Mixed chimerism in SCT: conflict or peaceful coexistence?

Stem cell transplants that follow both myeloablative and non-myeloablative conditioning regimens can result in states of mixed chimerism, which can be stable over time. With widespread availability of Y chromosome FISH in sex-mismatched transplantation and DNA-based methodologies for analysis of chimerism in other donor-recipient pairs, further insights have been gained regarding the implications of the mixed chimeric state. In transplants performed for inherited and acquired marrow failure disorders, disease status can be improved with only 10-20% donor cells, and it appears that stable mixed chimerism at that level is an acceptable outcome often leading to a state of tolerance, but an increasing level of recipient cells often precedes graft rejection. In transplants performed for malignant conditions, increasing levels of mixed chimerism may indicate disease relapse, but some cases with stable levels of mixed chimerism have been compatible with prolonged remission states. Understanding when mixed chimerism is an indication of secondary graft failure or impending graft rejection vs a state of tolerance and ongoing propensity for the establishment of a graft-vs-tumor effect is often difficult with currently available technologies and immunologic assays. The ability to understand the implication of mixed chimerism of multiple cell lineages and of varied lymphocyte subsets will remain important areas for future research to best harness the immunologic and other therapeutic benefits of allogeneic transplantation.

Pre-existing invasive fungal infection is not a contraindication for allogeneic HSCT for patients with hematologic malignancies: a CIBMTR study.

Patients with prior invasive fungal infection (IFI) increasingly proceed to allogeneic hematopoietic cell transplantation (HSCT). However, little is known about the impact of prior IFI on survival. Patients with pre-transplant IFI (cases; n=825) were compared with controls (n=10247). A subset analysis assessed outcomes in leukemia patients pre- and post 2001. Cases were older with lower performance status (KPS), more advanced disease, higher likelihood of AML and having received cord blood, reduced intensity conditioning, mold-active fungal prophylaxis and more recently transplanted. Aspergillus spp. and Candida spp. were the most commonly identified pathogens. 68% of patients had primarily pulmonary involvement. Univariate and multivariable analysis demonstrated inferior PFS and overall survival (OS) for cases. At 2 years, cases had higher mortality and shorter PFS with significant increases in non-relapse mortality (NRM) but no difference in relapse. One year probability of post-HSCT IFI was 24% (cases) and 17% (control, P<0.001). The predominant cause of death was underlying malignancy; infectious death was higher in cases (13% vs 9%). In the subset analysis, patients transplanted before 2001 had increased NRM with inferior OS and PFS compared with later cases. Pre-transplant IFI is associated with lower PFS and OS after allogeneic HSCT but significant survivorship was observed. Consequently, pre-transplant IFI should not be a contraindication to allogeneic HSCT in otherwise suitable candidates. Documented pre-transplant IFI is associated with lower PFS and OS after allogeneic HSCT. However, mortality post transplant is more influenced by advanced disease status than previous IFI. Pre-transplant IFI does not appear to be a contraindication to allogeneic HSCT.

One hundred autotransplants for relapsed or refractory Hodgkin's disease and lymphoma: value of pretransplant disease status for predicting outcome.

PURPOSE: One hundred autotransplants for Hodgkin's disease (HD) or non-Hodgkin's lymphoma (NHL) were examined prospectively to identify variables with prognostic significance. PATIENTS AND METHODS: Ninety-six patients with relapsed or refractory HD or NHL underwent 100 autotransplants. Patients received high-dose carmustine (BCNU), etoposide, cytarabine, and cyclophosphamide (BEAC) followed by unpurged autologous stem-cell rescue. RESULTS: The 3-year actuarial event-free survival (EFS) rate for the 47 HD patients is 49%, with a median followup duration of 2 years. For the 53 NHL patients, the 3-year actuarial EFS rate is 40%, with a median follow-up duration of 19 months. By multivariate analysis, minimal disease on admission (all areas < or = 2 cm) is associated with improved EFS (HD, P = .003, NHL, P = .03). The projected EFS rate for HD patients entering with minimal disease is 70% versus 15% for patients with bulky disease (P = .0001). The projected EFS rate for NHL patients with minimal disease is 48% versus 25% for patients with bulky disease (P = .04). Posttransplant involved-field radiotherapy, administered to 26 of the last 61 patients, was associated with an improved EFS rate for NHL patients (P = .015). The BEAC regimen was well tolerated by patients who entered the study with minimal disease (mortality rate, < 5%), but caused significant toxicity in patients with bulky disease (mortality rate, 25%). CONCLUSION: Disease burden before autotransplantation is an important predictor of regimen-related toxicity and EFS. Posttransplant involved-field radiotherapy may improve outcomes in select patients with NHL. The BEAC regimen is safe and effective, particularly for patients with minimal disease.

