Glutathione efflux into bile of the fluorocarbon-perfused isolated rat liver was altered with eight different agents (L-buthionine-[S,R]-sulfoximine, cefamandole, sodium arsenite, phenobarbital, furosemide, nitrofurantoin, aminopyrine, and benzylamine), and correlations were established between bile flow and biliary excretion of 1) glutathione, 2) endogenous bile acids, and 3) glutathione plus bile acids. Biliary efflux of endogenous bile acids was relatively low (0.5-5 nmol.min-1.g liver-1) and was minimally affected by these agents. Biliary glutathione excretion in control livers was between 4 and 9 nmol.min-1.g-1 and in treated livers ranged from 1 to 21 nmol.min-1.g-1. For each of the various interventions, an increase or decrease in glutathione excretion was always accompanied by a change in bile flow in the same direction; however, these changes were not always directly proportional when comparisons were made between treatment groups. Nevertheless, when bile flow (microliter.min-1.g-1; ordinate) was plotted against glutathione excretion into bile for the pooled data, a significant correlation was observed that was adequately described by a straight line: y = 0.071 chi + 0.72 (r2 = 0.62, P less than 0.001). A similar function described the relation between bile flow and the sum of bile acids and glutathione in bile: y = 0.077 chi + 0.55 (r2 = 0.62, P less than 0.001). In contrast, the taurocholate- or glycocholate-induced choleresis had only minimal effects on glutathione efflux. These findings support the hypothesis that glutathione is one of the osmotic driving forces in bile acid-independent bile formation.(ABSTRACT TRUNCATED AT 250 WORDS)