Environmental endotoxin measurement: the Kinetic Limulus Assay with Resistant-parallel-line Estimation.

A Limulus assay method was specifically designed for environmental endotoxin aerosols. Application of new statistical and sample preparation methods strengthened the validity and precision of the Limulus test. Statistically, the Kinetic Limulus Assay with Resistant-parallel-line Estimation (KLARE) differed from conventional analytic methods (as used in chromogenic assays and other kinetic methods) by routinely using a dilution series of the unknown sample as well as the standard to compute potency and an estimate of variance for each sample. Analysis of dose-response slopes for the standard and unknowns detected inhibition and enhancement effects--without multiple assay. Concentration-dependent interference and a more complex, concentration-independent interference with the Limulus assay were detected. Resistant regression and a standardized data analysis corrected for concentration-dependent interference. Sample preparation in a buffer eliminated concentration-independent interference and, thus, improved both the validity and the precision of potency measurements. The utility of a sample buffer and of parallel-line analysis, with both turbidimetric and chromogenic lysates, was demonstrated by assay of three control standard LPS and reference LPS (EC5). The limit of detection for endotoxin was less than 1 pg/ml in buffer. Samples containing greater than or equal to 10 pg/ml were measured with a coefficient of variation of approximately 6% in a single assay. Reproducibility of potency estimates for four samples over 3 days was compared on the basis of standard errors of the mean. The conventional method gave on average a CV of 65% while the resistant-parallel-line method gave, on average, a CV of 6%. Also, the conventional method failed to detect interference and, thus, included data from invalid assays. Conventional analysis of environmental aerosol samples was highly sensitive to the choice of dilution factor causing as much as 1000% variation in the result. By contrast, KLARE results changed by at most 30% with similar changes in initial dilution because KLARE was able to detect, and correct for, the influence of interferant compounds.