Allogeneic peripheral blood progenitor cells (PBPCs) were transplanted after immunoselection of CD34(+) cells. Two patient groups were studied: group I patients received immunoselected blood CD34(+) cells and unmanipulated marrow cells from the same donor. Group II patients were given immunoselected blood and bone marrow (BM) CD34(+) cells. One to 6 weeks before bone marrow transplantation (BMT), PBPCs from HLA-identical and MLC(-) sibling donors were mobilized with granulocyte colony-stimulating factor (G-CSF) (5 mu g/kg twice daily subcutaneously) for 5 days. Aphereses were performed at days 4 and 5 of G-CSF application. CD34(+) cells were separated from the pooled PBPC concentrates by immunoadsorption onto avidin with the biotinylated anti-CD34 monoclonal antibody 12.8 and then stored in liquid nitrogen. BM was procured on the day of transplantation. Patients were conditioned with either busulfan (16 mg/kg) or total body irradiation (12 Gy) followed by cyclophosphamide (120 mg/kg). Cyclosporin A and short methotrexate were used for graft-versus-host disease (GVHD) prophylaxis. After transplantation, all patients received 5 mu g G-CSF/kg/d from day 1 until greater than 500 neutrophils/mu L were reached and 150 U erythropoietin/kg/d from day 7 until erythrocyte transfusion independence for 7 days. Group I consisted of patients with acute myeloid leukemia (AML) (n = 2), chronic myeloid leukemia (CML) (n = 2), and T-gamma-lymphoproliferative syndrome and BM aplasia (n = 1). The patients received a mean of 3.3 x 10(6) CD34(+) and 3.7 x 10(5) CD3(+) cells/kg body weight of PBPC origin and 4.5 x 10(6) CD34(+) and 172 x 10(5) CD3(+) cells/kg body weight of BM origin. Group II consisted of five patients (two AML, two CML, one non-Hodgkin's lymphoma). They received a mean of 3.3 x 10(6) CD34(+) and 3.2 x 10(5) CD3(+) cells/kg from PBPC and 1.4 x 10(6) CD34(+) and 0.6 x 10(5) CD3(+) cells from BM. A matched historical control group (n = 12) transplanted with a mean of 5.2 x 10(6) CD34(+) and 156 x 10(5) CD3(+) cells/kg from BM alone was assembled for comparison. In group I, the median time to neutrophil recovery to >100, >500, and >1,000/mu L was 12, 15, and 17 days, respectively. Patients from group II reached these neutrophil levels at days 13, 15 and 17 post BMT. Neutrophil recovery in the control patient group occurred at days 17, 18, and 20 respectively. Group I patients were given platelet transfusions within 18 days and red blood cells within 10 days, whereas for group II patients, these time points were 26 and 17 days, respectively. These same transfusions could be ceased within 38 and 24 days, respectively, in control patients. The addition of about 2% more peripheral blood CD3(+) cells (group I patients) did not result in higher grades of acute GVHD (median grade II) as compared with the controls (median grade II). Four of five group II patients showed no signs of acute GVHD. These data suggest that the addition of immunoselected allogeneic CD34(+) progenitor cells to BM cells may accelerate hematopoietic recovery. (C) 1995 by The American Society of Hematology.