The purpose of this study was to investigate the effect of pentobarbital, halothane, and chloralose anesthesia on the endogenous release of cholecystokinin-33 (CCK-33) in dogs prepared with duodenal fistulas. Release of CCK-33 was induced by intraduodenal infusion of a medium-chain triglyceride (corn oil, 1 gm/kg/hr). Plasma CCK-33 concentrations were measured by means of a specific radioimmunoassay. Pentobarbital and chloralose were administered intravenously, and halothane was administered by a vaporizer (semiclosed technique), with O2 and N2O used as carriers. No incidence of hypotension was found with the use of these anesthetic agents. Basal concentrations of plasma CCK-33 were elevated, although not significantly, during pentobarbital or chloralose anesthesia. In conscious dogs (control study), peak plasma CCK-33 concentrations of 529 +/- 53 pg/ml were measured 30 minutes after intraduodenal infusion of fat. Under pentobarbital anesthesia, peak plasma CCK-33 concentrations of 452 +/- 264 pg/ml were found 80 minutes after infusion of fat. Under halothane anesthesia, fat-induced release of CCK-33 was abolished, whereas chloralose anesthesia did not influence fat-induced release of CCK-33. These findings may have implications for the design of future studies of gastrointestinal physiology. In CCK-33 studies that require anesthesia, chloralose appears to be an appropriate anesthetic agent.