Analysis of factors that correlate with mucositis in recipients of autologous and allogeneic stem-cell transplants.

PURPOSE: To identify predictors of oral mucositis and gastrointestinal toxicity after high-dose therapy. PATIENTS AND METHODS: Mucositis and gastrointestinal toxicity were prospectively evaluated in 202 recipients of high-dose therapy and autologous or allogeneic stem-cell rescue. Of 10 outcome variables, three were selected as end points: the peak value for the University of Nebraska Oral Assessment Score (MUCPEAK), the duration of parenteral nutritional support, and the peak daily output of diarrhea. Potential covariates included patient age, sex, diagnosis, treatment protocol, transplantation type, stem-cell source, and rate of neutrophil recovery. The three selected end points were also examined for correlation with blood infections and transplant-related mortality. RESULTS: A diagnosis of leukemia, use of total body irradiation, allogeneic transplantation, and delayed neutrophil recovery were associated with increased oral mucositis and longer parenteral nutritional support. No factors were associated with diarrhea. Also, moderate to severe oral mucositis (MUCPEAK > or = 18 on a scale of 8 to 24) was correlated with blood infections and transplant-related mortality: 60% of patients with MUCPEAK > or = 18 had positive blood cultures versus 30% of patients with MUCPEAK less than 18 (P =.001); 24% of patients with MUCPEAK > or = 8 died during the transplantation procedure versus 4% of patients with MUCPEAK less than 18 (P =.001). CONCLUSION: Gastrointestinal toxicity is a major cause of transplant-related morbidity and mortality, emphasizing the need for corrective strategies. The peak oral mucositis score and the duration of parenteral nutritional support are useful indices of gastrointestinal toxicity because these end points are correlated with clinically significant events, including blood infections and treatment-related mortality.

What would Karl Landsteiner do? The ABO blood group and stem cell transplantation.

ABO blood group antigens, of great importance in transplantation and transfusion, are present on virtually all cells, as well as in soluble form in plasma and body fluids. Naturally occurring plasma IgM and IgG antibodies against these antigens are ubiquitous. Nonetheless, the ABO blood group system is widely ignored by many transfusion services, except for purposes of red cell transfusion. We implemented a policy of transfusing only ABO identical platelets and red cells in patients undergoing stem cell transplantation or treatment for hematologic malignancies. Major bleeding episodes have occurred in about 5% of patients undergoing induction therapy for acute leukemia as compared with 15-20% in the literature. Overall survival times appear to be superior to that in historical cohorts. In 2002-2004, treatment-related mortality at 100 days in our Blood and Marrow Transplant Unit was 0.7% for autologous transplants (n=148), 13% for sibling allogeneic transplants (n=110), and 24% (n=62) for matched unrelated allogeneic transplants, suggesting that our approach is safe. We speculate that more rigorous efforts on the part of transfusion services to provide ABO identical blood components, and to remove incompatible supernatant plasma, when necessary, might yield reduced morbidity and mortality in patients undergoing stem cell transplantation.

Hematopoietic stem cell transplantation for multiply transfused patients with sickle cell disease and thalassemia after low-dose total body irradiation, fludarabine, and rabbit anti-thymocyte globulin.

Patients with sickle cell disease (N = 3) and thalassemia (N = 1) with high-risk features received hematopoietic stem cell transplantations (HCT) to induce stable (full or partial) donor engraftment. Patients were 9-30 years of age. Fludarabine, rabbit anti-thymocyte globulin (ATG), and 200 cGy total body irradiation were administered pre-transplant. Patients received bone marrow (N = 3) or peripheral blood stem cells (N = 1) from HLA-identical siblings, followed by mycophenolate mofetil and cyclosporine for post-grafting immunosuppression. Significant lymphopenia, but only moderate neutropenia and thrombocytopenia developed post transplant. No grade IV nonhematological toxicities or acute graft-versus-host disease (GVHD) were observed. At 3 months after transplantation, three of four patients had evidence of donor myeloid chimerism (range, 15-100%). However, after post transplant immunosuppression was discontinued, graft rejection occurred in all but one patient. This patient is now doing well 27 months post transplant with full donor engraftment. One patient died after a second transplant, and another patient experienced a stroke as her graft was being rejected. These results suggest that stable donor engraftment after nonmyeloablative HCT is difficult to achieve among immunocompetent patients with hemoglobinopathies and that new approaches will need to be developed before wider application of this transplantation method for hemoglobinopathies.

