Previous studies have suggested that exercise during hemodialysis (HD) could increase the efficacy of solute removal, although this hypothesis has not been conclusively evaluated. The goal of this study was to compare the removal of low-molecular weight solutes between HD sessions, with and without aerobic exercise. It was a controlled clinical trial, including HD patients in a randomly cross-over design, such that each patient received a HD session with exercise (intervention) and the next one without exercise (control), three times each. In the exercise sessions, patients pedaled on a cycle ergometer for 60 minutes. The total mass of removed urea, potassium, creatinine, and phosphate were calculated from the solutes concentration in dialysate (continuous spent sampling of dialysate). This was evaluated in a total of 132 HD sessions of patients with a mean age of 54 ± 15 years, 75% male and HD vintage of 3 (2-13) years. Phosphate removal in dialysate during intervention sessions was significantly higher (5.6 [2.5-18.9] vs. 5.1 [1.5-11.2] mg/min) than during control sessions, P = 0.04. The median mass of phosphate removed during control HD session was 1226 (367.8-2697.2) vs. 1348.6 (613.0-4536.2) mg/session during intervention sessions. The exercise did not modify the removal of urea (control 122.6 [61.3-286.0] vs. exercise 112.4 [51.1-250.3] mg/min, P = 0.44), creatinine (control 5.6 [2.5-13.8] vs. exercise 5.6 [2.5-12.8] mg/min, P = 0.49), or potassium (control 13.3 [11.2-15.8] vs. exercise 13.8 [6.6-15.8] mEq/min, P = 0.49). Aerobic exercise during HD increases the efficacy of phosphate removal, without changing urea, creatinine and potassium removal. The implications of this finding in mineral and bone disease and cardiovascular disease need to be evaluated on future clinical trials.