Comparison of haematological recovery times and supportive care requirements of autologous recovery phase peripheral blood stem cell transplants, autologous bone marrow transplants and allogeneic bone marrow transplants.

The haematological recovery time, infection rate and supportive care requirements of patients receiving recovery phase autologous peripheral blood stem cell transplants (APBSCT) (n = 38), autologous bone marrow transplants (autoBMT) (n = 13) and allogeneic bone marrow transplants (alloBMT) (n = 14) were compared with respect to the time post-transplant to reach 0.1, 0.5 and 2.0 x 10(9) neutrophils/l and 50 and 150 x 10(9) platelets/l, the length of hospitalization, fever and antibiotic use, the incidence of documented infection and the number of red cell and platelet transfusions. The APBSCT group had a significantly more rapid recovery of neutrophils and platelets and their supportive care requirements were significantly less than the autoBMT and the alloBMT groups. There was no difference between the latter two groups. The most significant variables contributing to the differences in haematological recovery times were the granulocyte-macrophage progenitor (CFU-GM) dose infused and, to a lesser extent, patient age. The APBSCT group received a higher CFU-GM dose of 87 +/- 12 x 10(4)/kg BW compared with 12 +/- 5 and 17 +/- 3 x 10(4)/kg BW in the autoBMT and the alloBMT groups, respectively (p = 0.0001). Patient age showed a negative correlation with the rate of recovery because the APBSCT group, which recovered faster was also older (48 +/- 2 years, compared with 33 +/- 3 and 31 +/- 2, respectively, p = 0.0001). On multivariate analysis, CFU-GM dose was the only variable to show a significant correlation with all the haematological recovery endpoints studied in these 65 patients.(ABSTRACT TRUNCATED AT 250 WORDS)     

Lack of correlation between nucleated bone marrow cell dose, marrow CFU-GM dose or marrow CFU-E dose and the rate of HLA-identical sibling marrow engraftment

There was no correlation between the rate of marrow engraftment and the number of nucleated bone marrow cells infused into 50 HLA-identical sibling marrow graft recipients with haematological malignancy conditioned with cyclophosphamide and fractionated total body irradiation, and immunosuppressed with either cyclosporin (42 patients) or methotrexate (eight patients). Similarly, there was no correlation between the number of marrow CFU-GM or CFU-e infused into recipients immunosuppressed with cyclosporin. The data show that recipients of HLA-identical sibling marrow allografts conditioned with cyclophosphamide and total body irradiation require less than 3 X 10(8) nucleated bone marrow cells/kg recipient weight to ensure engraftment, and throw doubt on the relevance of measuring committed progenitor cells in the donor marrow to assess the likelihood of subsequent haemopoietic reconstitution after matched sibling transplantation.     

Analysis of factors affecting hematopoietic recovery after autologous bone marrow transplantation for lymphoma

Thirty patients with relapsed lymphoma (14 Hodgkin's disease, 16 non-Hodgkin's lymphoma) who underwent autologous bone marrow transplantation (ABMT) were assessed for the effect of different parameters on the rate of hematologic recovery post-ABMT. There was no correlation between nucleated cells count or numbers of CFU-GM or BFU-E in the infused marrow on either neutrophil or platelet recovery times. Patients with lymphoma who received more salvage chemotherapy (greater than six cycles) prior to marrow harvest took significantly longer to recover neutrophils to greater than 0.5 x 10(9)/l (p = 0.0229) and platelets to greater than 20 x 10(9)/l (p = 0.0007) than patients who received less than five cycles of salvage chemotherapy prior to harvest. There was no significant correlation between underlying disease, use of total body irradiation or status of cytomegalovirus serology and recovery times post-ABMT. The results suggest that extensive salvage chemotherapy may result in considerable stem cell damage to marrow, and that potential ABMT candidates with lymphoma should undergo harvest of tumor-free bone marrow as soon as possible following first relapse.     

