It sounds like you’re referring to (TR 82), titled "Low Endotoxin Recovery (LER) – A Hidden Challenge in Endotoxin Testing for Biopharmaceuticals" .
Historically, USP <85> requires 50-200% recovery for spiked samples during validation. TR 82 acknowledges that for LER-prone products, early time-point recovery (e.g., 0-2 hours) may meet this, but later time points (24 hours) will not. The report suggests that validation protocols should define a "window of recovery" based on the product’s intended clinical handling (e.g., if the product is used immediately after thawing, LER at 24 hours may be less relevant).
For products in development, TR 82 provides guidance to formulation scientists: pda technical report 82
By defining these parameters scientifically, logistics teams can make data-driven decisions when delays occur, rather than defaulting to scrapping the product out of caution.
Under TR 82, manufacturers are encouraged to understand: It sounds like you’re referring to (TR 82),
Low concentrations of endotoxin can adsorb to the surface of the primary container (glass, plastic) or even to the product’s own aggregated proteins. The endotoxin is physically present but not in the solution phase sampled for the BET.
The core recommendation of TR 82 is to perform a systematic LER study. This involves: The report suggests that validation protocols should define
TR 82 categorizes LER into several mechanistic hypotheses. Understanding these is critical for implementing the report’s recommendations.
However, in the early 2010s, a disturbing phenomenon began to surface during method validation and stability studies. Analysts observed that even when endotoxin was deliberately spiked into a drug product sample, the recovery values would inexplicably drop to near-zero after a short period of storage—sometimes hours, sometimes days. Yet, when the same sample was diluted or chemically treated, the endotoxin activity returned.