In a double-blind placebo-controlled study we investigated the effect of recombinant human erythropoietin (r-HuEPO), on the perioperative hemoglobin concentration and the use of blood transfusions in patients undergoing elective colorectal surgery with a preoperative hemoglobin level =8.5 mmol/L. Altogether 100 were included, and 81 patients could be evaluated. A total of 38 patients received r-HuEPO in a dose of 300 IU/kg body weight on day 4 before surgery and 150 IU/kg daily for the following 7 days; 43 patients received placebo. In addition, all patients received daily doses of 200 mg iron orally for 4 days before surgery. There were no differences between the two groups with regard to sex, height, weight, serum electrolytes, and liver function tests at study entry. The preentry hemoglobin concentration was similar in the two groups, with a median value of 7.9 (range 5.3-8.5) mmol/L in the erythropoietin group and 7.6 (5.1-8.5) mmol/L in the placebo group. On the day of surgery the median hemoglobin concentration was 7.8 (5. 3-9.2) mmol/L in the erythropoietin group and 7.2 (4.6-8.5) mmol/L in the placebo group (p < 0.05). On postoperative days 3 and 7 the values were 7.2 (5.3-8.2) and 7.5 (5.4-9.4) mmol/L, respectively, in the erythropoietin group compared to 6.7 (5.2-7.8) and 6.9 (5.1-8.6) mmol/L in the placebo group (p < 0.01). At discharge the hemoglobin concentration was 7.8 (5.9-8.8) mmol/L in the erythropoietin group and 7.2 (5.4-8.6) mmol/L in the placebo group (p < 0.002). The blood loss during operation was similar in the two groups. In the erythropoietin group the median value was 280 ml (range 25-2000 ml), with the lower and upper quartiles 150 and 500 ml, respectively. In the placebo group the blood loss was median 300 ml (range 50-1800 ml), with the lower and upper quartiles 200 and 750 ml, respectively. The number of blood transfusions given was significantly lower in the erythropoietin group, with a mean of 0.3 (range 0-6) units compared to 1.6 (0-9) units in the control group (p < 0.05). In conclusion, the hemoglobin concentration at the time of surgery and during the week following surgery was significantly higher in the group of patients receiving r-HuEPO perioperatively compared to the placebo group together with a significant lower use of blood transfusions in the r-HuEPO group. However, the clinical implications of these findings has yet to be proven.