Progenitor cell assays predict hematopoietic reconstitution after syngeneic transplantation in mice

Hematopoietic reconstitution following syngeneic bone marrow transplantation with graded doses of untreated and drug-treated bone marrow was studied in B6D2F1 mice. Granulocyte-macrophage colony- forming units (CFU-GM) and spleen colony-forming units (CFU-S) showed similar in vitro drug sensitivities. Both the speed of hematologic recovery and survival of mice transplanted with untreated or drug- treated bone marrow were directly related to the number of CFU-GM or CFU-S transplanted. Similar hematologic recovery was seen for untreated marrow transplants and treated transplants that had similar CFU-GM or CFU-S content. There is a minimum number of transplanted CFU-GM or CFU- S that allows survival of lethally irradiated mice. This number is present in a marrow transplant containing the equivalent of 5 X 10(3) untreated cells or producing one to two spleen colonies. There also exists a maximum value for the number of hematopoietic progenitors in a marrow graft, above which the rate of hematologic recovery following transplantation is rapid and no detectable increase in the rate is seen with increasing CFU-GM or CFU-S content. The presence of this maximum value for transplanted progenitors and variations in culture techniques are probably the reasons previous studies have not always shown a correlation between CFU-GM content and hematologic recovery after bone marrow transplantation.     

Granulocytic and stromal progenitors in the bone marrow of patients with primary myelofibrosis

Quantitative and qualitative changes in granulocyte-macrophage (CFU-GM) and fibroblast colony-forming cells (CFU-F) were studied in 7 patients with primary myelofibrosis (MF). Marrow cells were collected from bone biopsy specimens after treatment with collagenase. The number of CFU-GM correlated with the amount of haemopoietic tissue noted in the bone marrow histology and ranged between 0-400/mg of bone. CFU-F were increased in 2 patients with moderate fibrosis. Circulating CFU-GM were increased in all patients studied (169-3749/ml of blood). There was no significant correlation between the number of CFU-GM in the bone marrow and that in the blood. Cytochemical studies showed a high incidence in eosinophil progenitors in the bone marrow and especially in the blood of patients with MF. These data suggest a functional abnormality of myeloid progenitors in this disease.     

Detection of HIV in haemopoietic progenitors

Peripheral blood cytopenias present a major problem in the management of patients with HIV infection. Their pathophysiology is likely to be multifactorial, although there is controversy as to whether haemopoietic progenitors are a target for HIV. In order to investigate the haemopoietic defect in HIV infection, we looked at bone marrow culture characteristics of marrow from eight HIV+ patients compared to normal controls. We performed long-term liquid culture (LTC) and colony forming assays for granulocyte-macrophage (CFU-GM) and granulocyte, erythroid, megakaryocyte, macrophage (CFU-GEMM). In LTC we found normal stromal appearance and haemopoietic focus formation. There was no difference in colony assays of CFU-GM and CFU-GEMM between HIV+ and normal controls. Colonies taken from CFU-GM and CFU-GEMM were analysed for HIV DNA sequences, and we were able to detect HIV DNA in colonies from all HIV+ patients. Our results indicate that despite infection of haemopoietic progenitor cells by HIV, bone marrow function is preserved. This suggests that HIV-related cytopenias may be due to alternative mechanisms not present in our in vitro system.     

Normal myeloid progenitors (CFU-GM) in multiple myeloma: a preliminary study in view of autologous BMT

