This prospective study was carried out in healthy donors and patients. The performance of the CliniMACS was evaluated with the comparison of the numbers of total nucleated cell (TNC) within and over the capacity of the normal scale column. In addition, large vs. normal scale column and manual vs. automated washing systems were also compared. A total of 44 selections were done. Eighteen normal scale selections were done with initial TNC numbers over 6 x 10(10) and 14 selections were performed below this number. None of the cases had CD34+ cell numbers over the capacity. Flow cytometry was used to check each separation performance for purity and recovery of CD34+ cells along with T- and B-cell depletion level parameters. All healthy donors had significantly better mean purity and recovery of CD34+ cells, and T- and B-cell depletion status than that of patients with values 95 vs. 85%, P: 0.006; 77 vs. 58%, P: 0.004; 4.55 log vs. 4.06 log, P: 0.004; 3.19 log vs. 2.63 log, P: 0.01, respectively. However, the performance of the system was not dependent on using the normal or large-scale column; automated or manual washing systems; and initial TNC numbers above (>6 x 10(10), range: 7.05-21.84 x 10(10), mean: 12.32 x 10(10)) or within (<6 x 10(10), range: 0.86-5.89 x 10(10), mean: 4.15 x 10(10)) the column capacity. In conclusion, the performance of the CliniMACS is more efficient in healthy donors than in patients. However, the performance of the system did not change as long as the numbers of CD34+ cells (range: 0.34-5.87 x 10(8)) were not exceeding the column capacity despite that more than 6 x 10(10) TNCs were used.