[Validation of three limulus tests for the determination of endotoxin in dialysate/substitute used in on-line hemodiafiltration].

The on-line hemodiafiltration, a newly developed technic of artificial kidney, requires a large mount of dialysate as a substitute for filtered plasma. The dialysate is prepared in the hospital, and is liable to be contaminated. To avoid adverse effects from contaminants, the dialysate needs to be checked at least for endotoxin (ET) with a highly sensitive, accurate and precise method. Three commercial kits, Endospecy (Seikagaku, Tokyo), ES-single (Wako, Osaka) and QCL (BioWhittacker, Maryland), were tested to see if they receive any interference from dialysate. The recoveries of Escherichia coli and Salmonella ET spiked in dialysate at final concentrations of 50, 100 and 150 EU/l were evaluated against those spiked in distilled water using simple linear regression analysis. The validation criteria was that the coefficient of a regression line which was forced to pass through the origin should fall between 0.75 and 1.25. No interference was observed with Endospecy. The QCL showed no interference with E. coli ET but enhancement with Salmonella ET. The ES-single was least sensitive for either ET. Dilution of dialysate affected the recoveries by QCL and ES-single but not those by Endospecy. The Endospecy was thus the only reagent that could detect ET level as low as 1.0 EU/l without interference from dialysate. The rise in body temperature after 4-hour on-line hemodiafiltration was correlated with ET level in the substitution fluid (r = 0.48, p < 0.05), and was significant when the level was 1.0 EU/l or more. These results suggest that the ET level of substitution fluid should be monitored with Endospecy and that the level should be kept below 1.0 EU/l.