Statistical determination of endotoxin content in influenza virus vaccine by the limulus amoebocyte lysate test

To determine laboratory-to-laboratory variability of the limulus amoebocyte lysate (LAL) test for evaluating endotoxin content in influenza virus vaccine, a collaborative study was designed. Participants were six influenza virus vaccine manufacturers and the Bureau of Biologics (BoB). Lysate lot 116, Reference Influenza Virus Vaccine for Endotoxin Assay E-1, and four test vaccines having different ratios of LAL activity relative to reference E-1 were supplied by the BoB. Each laboratory used its normal test procedure. All vaccines were coded. One pair (reference and test) of vaccines per day was tested for 4 days a week over a 4-week period. All data were analyzed at the BoB. The degree of variability experienced by testing laboratories was estimated by the study. This estimate did not conflict with experience gained from previous routine testing in any of the participating laboratories. A statistical approach to the evaluation of LAL data from testing influenza virus vaccine for endotoxin content was developed based upon the overall variation obtained from the collaborative study.