Influence of drug class and dose size on antibiotic-induced endotoxin/IL-6 release and impact on efficacy of anti-endotoxin antibody

Abstract

There are limited comparative data on antibiotic-mediated endotoxin release and cytokine liberation in septic hosts. To determine the effect of antimicrobial drug class and dose size on the relative concentration kinetics of free endotoxin and interleukin-6 (IL-6), CF-1 mice were made septic following a thermal injury and low-dose (103 CFU) subeschar challenge with Klebsiella pneumoniae serotype K2. Single intraperitoneal (i.p.) doses (in mg/kg) of ceftazidime (TAZ, 25, 200), aztreonam (AZT, 200), piperacillin (PIP, 200), meropenem (MER, 200), imipenem (IMI, 25, 100, 200), ciprofloxacin (CIP, 25) and gentamicin (GEN, 25) were administered at 72 h post burn and infection, when mice were septic with organ dysfunction. AZT, TAZ, MER, PIP (each at 200 mg/kg) and IMI (100 mg/kg) resulted in fold-increases in median endotoxin levels of 15.3, 14.9, 13.1, 8.2 and 12.4, respectively. All were significantly greater than predose baseline values (P < 0.01), however differences among agents did not reach statistical significance. The increases in free endotoxin levels for all of the beta lactams (8.2-15.3-fold) and CIP (7.7-fold) were significantly greater than for GEN (3.9-fold, P < 0.01). The fold-rise in median IL-6 concentrations from baseline for the beta lactams ranged from 3.0-7.7. All of the beta lactams resulted in statistically greater IL-6 release as compared with CIP (1.9-fold, P < 0.01) and GEN (1.4-fold, P < 0.01). The median endotoxin concentrations were significantly higher for the 100 mg/kg (668 EU/ml) and 200 mg/kg (862 EU/ml) doses of IMI compared to the 25 mg/kg dose (378 EU/mL, P < 0.05). There was also a significant increase in endotoxin levels with a dose increase of TAZ from 25 to 200 mg/kg (597 vs 1030 EU/ml, P= 0.017). The addition of antiendotoxin monoclonal antibody (E5, 2 mg/kg i.p. four times daily [qid]) to AZT (75 mg/kg qid) or TAZ (10 mg/kg qid) for 2 days significantly reduced the mortality by ∼20-40% for a 2-3 day period (P < 0.05) compared to AZT or TAZ alone. In contrast, the administration of E5 with either CIP (10 mg/kg) or GEN (10 mg/kg) by the same qid dosing schedule did not improve survival (P> 0.05). In conclusion, data from the present study, generated in a physiologically relevant model of sepsis, found relatively greater release of unbound endotoxin and IL-6 with the beta lactams, intermediate for CIP and lowest for GEN. The size of the dose of two beta lactams was also found to be a potentially important variable, suggesting that dose optimization might be possible to minimize antibiotic-associated endotoxin release. Furthermore, under our experimental conditions, the efficacy of E5 appears to be dependent upon the antibiotic with which it is administered. In order to understand better the potential significance of antibiotic-induced endotoxin release, the pharmacologic variables of the antimicrobial should be systematically evaluated in preclinical and clinical trials.