Hematopoietic growth factors in acute leukemia.

The hematopoietic growth factors are glycoproteins that can be produced by recombinant DNA technology. They have many potential clinical uses in acute leukemia; several areas have been explored extensively and much data are available from clinical trials. Other areas are of potential interest, but have a paucity of clinical information. The past decade has seen major strides in the development and clinical application of cytokines in acute leukemia and it is expected that this trend will continue over the next decade as further areas are explored and results of clinical trials mature to enable us to determine the precise role of cytokines in the clinical setting.

Characterization of the adherence of normal and leukemic CD34+ cells to endothelial monolayers.

The capacity of normal CD34+ marrow cells and CD34+ leukemic cell lines to adhere to human umbilical vein endothelial cells has been examined. Such interactions have importance since the processes of homing and egress within the marrow microenvironment involve the traverse of sinusoidal endothelium. Umbilical vein endothelial monolayers expressed CD44 and CD54 constitutively, and expression of both E-selectin (ELAM) and vascular cell adhesion molecule-1 (VCAM-1) were inducible with interleukin-1 (IL-1) alpha and beta and tumor necrosis factor (TNF). CD34+ marrow cells bound to unstimulated endothelial layers (33 vs. 16% to plastic), and their adhesion was significantly increased in the presence of IL-1 or TNF. This increased adhesion was not inhibited by functionally blocking antibodies to E-selectin or to CD54 but was partially inhibited by antibodies to VCAM. CD34+ KG1a cells also bound to endothelial monolayers (33 vs. 8% to plastic), and such adhesion was also upregulated by pretreatment of the endothelial cells with IL-1 or TNF. In contrast to normal CD34+ cells, this increased adhesion was inhibited by antibodies to E-selectin but not to VCAM. These findings indicate that adhesion of both normal CD34+ cells and leukemic blasts to endothelial cells can be upregulated by inflammatory mediators such as TNF and IL-1.

Effects of the farnesyl transferase inhibitor R115777 on normal and leukemic hematopoiesis.

Patients with acute myelogenous leukemia or myelodysplastic syndrome may respond to farnesyl transferase inhibitors (FTIs) with partial or complete response rates noted in about 30% of such patients. FTIs prevent the attachment of a lipid farnesyl moiety to dependent proteins prior to their insertion into the plasma membrane and thereby prevent activity of these prenylation-dependent proteins, but their mechanism of tumor suppression remains unknown. Many patients receiving FTIs do experience myelosuppression. In this work, the in vitro effects of the FTI, R115777 on normal and leukemic hematopoiesis have been examined as have its effects on apoptosis induction and cell cycle profile in both leukemic blasts and normal CD34+ cells. R115777 was inhibitory to normal CD34+ cell proliferation and to leukemic blast cells, but did not affect long-term culture initiating cell frequency nor NOD-SCID reconstituting capacity. No induction of apoptosis or cell cycle changes were noted in AML blasts. These data suggest that myelosuppression with R115777 occurs largely at the intermediate to late progenitor stage of hematopoiesis and that cyclic use might avoid long-term marrow suppression.

Effect of stem cell factor on myelopoiesis potential in human Dexter-type culture systems.

