In 6 hypertensive patients with terminal renal failure maintained on hemodialysis, the effects of 'salt subtraction' and of sequential ultrafiltrating were evaluated. Following each of 3 weekly hemodialysis sessions, salt subtraction was carried out by ultrafiltrating 1 liter and simultaneously infusing an equal volume of 5% dextrose. This resulted in a net sodium loss without hypovolemia. After a 2-week period of this procedure, the blood pressure prior to dialysis was lower (156/76 +/- 12/5 mm Hg) than after a comparable number of sequential ultrafiltration sessions (181/88 +/- 10/6 mm Hg; mean +/- SEM). This difference was not statistically significant. At the same time, body weight was comparable at 64.4 +/- 3 and 64.7 +/- 4 kg, respectively. Neither plasma renin activity nor plasma catecholamines responded with a clear increase to either procedure. The limited effect on blood pressure and the renin system of a marked sodium removal during salt subtraction suggests that sodium must still be present in excess in these patients. The procedure of salt subtraction appears safe and subjectively well tolerated, but it can probably not be used as the sole means of decreasing total body sodium without associating dietary measures to reduce sodium intake.