Normal granulocyte-macrophage precursors (CFU-GM) were studied in 65 multiple myeloma patients by means of culture assays. The patients were divided into separate groups on the basis of previous therapy (i.e. analysis performed at diagnosis or after chemotherapy), time elapsed from the last therapy (i.e. more or less than 1 month) and clinical features of the disease (i.e. tumor stage, immunoglobulin type, bone marrow plasma cell infiltration). The results were evaluated by Wilcoxon rank sum test and linear regression analysis. There was no statistical difference in CFU-GM cloning efficiency or in the number of CFU-GM/ml of bone marrow, even though a larger CFU-GM recovery was found in patients evaluated at diagnosis or at least 1 month or more from previous chemotherapy. In addition, no correlation was demonstrated between bone marrow plasma cell percentage and CFU-GM cloning efficiency. This finding was confirmed by the number of myeloid bone marrow cells in S-phase, assessed by the bromodeoxyuridine labeling index, which showed similar results in patients with different degrees of plasma cell infiltration. In conclusion our data indicate that the granular-monocytic lineage keeps its cell-line potentiality regardless of the degree of marrow plasma cell infiltration and the type of therapeutic approach. These data suggest that autologous bone marrow transplantation might be feasible even in patients with a large neoplastic infiltration.     

Use of recombinant human granulocyte-macrophage colony-stimulating factor in patients with lymphoid malignancies transplanted with unpurged or adjusted-dose mafosfamide-purged autologous marrow.

The neutropenia-related morbidity and mortality occurring after autologous bone marrow transplantation (ABMT) is increased by marrow purging procedures. While phase I through III clinical trials showed the enhancing activity of recombinant human granulocyte-macrophage colony-stimulating factor (rhGM-CSF) on neutrophil recovery after ABMT with unpurged marrow, controversial results have been reported when purged marrow was used. Therefore, it was the aim of the present study to evaluate the efficacy of rhGM-CSF administration in a group of patients (n = 15) with lymphoid malignancies transplanted in complete remission with mafosfamide-purged (n = 10) or unpurged (n = 5) marrow. Mafosfamide concentrations used for marrow purging were evaluated on an individual basis by means of a recently described technique that destroys the granulocyte-macrophage (granulocyte-macrophage colony-forming units [CFU-GM]) compartment, but spares 50% of the more primitive stroma adherent colony-forming cells (CFU-Blast). rhGM-CSF (10 micrograms/kg/d) was started within 24 hours of ABMT and administered in a 4-hour infusion daily until the absolute neutrophil count (ANC) reached 500 x 10(6)/L and then for 7 more days. Patients receiving mafosfamide-purged or unpurged marrow failed to show any difference in terms of median number of days required to achieve an ANC > or = 500 x 10(6) (13 v 14.0, P > .4) cells/L. As compared with retrospective controls, granulocytic recovery was reduced by a median time of 11 (P < or = .0005) and 5 (P < or = .0005) days for patients grafted with purged and unpurged marrow, respectively. The number of CFU-GM (mean +/- SD) infused per kilogram of body weight was significantly lower in patients who received purged autografts as compared with those receiving unpurged autografts (0.85 +/- 0.79 x 10(4) v 15.7 +/- 9.2 x 10(4), P < or = .0005). The dose of CFU-GM progenitors infused per kilogram of body weight did not correlate (r = .031, P > .05) with the time required to reach an ANC > or = 500 x 10(6) cells/L. The number of CFU-Blast (mean +/- SD) infused per kilogram of body weight was not significantly different between patients who received purged or unpurged autografts (5.05 +/- 2.51 x 10(3)/kg v 6.18 +/- 2.66 x 10(3)/kg, P < or = .375). A statistically significant correlation (r = -.658, P < or = .05) was observed between the number of CFU-Blast infused and the number of days required to reach an ANC > or = 500 x 10(6) cells/L.(ABSTRACT TRUNCATED AT 400 WORDS)     

