The potential barrier effect of erythrocytes (RBC) on renal excretion (mainly by tubular secretion) of hydrochlorothiazide (HCTZ) was evaluated in nine anesthetized rats during steady-state iv infusion. Drug concentrations in plasma and blood from the carotid artery and renal vein were assayed by a simple modified HPLC method. Renal extraction ratios were concentration-independent with a mean of 0.17 +/- 0.05 (SD). The renal excretion was found to occur primarily from the drug in plasma; the mean net fractional removal from plasma was 0.57 +/- 0.12, while that from RBC was less than 0.008 +/- 0.041. The virtual total unavailability of HCTZ from RBC (containing approximately 70% of drug in arterial blood) for renal excretion is attributed to relatively slow efflux of drug from RBC to plasma during each passage through the kidney compared with the blood transit time (in seconds). Preliminary in vitro influx and efflux kinetics of HCTZ across RBC membranes were studied using rat and human blood. The flux data could be adequately described by a linear, reversible, closed two-component system model, and the mean equilibration half-times (ET1/2) in rat and human blood were 10.9 and 20.5 min, respectively. The mean residence time of drug in blood circulation of rats was estimated to be 8.32 +/- 1.06 min, which is shorter than the ET1/2. This is consistent with data indicating that distribution equilibrium of HCTZ in arterial blood might not be reached in vivo even at steady state. Other implications of slow transport kinetics of drugs across RBC membranes are discussed.