Abstract
It has been 40 years since the first description of bacteremia due to gram-negative rods appeared.1 It has always been assumed that the endotoxins (lipopolysaccharides) of gram-negative bacteria were responsible for the clinical manifestations of infections caused by these bacteria. This view was based on the observation that many of the characteristics of the infections could be reproduced by the infusion of endotoxin in laboratory animals and humans.2 Since the 1960s, it has been well known that the seriousness of the underlying condition was one of the most important predictors of outcome for patients with gramnegative bacteremia. Furthermore, it is . . .