供者用粒细胞集落刺激因子后骨髓细胞成分的变化及移植后临床分析

目的了解供者应用粒细胞集落刺激因子(G-CSF)后骨髓细胞成分的改变及移植后促造血重建和减轻移植物抗宿主病(GVHD)的疗效.方法供者应用(研究组)和未用G-CSF(对照组)各12例进行异基因骨髓移植,研究组供者接受G-CSF(Lenograstim) 250 μg/d 连用7 d后采髓,比较研究组和对照组植入物造血成分CD+34、粒-单细胞集落形成单位(CFU-GM)、巨核细胞集落形成单位(CFU-MK)和CD+3及亚群的改变,及移植后对造血重建和急性GVHD的影响.结果两组在采集大致相同体积的骨髓植入物中,研究组采集的有核细胞数(TNC),CD+34,CFU-GM和CFU-MK明显高于对照组(P＜0.01),两组CD+3类同,CD+4减少,CD+8增加(P＞0.05),CD+4/CD+8明显减少(P＜0.01).研究组和对照组骨髓CD+34和T淋巴细胞亚群百分数及CFU-GM与CFU-MK增殖进行分析,显示类似变化的结果.研究组移植后中性粒细胞>0.5×109/L的时间是16 d(11～23),血小板>20×109/L的时间为17 d(14～25),对照组分别为20.5 d(14～29)和23.0 d(17～32)(P＜0.05).研究组无1例发生急性Ⅱ～Ⅳ GVHD,对照组3例发生急性Ⅱ～Ⅳ GVHD,两组比较差异无显著性(P＞0.05).结论异基因骨髓移植供者应用G-CSF 可促进造血恢复作用,这与增加植入物中CD+34、CFU-GM和CFU-MK数量有关,具有降低急性重度GVHD发生的倾向.     

In vivo administration of recombinant human granulocyte/macrophage colony-stimulating factor in acute lymphoblastic leukemia patients receiving purged autografts [see comments]

Based on the recent reports that recombinant human granulocyte/macrophage colony-stimulating factor (rhGM-CSF) accelerates the rate of engraftment in a variety of autologous bone marrow transplantation settings, we have investigated its effects on hematopoietic recovery of patients with acute lymphoblastic leukemia (ALL) undergoing autologous bone marrow transplantation. Our studies, which involved 25 autologous ALL recipients who received rhGM-CSF and 27 controls similar for disease status (remission or relapse) and disease type (B- or T-lineage) differed from previous studies in one important aspect: the bone marrows were purged with 4- hydroperoxcyclophosphamide (4HC) and anti-T or anti-B-cell lineage- specific antibodies before transplantation. Such treatments frequently lead to a reduction in the CFU-GM content of the transplanted marrow. Eighteen of 25 patients completed the entire course of rhGM-CSF. Of the 16 patients who received greater than or equal to 64 micrograms/M2/d for at least eight days, there were five patients who had an apparent rhGM-CSF response and 11 patients who did not respond. Of the parameters analyzed, only the number of CFU-GM progenitor cells infused per kilogram was significantly associated with an rhGM-CSF response. All patients receiving greater than or equal to 1.2 x 10(4) CFU-GM progenitors per kilogram achieved an absolute neutrophil count (ANC) greater than or equal to 1,000/microL by day 21 and had a greater than 50% decrement in ANC within 48 to 72 hours of discontinuing rhGM-CSF, as contrasted to none of the patients receiving less than or equal to 7.2 x 10(3) CFU-GM progenitors per kilogram.(ABSTRACT TRUNCATED AT 250 WORDS)     

Prolonged thrombocytopenia after autologous bone marrow transplantation

Thirty two patients with hematologic malignancies and solid tumors were treated with intensive therapy and autologous bone marrow transplantation. In nine out of 32 patients, it took more than 50 days to achieve a sustained platelet count of 50,000/microliter or greater. Significant associations with poor platelet recovery were found for patient age, diseases, period of cryopreservation, the kinds of eradicative therapy and in vitro purging. But most of these factors overlapped each other in the same patients. No correlation was found between platelet recovery and number of cells or CFU-GM infused.     