Hematopoiesis is influenced by the presence of the hematopoietic microenvironment, and Dexter-type liquid culture systems represent an in vitro representation of some aspects of the microenvironment that are optimal for the propagation of myeloid progenitors. Marrow stromal layers, which constitute part of these culture systems, produce growth factors, including stem cell factor (SCF), a ligand for the c-kit proto-oncogene that has been found to increase detection of myeloid, erythroid, and megakaryocytic progenitors in short-term marrow colony assays. In this work, the role of SCF in Dexter-type culture systems was examined to better define its contribution to steady-state myelopoiesis. When cultured in the continued presence of 100 ng/mL SCF, both primary and recharged cultures demonstrated significantly greater CFU-GM output, with quantitative differences noted throughout culture duration (up to 6 weeks). This increase in CFU-GM could be inhibited specifically with the addition of 1:1500 SR-1, a neutralizing anti-c-kit monoclonal antibody (MAb) that neutralizes the biological effects of SCF, and the increase was noted both with recharged light-density marrow cells and purified CD34+ progenitor cells. On the other hand, when primary or recharged marrow cultures were established in the absence of exogenous SCF, but in the continuous presence of SR-1, no inhibition of CFU-GM output was observed. When light-density marrow cells were purged of pre-existing CFU-GM by 4-hydroperoxycyclophosphamide (4-HC) and were seeded over irradiated stromal layers, exogenous SCF resulted in detection of CFU-GM from 4-HC-treated cells as early as 1 week of culture, as compared to the lack of significant emergence of CFU-GMs at 4 weeks in the control cultures. This SCF effect was also inhibited by SR-1. Purified CD34+ progenitor cells did not adhere to SCF immobilized to tissue culture plates, and the adhesion of such progenitors to murine Steel lines transfected with membrane-bound SCF was not greater than to the parent nontransfected Steel line, suggesting that the effect of SCF was not on CD34+ cell adhesion. These studies confirm the action of SCF at a pre-CFU level, and they demonstrate the ability of SCF to stimulate increased production of myeloid progenitors in long-term liquid culture systems.

Phenotypic characterization of the human fibrous histiocytoma giant cell tumor (GCT) cell line and its cytokine repertoire.

The pleiotropic nature of malignant fibrous histiocytomas (MFH) is manifested as mixed cellular infiltrates consisting of myofibroblasts, histiomonocytes, and neutrophils. We detail in this report the phenotypic characteristics of the human fibrous histiocytoma giant cell tumor (GCT) cell line that establish its mesenchymal origin. The latter is underscored by the ability of GCT cells to express mRNA for transforming growth factor beta (TGF-beta) as well as both A and B chains of platelet-derived growth factor (PDGF). GCT cells also support the binding of CD34+ cells, but less efficiently than do normal marrow stromal cells. Since cytokines elaborated by MFH may mediate in part the recruitment of monocytes and neutrophils into tumor-infiltrated tissues, we have determined the cytokine repertoire of the GCT cell line, already known for its ability to elaborate colony-stimulating factors (CSFs) and interleukin-1 (IL-1). GCT cells express IL-1 alpha, IL-1 beta, IL-6, macrophage colony-stimulating factor (M-CSF or CSF-1), granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), and IL-8. No detectable mRNA for IL-3, IL-4, IL-7, and tumor necrosis factor-alpha (TNF-alpha) was detected in GCT cells by polymerase chain reaction (PCR). Expression of cytokine mRNAs was responsive to agents such as dexamethasone (dex), 12-O-tetradecanoyl phorbol 13-acetate (phorbol diester or TPA), and TNF-alpha. Thus, this cell line provides a useful model for understanding the pathobiology of MFH and hematopoietic progenitor interactions with mesenchymal/stromal cells.

Adhesive interactions of normal and leukemic human CD34+ myeloid progenitors: role of marrow stromal, fibroblast, and cytomatrix components.

Adhesive interactions between CD34+ myeloid progenitors, cytomatrix components, and marrow fibroblast and stromal monolayers are described and compared to the binding interactions of the CD34+ myeloid leukemic cell lines KG1a and KG1. Both normal precursors and their leukemic counterparts showed adhesion to marrow stroma and fibroblasts. CD34+ myeloid progenitors bound to the extracellular matrices of marrow stromal cell and fibroblast monolayers and to laminin and fibronectin to a lesser extent than to cellular stromal layers. These adhesive interactions were not inhibited by polyclonal antibodies to laminin or fibronectin, nor by 1 mM Arg-Gly-Asp-Ser (RGDS)-containing peptides. Also, although both normal and leukemic cells expressed the CD18 antigen, binding of these cells to stroma was not inhibited by blocking anti-CD18 monoclonal antibodies. Finally, KG1a adhesion was not blocked in the presence of anti-CD54 (ICAM) antibody, nor was it blocked when galactosyl or mannosyl pyranosides were added. KG1a binding was trypsin sensitive and enhanced in the presence of neuraminidase. These studies serve to characterize adhesive properties of normal and leukemic myeloid progenitors and begin to establish interactions important for the lodgement of early progenitor cells in human marrow.

