Cephalosporin and carbacephem nephrotoxicity: Roles of tubular cell uptake and acylating potential

Abstract

Three beta-lactams, desacetylcephaloglycin, ampicillin, and loracarbef, were studied to test a hypothesis derived from retrospective analysis of previously studied cephalosporins: that beta-lactam nephrotoxicity develops in approximate proportion to tubular cell antibiotic concentrations and lactam ring reactivities. Concentrations of each beta-lactam (and inulin) in rabbit renal cortex and serum were measured at the end of 0.5-hr infusions of 100 mg antibiotic/kg body weight and 0.5 to 0.67 hr later. Total cortical AUCs (total areas under the curve of concentration and time in renal cortex) and transported cortical AUCs (total minus inulin-space beta lactam) were calculated from these measurements. Reactivities, determined by the rate constants of lactam-ring opening at pH 10, were taken from the literature. Nephrotoxicity was quantified by grades of proximal tubular cell necrosis and by serum creatinine concentrations 2 days after infusion of 100–1500 mg/kg of the antibiotics. Desacetylcephaloglycin was slightly less nephrotoxic than cephaloglycin; the AUCs, reactivities, and toxicities of these two cephalosporins fit the proposed model, particularly when allowance is made for hepatic and renal deacetylation of cephaloglycin. The very low AUCs, limited reactivity, and absence of nephrotoxicity of ampicillin also fit the model. Loracarbef had a transported AUC less than three times, and reactivity one-thirtieth, those of cefaclor, respectively. Although only at 1500 mg/kg, loracarbef was significantly more nephrotoxic than cefaclor. If the reactivity of loracarbef with its targeted bacterial proteins, which is essentially the same as that of cefaclor, is considered instead of the base hydrolysis rate constant, then loracarbef also fits the model. By the same analysis, the comparatively high in vitro stability of other carbacephems, although pharmaceutically convenient, may not limit their nephrotoxicity.