We conducted a randomized unblinded parallel clinical trial to compare the effectiveness, side effects and tolerance between calcium acetate (CA) and calcium carbonate (CC) in 80 stable chronic hemodialysis patients selected on the basis of their acceptable control of serum phosphorus (P) levels with aluminum hydroxide (AH). All patients were dialyzed against the same calcium dialyzate (1.62 mmol/l). The serum analytical tests included: calcium corrected to total protein, P, PTH (intact molecule) and bicarbonate. The study was divided into the following periods: P0: baseline measurements; P1: washout (winthdrawal of AH for 15 days); P2: random allocation to CA and CC treatment at doses equivalent to 75 mEq of elemental calcium, stratified according to previous doses of AH (2 months); P3: adjustment of doses until control P (2 months). CA was poorly tolerated in 7 patients and CC in 2 (NS). The changes in serum P levels between P0 and P2 periods were lower in the CA group (1.73 ± 0.25 vs. 1.80 ± 0.50 mmol/l; p = 0.26) than in the CC group (1.77 ± 0.35 vs. 1.93 ± 0.48 mmol/l; p = 0.03, paired t test). Serum calcium was hardly modified by CA (2.42 ± 0.20 vs. 2.47 ± 0.17 mmol/l; NS) while in the CC group, it rose significantly (2.40 ± 0.12 vs. 2.55 ± 0.22 mmol/l; p = 0.0004). There were no differences in the control of PTH or bicarbonate. At the P3 period, patients on CC treatment needed higher doses of elemental calcium to control P than the CA group (99.4 ± 31 vs. 83 ± 27 mEq; p = 0.01), and at these doses, hypercalcemia ( > 2.86 mmol/l) was more frequent in the CC group (30.5 vs. 16%; NS). We conclude that CA is slightly more effective and safer than CC, albeit less well tolerated.