State of the art; the hypereosinophilic syndromes.

Many disease states such as parasitic infestations, malignancies, collagen vascular diseases, and allergies are associated with eosinophilia. The diagnosis of idiopathic hypereosinophilic syndrome (HES) requires a persistent elevation in the total eosinophil count (greater than 1500/mm3) for over 6 months, associated organ damage and no detectable underlying cause. This review provides an updated summary of the cytokine cascade that controls eosinophil production and delineates our current understanding of the clinical features of hypereosinophilic states. We also examine the central role of T-lymphocyte activation in eosinophilia, and have attempted to integrate current treatment strategies for HES with the physiology of eosinophilopoiesis.

Mechanisms of cytopenia in human immunodeficiency virus infection.

Human immunodeficiency virus (HIV) infection often has effects on the hematopoietic system which can be distinguished from the concurrent effects of medications or opportunistic infections. Exactly how the virus mediates these effects remains uncertain, but both in vivo and in vitro studies have pointed up possible direct and indirect modes of hematopoietic suppression. Whether a significant fraction of CD34+ cells in vivo are infected with HIV remains controversial, but most studies using in situ polymerase chain reaction techniques would suggest not. Other more indirect modes of hematopoietic cell suppression such as production of autoantibodies, production of other humoral inhibitory factors, T-cell mediated suppression of hematopoiesis, or production of inhibitory or stimulatory cytokines may also be contributory. It is probable that several of these mechanisms may occur simultaneously, and an increased understanding of their role may lead to improved strategies to correct the cytopenias which often accompany HIV disease.

Phase I study for poor-prognosis lymphoma: augmentation of the "BEAM" regimen with escalating dose melphalan using amifostine cytoprotection and autologous hematopoietic stem cell transplantation--a preliminary report.

We and others have previously shown that the use of amifostine (Ethyol; MedImmune Inc, Gaithersburg, MD) can ameliorate certain regimen-related toxicities of high-dose melphalan (HD-MEL) in the autologous hematopoietic stem cell transplant setting. Our recent experience indicated that the maximum tolerated dose of HD-MEL plus autologous hematopoietic stem cell transplant could be increased from approximately 200 mg/m2 to at least 280 mg/m2 with amifostine. Although a dose-limiting toxicity was not clearly identified, atrial fibrillation was noted in several patients. Phase II trials using this regimen have been reported in lymphoma and myeloma. Nonetheless, it is unlikely that single agent therapy, regardless of dose, will be highly curative in advanced hematologic malignancy. Thus, we used amifostine to permit dose escalation of HD-MEL within the BEAM (BCNU/etoposide/arabinosylcytosine/HD-MEL) combination chemotherapy regimen before autologous hematopoietic stem cell transplant in selected patients with lymphoma. Patient entry at the starting dose (ie, HD-MEL 140 mg/m2) has been completed without the development of severe regimen-related toxicities. This trial is ongoing.

Hematopoietic stem cell compartment: acute and late effects of radiation therapy and chemotherapy.

The bone marrow is an important dose-limiting cell renewal tissue for chemotherapy, wide-field irradiation, and autologous bone marrow transplantation. Over the past 5-10 years a great deal has been discovered about the hematopoietic stem cell compartment. Although the toxicity associated with prolonged myelosuppression continues to limit the wider use of chemotherapy and irradiation, ways are being discovered to circumvent this toxicity such as with the increasing use of cytokines. This review describes what is known of how chemotherapy and irradiation damage stem cells and the microenvironment, how cytokines protect hematopoietic cells from radiation damage and speed marrow recovery after chemotherapy or marrow transplantation, and how various types of blood marrow cells contribute to engraftment and long-term hematopoiesis after high doses of cytotoxic agents and/or total body irradiation.

Stem cell factor and stromal cell co-culture prevent apoptosis in a subculture of the megakaryoblastic cell line, UT-7.

