The complete metabolic fate of propranolol has been difficult to assess because of its complexity. The aim of this study was therefore to develop an analytical approach that would permit separation and quantitation of all major propranolol metabolites. This was accomplished in the urine of dogs, rats, and hamsters using a combination of techniques. Propranolol, 25 mg/kg, was given together with [3H]propranolol po to dogs and ip to rats and hamsters. After hydrolysis with beta-glucuronidase and sulfatase, the 0-24-hr urine was extracted with diethyl ether first at pH 9.6 to yield basic, phenolic, and neutral metabolites. This fraction contained 35 +/- 1% (mean +/- SE; N = 3) in the dog, 59 +/- 2% in the rat, and 53 +/- 5% in the hamster (based on the total radioactivity in urine). The urine was then extracted at pH 2 for acidic metabolites, yielding 32 +/- 3% in the dog, 4 +/- 1% in the rat, and 5 +/- 1% in the hamster. The pH 9.6 and pH 2 extracts were each separated into its components by HPLC. Liquid scintillation spectrometry and GC/MS allowed quantitation and identification of about 15 phase I metabolites. Major species differences existed, with glucuronidation and side chain oxidation dominating in the dog and ring oxidation in the rat and hamster. A large nonextractable fraction constituting previously unrecognized propranolol metabolites was discovered both in dogs (34 +/- 2%) and in rats (37 +/- 1%) and hamsters (40 +/- 8%). This fraction, which appeared to have a long half-life in the dog and rat, could be separated by HPLC into several distinct radioactive peaks. Preliminary attempts to identify these metabolites, which differed among species, indicated them to be highly polar and labile, some of them probably conjugates of known ring-hydroxylated metabolites.