Clonogenic cells in acute myeloblasts leukaemia

Quantitative and qualitative changes in granulocyte-macrophage (CFU-GM) and fibroblast colony-forming cells (CFU-F) were studied in 7 patients with primary myelofibrosis (MF). Marrow cells were collected from bone biopsy specimens after treatment with collagenase. The number of CFU-GM correlated with the amount of haemopoietic tissue noted in the bone marrow histology and ranged between 0–400/mg of bone. CFU-F were increased in 2 patients with moderate fibrosis. Circulating CFU-GM were increased in all patients studied (169–3749/ml of blood). There was no significant correlation between the number of CFU-GM in the bone marrow and that in the blood. Cytochemical studies showed a high incidence in eosinophil progenitors in the bone marrow and especially in the blood of patients with MF. These data suggest a functional abnormality of myeloid progenitors in this disease.     

Clinical application of Percoll gradient-separated bone marrow

We have evaluated the transplantation potential of bone marrow stem cell concentrates isolated from the 40/60% interface of discontinuous Percoll gradients. This mononuclear fraction is free from platelets and depleted of granulocytes, and contains the majority of granulocyte-macrophage colony-forming cells (GM-CFC), erythroid burst-forming units (BFU-E), and granulocyte, erythroid, macrophage, megakaryocyte colony-forming cells (GEMM-CFC) in less than 10% of the cell number of the original buffy coat. This preparation allows further manipulation without the clumping and cell loss associated with buffy coat cell preparations. Cells isolated by this technique were evaluated for hematopoietic restoration potential in 14 patients who received allogeneic bone marrow transplants as supportive therapy after high dose cytoreduction to treat leukemias or lymphoma. The number of nucleated cells infused varied from 1.6-5.5 X 10(7)/kg, and the number of GM-CFC infused ranged from 0.4 to 3.7 X 10(5)/kg. There was an inverse relationship between the time to recovery of granulocytes and platelets and the number of GM-CFC infused when fewer than 10(5) GM-CFC/kg were transplanted. Above this dose, there was recovery within 10-15 days after transplantation. The stem cell-enriched fraction contained 30-40% of the original number of T lymphocytes, and acute graft-versus-host disease was observed in seven of these patients.     

Granulocyte/macrophage progenitor cells from peripheral blood and bone marrow differ in their response to prostaglandin E1

Prostaglandins of the E series (PGE) inhibit proliferation of normal bone marrow granulocyte/macrophage progenitors (CFU-GM). Circulating CFU-GM are known to differ from marrow CFU-GM in many characteristics, and in the present study, we compared the effect of PGE1 on circulating and bone marrow progenitors in normals and in patients with chronic myelogenous leukemia (CML). PGE1 caused a dose-dependent inhibition of normal marrow CFU-GM. Circulating CFU-GM were inhibited only at concentrations of 10(-5) mol/L or greater, and progenitor proliferation was, in fact, significantly stimulated at PGE1 concentrations between 10(-8) and 10(-6) mol/L. Bone marrow CFU-GM from patients with CML were inhibited in a manner similar to that of normal bone marrow. Circulating cells from patients with CML were, however, less sensitive to PGE1 inhibition than CML bone marrow cells and demonstrated a pattern intermediate between normal circulating and normal marrow progenitors. These studies suggest that peripheral blood and bone marrow contain different progenitor cell populations.     

A comparison between regimens containing chemotherapy alone (busulfan and cyclophosphamide) and chemotherapy (V. RAPID) plus total body irradiation on marrow engraftment following allogeneic bone marrow transplantation

The effect of two conditioning regimens given prior to allogeneic bone marrow transplantation (BMT) on the kinetics of engraftment were compared. 5 patients received busulfan and cyclophosphamide: 7 patients received daunorubicin, vincristine, cytosine arabinoside, methylprednisone and VM-26 plus total body irradiation (TBI). Bone marrow progenitors (BFU-E, CFU-E, CFU-GM, CFU-F) were assayed up to 3 months post-BMT. All progenitors were severely depressed in spite of peripheral blood recovery. There was no stromal recovery in any adult patient post-BMT. There was no significant difference in time to engraftment, or colony forming units, or between patients conditioned with chemotherapy alone or chemotherapy plus TBI. We were unable to detect effects of graft-versus-host disease or cytomegalic viral infection on bone marrow progenitors or peripheral blood recovery in this study.     