The megakaryoblastic cell line, UT-7, is dependent for its growth upon interleukin-3 (IL-3), erythropoietin, or granulocyte-macrophage colony stimulating factor (GM-CSF). A subculture of this line can be maintained in recombinant human c-kit ligand [stem cell factor (SCF)] at 100 ng/ml without requirement for other growth factors. Removal of this subculture from SCF results in rapid loss of viability and decreased proliferation. Cells grown in SCF also can be maintained in GM-CSF but not vice versa. In this work, we have characterized the SCF dependence of this UT-7 subculture. Stem cell factor removal results in apoptosis and a decline in viability which can be restored partially by re-addition of SCF, GM-CSF, or co-culture with adherent marrow stromal cells. Apoptosis in the factor-starved UT-7 population has been documented by light microscopy, electron microscopy and DNA analysis, showing the typical 180 base pair laddering characteristic of apoptosis. To quantitate the degree of apoptosis in the cell populations, and to assess whether apoptosis decreased with re-exposure of starved cells to growth factors or stroma, we utilized flow cytometry. This confirmed that exposure of previously factor-starved cells to stroma decreased the percentage of cells undergoing apoptosis. Co-culture with an SCF-deficient murine stromal cell line was also able to prevent apoptosis, suggesting contribution of other stromal cell factors. Experiments performed using trans-well inserts which do not allow cell passage, showed greatest viability of cells in contact with stroma, but viability was also improved in cells cultured in the presence of, but not in contact with, stromal cells compared to those cultured above plastic, suggesting a role for soluble stroma-produced substances. These data demonstrate that SCF alone can prevent apoptosis in cells dependent upon its presence for proliferation. Also, marrow stromal cells can serve as a partial substitute for growth factor in the prevention of apoptosis in these cells, probably due to constitutive presentation of SCF and other hematopoietic growth factors in both soluble and surface-bound forms.

Effects of GM-CSF on Ki67 expression and cell cycle traverse in acute myelogenous leukemia specimens and cell lines.

A new strategy in the treatment of acute myelogenous leukemia is to attempt to increase the growth fraction of clonal leukemic cells prior to administration of chemotherapeutic agents by the administration of hematopoietic growth factors. We have studied the effect of GM-CSF on the cell cycle status and Ki67 nuclear antigen expression of AML blasts in vitro. The effect of growth factors and stromal cell co-culture on Ki67 expression in leukemic cell lines was also examined. Neither stromal cell co-culture nor exposure of factor-dependent and factor-independent cell lines to GM-CSF, IL-3, SCF, or combinations thereof significantly changed the percentage of cells expressing Ki67. In the AML population analyzed as a whole, exposure of blasts to GM-CSF for up to 96 h did not significantly change the percentage of cells in S phase or expressing Ki67.

CD34+ progenitor cell isolation from blood and marrow: a comparison of techniques for small-scale selection.

The isolation and characterization of primitive hematopoietic cells and their purification in sufficient numbers is important in clinical and research marrow transplantation settings. As systems for large-scale isolation and amplification of such cells are developed, they may assume importance in transplantation, treatment of marrow failure and for gene therapy applications. Such cells have been isolated by numerous techniques and in this work, small-scale isolation of CD34+ cells by two immunoadsorption purification methods is compared with isolation by flow cytometry. While the immunoadsorption techniques allow for the processing of large numbers of density gradient-separated or unseparated cells for progenitor isolation, such techniques do not achieve the purity afforded by fluorescence activated cell sorter separation.

Survival after HLA-identical allogeneic peripheral blood stem cell and bone marrow transplantation for hematologic malignancies: meta-analysis of randomized controlled trials.

The impact of peripheral blood stem cell transplantation (PBSCT) on survival relative to bone marrow transplantation (BMT) remains poorly defined. Several randomized controlled trials (RCTs) comparing HLA-matched related PBSC- and BMT for patients with hematologic malignancies have been published, yielding differing results. We conducted a meta-analysis of published RCTs to more precisely estimate the effect of PBSCT on survival. Seven trials that assessed survival were identified and included in our analysis. Using a fixed effects model, and combining the results of all seven trials, the summary odds ratio for mortality after PBSCT was 0.81 (95% CI, 0.62-1.05) when compared to BMT. Subgroup analysis revealed no association between the median PBSCT 34+ cell dose and relative risk for morality after PBSCT. However, there was an association between the proportion of patients enrolled with advanced-stage disease and the summary odds ratio for mortality. The pooled estimate was 0.64 for studies where patients with intermediate/advanced disease comprised at least 25% of enrollment, and was 1.07 for the studies enrolling a smaller proportion. This finding substantiates results from previously published studies that have demonstrated a survival advantage with PBSCT limited to patients with advanced disease.