Prediction of drug dosage in patients with renal failure using data derived from normal subjects.

A one-compartment model designed to predict alterations in persistence of drugs in uremic patients was constructed using information obtained from normal subjects. Data obtained from the literature in which the fraction of absorbed drug eliminated unchanged in the urine and the apparent elimination rate constants were compared in both normal control subjects and in severely uremic patients. Twenty-two drugs were examined. Despite changes in apparent volume of distribution and metabolism reported in uremia, the model was able to predict overall elimination rate constants in severe uremia with an error under 10% for 12 and under 20% for 7 additional drugs. The method appears to be most informative for drugs that tend to be retained in the presence of renal failure. Great error was observed with doxycycline and erythromycin. Studies with erythromycin lactobionate and a cupplate assay reduced the error for this drug from 56% for the glucoheptonate to 18%. Doxycycline is known to have a complex enterohepatic circulation. The model is offered as a useful approach to predict dosage adjustment in uremic patients with drugs for which data are not available.