The in vivo significance of antibiotic-induced endotoxin release in experimental Gram-negative sepsis

Open Access
Publisher Website
Google Scholar
Abstract
Considerable evidence now supports the experimental findings that penicillin-binding protein (PBP)-2 specific antimicrobial agents such as imipenem generate less endotoxin than PBP-3 specific agents such as ceftazidime during the process of bacteriolysis of Gram-negative bacteria. To determine if differences in endotoxin release have pathophysiologic significance in vivo, Sprague-Dawley rats were experimentally challenged with intraperitoneal injections of virulent, serum-resistant clinical strains of the following Gram-negative bacilli: Escherichia coli 018:K1, Klebsiella pneumoniae K2, and Pseudomonas aeruginosa 12.4.4 (immuno type 6). After intravenous administration of imipenem (25 mg/kg), ceftazidime (50 mg/kg) or saline control, imipenem and ceftazidime-treated animals had rapid reductions in the quantitative level of bacteremia from all three pathogens. Peritoneal fluid samples revealed spherical forms with imipenem and long, filamentous forms with ceftazidime. Circulating plasma endotoxin levels were consistently higher ( P < 0.05) with ceftazidime than imipenem for 6 h after administration of E. coli or P. aeruginosa intraperitoneal challenge. Endotoxin levels were unchanged to slightly higher with imipenem than ceftazidime following K. pneumoniae intraperitoneal challenge. TNF levels peaked 2 h post-therapy and were consistently higher with ceftazidime-treated animals ( P < 0.05). D-galactosamine-treated animals had LD50 values that were 0.5-2 log higher ( P < 0.001) with imipenem for E. coli and P. aeruginosa but did not differ from ceftazidime in animals challenged with the K. pneumoniae strain. These results indicate that the PBP-2 specific agent imipenem led to significantly less endotoxin release than did ceftazidime with its great affinity to PBP-3. Differential endotoxin release was found after antimicrobial therapy with the E. coli and P. aeruginosa strains but not with the K. pneumoniae strain tested in this study. The clinical relevance of these findings with treatment of systemic Gram-negative infections in humans will require further clinical investigation.