Interleukin-1 accelerates murine granulocyte recovery following treatment with cyclophosphamide

This study investigated the effects of recombinant human interleukin-1 (rhIL-1 alpha) on granulocyte recovery following treatment of mice with cyclophosphamide (CPM). CF1 mice were injected with 0.5 microgram rhIL- 1 alpha or heat-inactivated rhIL-1 alpha according to five different regimens, before and/or following 200 mg/kg CPM. Significant neutrophilia initially developed in treatment mice of all five regimens and accelerated granulocyte recovery occurred in treatment mice of four IL-1 regimens. Significant elevations in serum colony stimulating activity (CSA) occurred in treatment mice at a number of time points studied. In addition, marked increases in the percentage of maturing granulocyte precursors and in the proportion of cells cycling in S and G2/M were observed in treatment marrow throughout the IL-1 regimen. Before granulocyte recovery, premature nuclear segmentation was noted in metamyelocytes of treatment marrow. Concomitant with granulocyte recovery, treatment marrow was significantly more cellular and contained more total CFU-GM, more CFU-GM in S phase, more cells in S and G2/M, and more mitotic figures than control marrow. Splenic myelopoiesis was also enhanced in treatment mice. These data suggest that IL-1 significantly hastens granulocyte recovery following treatment with CPM by enhancing both proliferation and maturation of myeloid precursors.     

Detection of myeloid precursors (granulocyte/macrophage colony forming units) in the bone marrow adjacent to rheumatoid arthritis joints.

Various cytokines were recently found to be involved in the pathogenesis of rheumatoid arthritis (RA) and particularly, cytokines with hematopoietic activity have been detected in synovial tissues. We counted the number of myeloid precursors in terms of granulocyte/macrophage colony forming units (CFU-GM) and the number of stromal cell progenitors in terms of fibroblast colony forming units (CFU-F) in the tibial bone marrow adjacent to the joints affected by RA (n = 21), osteoarthritis (OA) (n = 10), and trauma (n = 2) using the colony formation unit assay. We also quantitated the amounts of interleukin 1 beta (IL-1 beta), IL-6, and granulocyte/macrophage colony stimulating factor (GM-CSF) in the culture supernatant of synovial tissue explants of these patients by enzyme linked immunosorbent assay (ELISA). The mean number (+/- SEM) of CFU-GM in patients with RA (7.4 +/- 4.9) was greater than that in patients with OA (0.5 +/- 0.2), while CFU-GM was not detected in trauma patients. The number of CFU-GM in the tibial bone marrow of patients with RA correlated well with the amount of IL-1 beta (r = 0.64, p < 0.01), but not with GM-CSF or with IL-6 from synovial tissues. These findings suggest that active bone marrow is present adjacent to the affected joints in patients with RA and that hematopoietic activity is influenced by IL-1 beta produced in nearby synovial tissues.     

Expression of Ia-like antigens defined by monoclonal OKIal antibody of hemopoietic progenitor cells in cord blood: a comparison with human bone marrow

Expression of Ia antigens on granulocyte/macrophage colony-forming cells (CFU-GM) in human cord blood was compared with that in bone marrow with the use of monoclonal OKIal antibody. Mononuclear cells prepared from cord blood and bone marrow were pretreated with OKIal antibody plus complement, and, thereafter, the ability of cord blood and bone marrow cells to form colonies of CFU-GM was assayed in semisolid agar culture. Consistent reduction in the number of CFU-GM in cord blood to 58.8% +/- 13.0% (mean +/- SD) of controls treated with complement alone was shown after elimination of Ia-antigen-bearing CFU- GM, but was significantly remarkable than that in bone marrow (18.0% +/- 5.6%). Although the reduction of both granulocyte (CFC-G) and macrophage colony (CFC-M) types of cord blood, characterized by the double staining for esterase activity, was shown following treatment with OKIal antibody plus complement, the relative inhibition of CFC-G weas significantly greater than that of CFC-M (p less than 0.02). These results suggest some differences in the characteristics of Ia-antigen- bearing CFU-GM between cord blood and bone marrow cells. Furthermore, it is suggested that Ia-dependent regulatory mechanisms might participate in the differentiation of CFU-GM to CFC-G and CFC-M.     

