In a prospective study, the clinical relevance of a quantitative chromogenic endotoxin assay in plasma (detection limit 10 ng/L, assay time 2.5 hours) versus blood cultures was evaluated in 51 critically ill patients with increased susceptibility for infectious complications. Of the 400 samples tested, the endotoxin assay and bacterial culture both were negative in 342 samples. In 21 samples from 15 patients, gram-negative aerobic microorganisms were cultured. Corresponding endotoxin assays were positive in 14 samples (mean 100 ng/L). Twenty-three samples grew gram-positive bacteria. The associated endotoxin assays all were negative. Twelve samples were found to be endotoxin-positive without a corresponding gram-negative bacterial culture. In 7 of these 12 positive endotoxin assays, a laboratory or clinical explanation for these positive tests could be provided. In view of the high sensitivity, specificity, and predictive values obtained, the authors conclude that the endotoxin assay used is a useful clinical adjunct for both the detection and exclusion of gram-negative septicemia.