Liver disease may produce significant, albeit highly variable, effects on the pharmacokinetic behaviour of antibiotics in serum. Drug disposition may be altered through several pathophysiological mechanisms including reduced hepatobiliary clearance, and modifications in the volume of distribution induced by albumin synthesis deficiency or portal hypertension-related ascites. Antibacterial agents are not affected by potential alteration in hepatic first-pass effects. Only liver cirrhosis-induced effects on serum pharmacokinetics of antibiotics have been extensively studied, unlike those possibly produced by other forms of liver disease. In liver cirrhosis, pharmacokinetic alterations of nearly all beta-lactam or quinolone agents appear not to be marked enough to require dosage adjustment, provided that renal function stays normal. Adaptation in therapeutic schedule, however, is warranted for those drugs that are substantially cleared by the hepatobiliary system, namely mezlocillin, clindamycin, erythromycin, pefloxacin, enoxacin, antituberculous agents or nitroimidazole derivatives. Special caution should also be exercised when using aminoglycosides or vancomycin because of the wide interpatient variability of their pharmacokinetic disposition and their toxic potential. When renal function is impaired and there is an increased volume of distribution due to ascites, as frequently observed in severe liver insufficiency, the elimination half-life of most antibiotics is markedly prolonged, resulting in potential side effects due to drug accumulation. Accordingly, dosage adjustment applies to all drugs. In this regard, it should be remembered that delineating the dosage guidelines for a given antibiotic on the basis of reported pharmacokinetic parameters in patients with liver cirrhosis is awkward and probably of limited value. This pattern is ascribed to large interpatient variability in the active hepatic cell mass, the degree of portal hypertension and the alteration of serum binding capacity. Furthermore, there is no way of predicting accurately the extent of liver insufficiency in an individual patient. Dosage reduction is thus done empirically in most cases. Whenever possible, direct measurements of serum antibiotic concentrations should be the reasonable approach to manage antibiotic therapy in this kind of clinical condition.