Uremic patients with suppressed food intake may regain appetite soon after starting dialysis, presumably because of the removal of one or more toxic factors that suppress appetite. To investigate this matter, this study used a new experimental model in free-moving, unstressed male Wistar rats (300 to 350 g) with feeding catheters channeled from the top of the skull to the oral cavity. When the rats recovered from surgery, they were tested under standardized conditions by being given an intraoral infusion (1 mL/min) of a 1 M sucrose solution or a 97 g/L protein solution or a mixed solution of carbohydrate, protein, and fat (Fortimel (Nutricia Nordica AB, Stockholm, Sweden)) while the time (volume) of ingestion was recorded. Solutions to be tested for their ability to inhibit ingestion were injected intraperitoneally (lp) and the intraoral infusion was started 20 min later. Plasma ultrafiltrate was collected from end-stage renal failure patients by isolated ultrafiltration at the beginning of their first hemodialysis and pooled. Ultrafiltrate was also obtained by filtering pooled plasma from healthy volunteers in vitro, using the same type of dialyzer and cellulose acetate membranes as those used in the uremic patients. Morning urine samples from healthy volunteers were pooled and subjected to the same in vitro filtration procedure as the normal plasma. Intraperitoneal injection of 20 mL normal ultrafiltrate had no effect on sucrose ingestion, whereas injection of 20 mL uremic ultrafiltrate reduced the ingestion of sucrose solution by 23% and the ingestion of Fortimel by 17%. Ten mL of ultrafiltrate from normal urine reduced the sucrose intake by 42%. The pooled ultrafiltrates from normal and uremic plasma and normal urine were subjected to molecular filtrations using a series of membranes with known cut-off points. The filtrations yielded four concentrated fractions with molecular weight ranges of 0.1 to 0.5 kilodaltons (kd), 0.5 to 1 kd, 1 to 5 kd, and 5 to 10 kd, respectively; the plasma fractions were concentrated a factor of about 25:1 and the urine fractions by about 15:1. After an ip injection of 2 mL of each concentrated plasma fraction, only the 1 to 5 kd fraction from the uremic ultrafiltrate inhibited sucrose intake, whereas the corresponding fraction from the normal ultrafiltrate had no effect. After injection of 1, 3, and 5 mL of the concentrated fractions of uremic ultrafiltrate, a dose-dependent inhibition of sucrose intake was achieved with the 1 to 5 kd fraction and, to a lesser extent, with the 5 to 10 kd fraction. Intraperitoneal injection of 0.5, 1.0, and 2 mL of the concentrated 1 to 5 kd fraction, but not of the other fractions from normal urine, also resulted in a dose-dependent inhibition of sucrose intake. The 1 to 5 kd fractions from the uremic ultrafiltrate and the normal urine ultrafiltrate also inhibited protein intake in a dose-dependent manner. These results suggest that one or more toxic compounds in the middle-molecule weight range, which are normally excreted in the urine, accumulate in uremia and suppress food intake.