To test the hypothesis that eating may adversely affect the hemodynamic response to ultrafiltration-dialysis, we studied the effect of a standard snack (about 400 Kcal) in 13 patients on RDT. Each patient was studied in random order during two standard hemodialysis sessions (snack-HD and control-HD) performed at identical UF rate. Arterial pressure fell significantly (p less than 0.01) during both the Control-HD (from 135 +/- 8/76 +/- 3 to 121 +/- 10/68 +/- 5 mmHg) and the Snack-HD (from 137 +/- 7/77 +/- 3 to 105 +/- 8/59 +/- 4 mmHg). The rate of fall, however, was significantly higher (p less than 0.01) after the snack than during the corresponding period in the control HD. Consequently, there were more hypotensive episodes requiring saline infusion during Snack-HD (23 in 10 patients) than during Control-HD (12 in 6 patients) (p less than 0.025). In spite of the greater number of interventions, the average fall in arterial pressure after a snack (-22 +/- 3/-13 +/- 2 mmHg) was more marked than during the corresponding period in the control-HD (-13 +/- 3/-9 +/- 2 mmHg). The hypotensive effect of snack was more pronounced in the presence of advanced autonomic neuropathy. Food ingestion impairs the arterial pressure response to UF in patients on RDT. Fasting during hemodialysis may in part prevent hypotension.