Abstract

Summary: The antiendotoxin effect of polymyxin B was investigated in experimentally induced septicemia in rabbits. The Pasteurella multocida organisms were sensitive to the antibacterial action of penicillin but not to polymyxin B. Animals pre-treated with polymyxin showed positive blood cultures and significantly reduced plasma endotoxin levels (Limulus test) with normal white blood cell and platelet counts when analyzed 6 hr after the injection of live organisms. Polymyxin therapy given after the animals had established septicemia-endotoxemia reduced the plasma endotoxin levels and improved the survival, but had no effect on the leukopenia and thrombocytopenia. The best survival data were obtained in rabbits who were treated with both penicillin and polymyxin. The data suggest that polymyxin is effective in neutralizing the endotoxic effects from live organisms and that the timing and perhaps duration of the polymyxin treatment is of critical importance. In addition, a modified Limulus lysate method was developed which showed that quantitative plasma endotoxin determination could be made more sensitive by prior heating of the plasma to remove the natural inhibitors. Speculation: This investigation and other reports have shown that polymyxin B sulfate neutralized endotoxin effects in animals given purified endotoxin, dead endotoxin-containing organisms, and gram-negative septicemic animals (3, 6, 8, 12–15). The antiendotoxin effect of the antibiotic appears to be separate from its antimicrobial effects. Because many of the manifestations of gram-negative septicemia in humans are reputed to be the result of endotoxemia, a clinical study focused on neutralization of the endotoxin effects would appear appropriate. The modified Limulus lysate test, as described in the report, was found to be accurate in measuring quantitative plasma endotoxin levels in small quantities of plasma by removal of the natural inhibitors by simply heating the plasma. The data also indicated that the lysate test sensitivity was directly related to the number of infectious organisms per milliliter of blood in the septicemic rabbit. Therefore, its sensitivity in humans could similarly be affected by the presence of inhibitors and the number of organisms. Because most human studies have failed to correlate the positivity of the test with the number of infecting organisms, these data would suggest that re-evaluation of its usefulness in clinical disease should be performed.