1 Eight patients with end-stage renal insufficiency on maintenance haemodialysis, and seven patients with newly diagnosed hyperthyroidism, received a single intravenous dose of diazepam, followed by blood sampling over the next 7 days. Fifteen healthy volunteer controls, matched with patients for age and sex, were similarly studied. 2 Diazepam half-life in renal failure patients (mean 37 h) was greatly reduced compared to controls (mean 92 h, P less than 0.05) and clearance of total (free plus bound) diazepam correspondingly increased (0.94 v 0.34 ml min-1 kg-1, P less than 0.01). 3 However, differences were largely related to disease-related changes in drug binding and distribution. Mean unbound fraction of diazepam in plasma of renal patients (7.0%) was greatly increased over controls (1.4%, P less than 0.01) and Vd of unbound diazepam greatly reduced (57 v 157 l/kg, P less than 0.01). 4 Clearance of pharmacologically active unbound diazepam (intrinsic clearance) was not significantly different between renal patients and controls (23 vs 30 ml min-1 kg-1). 5 None of the kinetic variables for total or unbound diazepam in thyrotoxic patients differed significantly from those in controls matched for age and sex. 6 End-stage renal failure (or its associated drug therapy) alters diazepam protein binding and distribution, but does not significantly change clearance of unbound drug. Thyrotoxicosis does not influence diazepam kinetics.