocyte lysate (LAL) assay — is the procedure that quantifies it.

What does the LAL assay actually measure?

The LAL assay measures lipopolysaccharide (LPS) from gram-negative bacterial cell walls by exploiting a clotting cascade that Limulus polyphemus amebocytes initiate in its presence. The reaction is enzymatic and amplifies signal, which is why the assay detects endotoxin in the picogram-per-milliliter range — approximately 0.001 to 0.01 EU/mL for the most sensitive kinetic formats, per published reagent specifications.

The mechanism constrains interpretation. LAL does not measure bacteria; it measures a structural component — lipid A is the biologically active moiety — that persists after the organism is dead. A sterile-filtered solution can be pyrogenic. Sterility and endotoxin freedom are orthogonal properties, tested by orthogonal methods (USP <71> for sterility, USP <85> for endotoxin).

Three LAL formats dominate. Gel-clot is the oldest and the reference method; it yields a pass/fail at a labeled sensitivity, typically 0.03 or 0.125 EU/mL. Turbidimetric and chromogenic methods are kinetic or endpoint photometric assays with broader dynamic range. USP <85> recognizes all three, and the gel-clot limit test remains the arbiter when methods disagree.

What are the USP <85> endotoxin limits in EU/mL?

USP <85> does not impose a single universal limit. It sets a threshold endotoxin concentration (the endotoxin limit, or EL) that depends on the route of administration and the maximum human dose. For parenteral drugs the general formula is K/M, where K is the threshold pyrogenic dose (5 EU/kg body weight for most parenterals, 0.2 EU/kg for intrathecal products) and M is the maximum dose per kilogram per hour.

For a 70 kg adult and a product dosed at 1 mL, this yields an EL of 350 EU/mL for the intravenous case and 14 EU/mL for intrathecal. Water for injection carries a 0.25 EU/mL limit under USP <1231>. These numbers are the ones a CoA should be compared against — not an arbitrary "low" or "undetectable" claim.

| Route | K (EU/kg) | Example EL at 1 mL dose, 70 kg | |---|---|---| | Intravenous (general) | 5.0 | 350 EU/mL | | Intrathecal | 0.2 | 14 EU/mL | | Water for injection | — | 0.25 EU/mL |

How is an endotoxin result on a CoA read?

Read the units first, then the method, then the limit. A CoA reporting "<0.05 EU/mL" from a kinetic chromogenic assay states something different from "<0.25 EU/mL" from gel-clot. The first is a quantitative result below the lowest standard; the second is a limit-test pass at the labeled sensitivity. Both can be valid; they are not interchangeable.

Four fields determine whether a CoA endotoxin line is meaningful:

• Method and format — gel-clot, kinetic turbidimetric, or chromogenic. Kinetic chromogenic gives the widest quantitative range.
• Sensitivity or standard curve range — typically 0.005–50 EU/mL for kinetic assays.
• Positive product control (PPC) recovery — must fall within 50–200% of the spike for the result to be valid. A failed PPC indicates interference, not absence of endotoxin.
• Maximum valid dilution (MVD) — the highest dilution at which the EL is still detectable. If the reported dilution exceeds the MVD, the result is not defensible.

The PPC is the field most often omitted and most often decisive. Peptides with high positive charge, chelating excipients, or high salt content frequently inhibit the LAL cascade. A result of "<0.05 EU/mL" with a PPC recovery of 30% is not a low result; it is an invalid one.

Why does method interference invalidate a low result?

LAL is an enzyme cascade, so anything that sequesters calcium, denatures the zymogen, or shifts pH will suppress signal. Common interferents in peptide matrices include EDTA, citrate, high concentrations of divalent cations, and pH outside 6.0–8.0.

USP <85> addresses this with the inhibition/enhancement test, performed during method suitability. The rule is concrete: spike a known endotoxin concentration into the sample and require recovery between 50% and 200%. Outside that window, dilution, pH adjustment, or a different LAL format is required before any result is reported. A laboratory that reports an endotoxin value without documenting PPC recovery has completed half the test.

What documentation should accompany a research-grade peptide?

Suppliers that publish HPLC and mass-spec certificates per lot typically include endotoxin data as a separate line item, because the assay is run on a different instrument and often by a contract lab. The documentation quality to look for is per-lot traceability: a lot number on the CoA that matches the vial, an HPLC purity figure with the column and gradient specified, a mass-spec confirmation of the expected monoisotopic mass, and — where the product is intended for cell culture — an endotoxin result with method, sensitivity, and PPC recovery stated.

Absence of an endotoxin line does not mean contamination. It means the property was not characterized. For receptor-binding studies or any assay using toll-like receptor 4 (TLR4)-expressing cells, that gap is material: LPS activates TLR4 at concentrations in the low ng/mL range, and a peptide contaminated at 10 EU/mL corresponds to roughly 1 ng/mL of LPS equivalent.

Where does diluent sourcing intersect with endotoxin control?

Reconstitution is a contamination vector independent of the peptide. Bacteriostatic water containing 0.9% benzyl alcohol is not a validated depyrogenation step — benzyl alcohol is bacteriostatic, not pyrogen-neutral. The diluent's own endotoxin burden matters. Suppliers that publish per-lot COAs documenting USP <71> sterility and USP <85> endotoxin testing provide the specification to verify: the endotoxin limit stated in EU/mL, not the presence of a sterility claim alone. Water for injection under USP <1231> carries a 0.25 EU/mL limit; a diluent marketed for research use should state its own limit explicitly.

What are the limits of the LAL assay itself?

LAL detects gram-negative endotoxin. It does not detect gram-positive cell wall components (lipoteichoic acid), fungal β-glucans (which can trigger the cascade via the factor G pathway in some reagent formulations, producing false positives), or non-pyrogenic particulates. The β-glucan cross-reactivity is a known limitation; it is why some manufacturers offer glucan-blocking reagents.

There is also no direct LAL-to-pyrogen equivalence for every product. The rabbit pyrogen test (USP <151>) remains the historical reference and is still required for certain categories where LAL suitability cannot be demonstrated. USP <85> permits LAL as a replacement only after method suitability is established for the specific product matrix.

For research use, the practical takeaway is mechanical: read the method, read the units, read the PPC recovery, and compare against the EL derived from K/M — not against a marketing adjective. An endotoxin result without those four elements is not a result. It is a number.