Potential causes for reported incomplete (usually 40-60%) and often highly variable (e.g., 11-79%) bioavailability of furosemide in humans were investigated. The drug was found to be fairly stable in gastric fluids and its hepatic first-pass elimination (HFPE) was estimated to be much less than 6% based on published i.v. data. The rat was used as the main model for extensive evaluation. About 4% (n = 4) of dose was recovered unchanged in the GI tract after i.v. injection while about 40% (n = 12) was recovered after a 120-fold (0.05-6 mg) dose range of oral administration. In another study 70% of the oral dose eventually disappearing (presumably due to absorption and first-pass elimination) from the GI tract was estimated to occur in just 20 min. These data indicate an unsaturable, incomplete, site-specific absorption as well as a lack of dissolution-rate-limited absorption at the doses studied. Based on plasma data, oral bioavailability in four rats was only 30%, and the HFPE much less than 10%. After oral administration, 61% of the dose was absorbed and/or metabolized in the GI recovery study. Thus, 20-30% of oral dose in rats must be metabolized in the GI wall during absorption. The metabolic activity of stomach (homogenate) from 5 rats was found to be much (e.g., 5-10.5-fold) greater than those of liver and small intestine. This was also confirmed in preliminary studies with 3 rabbits and 1 dog. Large intersubject variability in enzyme activity was found in rats and rabbits. The phenomenon of a presystemic first-pass effect was also substantiated by urinary excretion data of a metabolite. It is postulated that variable gastric and intestinal first-pass metabolism may be a major factor causing incomplete and irregular absorption of furosemide in humans.