BACKGROUND Various open and semi-closed methods are used for red cell (RBC) depletion and hematopoietic progenitor cell (HPC) enrichment of bone marrow (BM) in vitro, but with variable efficacy. A simple, efficient, and safe method using dextran 110k was developed. METHODS An equal volume of 4.5-percent dextran was applied to major ABO-incompatible BM in transfer bags and sedimentation was allowed for 30 minutes. RBCs, nucleated cells (NCs), and mononuclear cells (MNCs) from BM allografts before and after dextran sedimentation (DS) were counted. Flow cytometry, short-term cultures, and long-term cultures were performed to assay the respective recovery of CD34+ cells, colony-forming units (CFUs), and long-term culture-initiating cells (LTC-ICs). RESULTS Sixteen BM collections were processed. The mean volume was 666 mL (range, 189-1355 mL). The mean +/-1 SD post-DS NC, MNC, CD34+ cell, and CFU counts per kg of the recipient's body weight were 4.11 +/-1.74 x 10(8), 8.98 +/- 3.68 x 10(7), 2.90 +/- 1.95 x 10(6), and 2.03 +/- 2.01 x 10(5), respectively, with the corresponding post-DS recovery being 90.6 percent, 90 percent, 92.4 percent, and 100.8 percent. The numbers of LTC-ICs in cultures (up to 12 weeks) of pre-DS and post-DS samples of five BM allografts were comparable (p = 0.91). Residual RBCs were 5.1 +/- 4.6 (0.1-14) mL with depletion of 96.5 +/- 3.2 percent. There was no significant difference in the mean absolute RBC count in post-DS BM allografts and in four ficoll-treated BM allografts (8.09 x 10(10) vs. 4.9 x 10(9); p = 0.206) and in eight major ABO-incompatible peripheral blood HPC collections (8.09 x 10(10) vs. 9.81 x 10(10); p = 0.87). No posttransplant hemolysis was encountered. Engraftment occurred at 22 +/- 7 days, which is similar to that of four transplants with ficoll-treated BM allografts (22 +/- 9; p = 0.611) and 54 unprocessed BM allografts (19 +/- 6; p = 0.129). CONCLUSIONS DS is an efficient method of depleting RBCs in major ABO-incompatible BM allografts without significant loss of HPCs.