Recovery of hematopoiesis after high-dose chemotherapy and peripheral blood stem cell autografts in children]

Hematopoietic recovery kinetics were evaluated in 34 children with therapy-refractory malignant tumors who underwent a total of 35 peripheral blood stem cell autografts (PBSCT) after marrow-ablative chemotherapy without total body irradiation. A negative correlation was found between the numbers of colony-forming units of granulocyte-macrophage (CFU-GM) infused per kilogram of the patients' body weight and the time of achieving an absolute granulocyte count (AGC) of 0.5 x 10(9)/l or a platelet count of 50 x 10(9)/l (granulocyte: r = -0.631, p less than 0.001, platelet: r = -0.590, p less than 0.001). The patients were classified into three groups; 14 patients who received less than 1 x 10(5) CFU-GM/kg (group A), 7 patients who received 1-3 x 10(5) CFU-GM/kg (group B), and 14 patients who received greater than or equal to 3 x 10(5) CFU-GM/kg (group C). The AGC recovered to 0.5 x 10(9)/l by 21 day in group A, 14 day in group B, and 10 day in group C. The platelet count recovered to 50 x 10(9)/l 102 in group A, 23 in group B, and 16 day in group C patients. The final platelet infusion was on day 60 in group A, day 12 in group B, and day 12 in group C. Transient decrease in the blood cell count developed in all patients 3 to 7 weeks after transplantation, and two cases developed reversible ITP. In the remaining patients, the recovered hematopoietic function was sustained for 1-48 months after transplant action.(ABSTRACT TRUNCATED AT 250 WORDS)     

Characteristics of engraftment after repeated autologous bone marrow transplantation

The rate of engraftment after autologous bone marrow transplantation (ABMT) is extremely variable and largely unpredictable. To identify factors influencing engraftment, we studied 35 patients with refractory germ cell tumors undergoing high-dose chemotherapy with carboplatin (900-2000 mg/m2) and etoposide (1200 mg/m2) with bone marrow rescue. Prior to the initiation of chemotherapy, bone marrow sufficient for two marrow infusions was harvested (range 0.86-4.82 x 10(8) nucleated cells per kg). All 35 patients received half of the collected bone marrow 3 days after the last dose of chemotherapy; 23 responders received a second round of the same chemotherapy followed by infusion of the second half of the bone marrow. Eighteen patients could be compared for the two transplant episodes. The "rate of engraftment" was defined as the unweighted mean of four parameters: 1) the number of days until the absolute granulocyte count surpassed 0.2 x 10(9)/liter, 2) the number of days until the absolute granulocyte count surpassed 0.5 x 10(9)/liter, 3) the number of days until the last platelet transfusion, and 4) the number of days until the reticulocyte count surpassed 25 x 10(9)/liter. No significant correlation was found between rate of engraftment and such factors as the number of nucleated cells per kg infused, the dose of chemotherapy, extent of prior chemotherapy, tumor response to the high-dose chemotherapy, age of the patient, or the days of granulocytopenic fever (all p greater than 0.20). In contrast, a close correlation was found for the number of units of platelets (p = 0.005) and red blood cells (p = 0.006) transfused following each of the two transplants. There was no significant difference between rate of engraftment after first and second transplantation. Comparison of these data with the results obtained in reported ABMT with separate harvests suggests that the characteristics of the infused marrow determine the rate of engraftment after ABMT. This model of repeated transplantation could provide an important tool for assessing the therapeutic efficacy of hematopoietic growth factors.