Effect of CD8⁺ cell content on umbilical cord blood transplantation in adults with hematological malignancies

Graft CD8 T-cell-based risk system predicts survival in antithymocyte globulin-based myeloablative haploidentical peripheral blood stem cell transplantation

Objective

This study investigated the cellular composition of peripheral blood grafts for anti-thymocyte globulin (ATG)-based myeloablative haploidentical haematopoietic stem cell transplantation (haplo-HSCT).

Methods

Clinical characteristics were retrospectively evaluated in a training cohort with ATG-based myeloablative haplo-HSCT between January 2016 and February 2020 and confirmed in a validation cohort between March 2020 and June 2021.

Results

A higher dose of graft CD8+ T cells (≥ 0.85 × 108 kg-1) was significantly improved overall survival (OS; hazard ratio [HR], 1.750; P = 0.002) and disease-free survival (DFS; HR, 1.751; P < 0.001) in the training cohort, according to multivariate Cox regression analysis. Higher doses of mononuclear cells (MNCs) demonstrated better OS (HR, 1.517; P = 0.038) and DFS (HR, 1.532; P = 0.027). Older patient age (> 46 years), older donor age (≥ 50 years) and a higher refined disease risk index (rDRI) were also related to OS. A graft CD8+ T-cell risk system based on graft CD8+ T-cell dose, donor age and rDRI was constructed using a nomogram model after LASSO Cox regression analysis. It showed acceptable discrimination, with a C-index of 0.62 and 0.63, respectively. Graft CD8+ T-cell dose was negatively correlated with donor age (P < 0.001) and positively correlated with a higher lymphocyte percentage in the peripheral blood before mobilisation (P < 0.001).

Conclusion

A higher CD8+ T-cell dose in peripheral blood-derived grafts improves patients' survival with ATG-based myeloablative haplo-HSCT. Younger donors with higher lymphocyte percentages improved patients' survival with an intermediate rDRI risk.

Transfusion Burden in Allogeneic Hematopoietic Stem Cell Transplantation over Time: Experience from a Single Institution

INTRODUCTION

Transfusion plays a main role in supportive treatment for patients who receive an allogeneic hematopoietic stem cell transplantation (HSCT). In this study, we compare the transfusion requirements of patients undergoing different modalities of HSCT according to different time periods. The objective is to assess the evolution of HSCT transfusion requirements over time, from a single institution.

METHODS

The clinical charts and transfusion records of patients who underwent HSCT of different modalities at La Fe University Hospital during a twelve-year period were reviewed (2009-2020). For analysis, we divided the overall time into three periods: 1 from 2009 to 2012, 2 from 2013 to 2016 and 3 from 2017 to 2020. The study included 855 consecutive adult HSCT: 358 HLA-matched related donors (MRD), 134 HLA-matched unrelated donors (MUD), 223 umbilical cord blood transplantation (UCBT) and 140 haploidentical transplants (Haplo-HSCT).

RESULTS

There were no significant differences in RBC and PLT requirements or transfusion independence among the three time periods for MUD and Haplo-HSCT. However, the transfusion burden increased significantly for MRD HSCT during the 2017-2020 period.

CONCLUSION

despite HSCT modalities having evolved and changed over time, overall transfusion requirements have not significantly decreased and continue to be a cornerstone of transplantation-supportive care.

Umbilical cord blood: an undervalued and underutilized resource in allogeneic hematopoietic stem cell transplant and novel cell therapy applications

PURPOSE OF REVIEW

The purpose of this review is to primarily discuss the unwarranted decline in the use of umbilical cord blood (UCB) as a source of donor hematopoietic stem cells (HSC) for hematopoietic cell transplantation (HCT) and the resulting important implications in addressing healthcare inequities, and secondly to highlight the incredible potential of UCB and related birthing tissues for the development of a broad range of therapies to treat human disease including but not limited to oncology, neurologic, cardiac, orthopedic and immunologic conditions.

RECENT FINDINGS

When current best practices are followed, unrelated donor umbilical cord blood transplant (CBT) can provide superior quality of life-related survival compared to other allogeneic HSC donor sources (sibling, matched or mismatched unrelated, and haploidentical) through decreased risks of relapse and chronic graft vs. host disease. Current best practices include improved UCB donor selection criteria with consideration of higher resolution human leukocyte antigen (HLA) typing and CD34+ cell dose, availability of newer myeloablative but reduced toxicity conditioning regimens, and rigorous supportive care in the early posttransplant period with monitoring for known complications, especially related to viral and other infections that may require intervention. Emerging best practice may include the use of ex vivo expanded single-unit CBT rather than double-unit CBT (dCBT) or 'haplo-cord' transplant, and the incorporation of posttransplant cyclophosphamide as with haploidentical transplant and/or incorporation of novel posttransplant therapies to reduce the risk of relapse, such as NK cell adoptive transfer. Novel, non-HCT uses of UCB and birthing tissue include the production of UCB-derived immune effector cell therapies such as unmodified NK cells, chimeric antigen receptor-natural killer cells and immune T-cell populations, the isolation of mesenchymal stem cells for immune modulatory treatments and derivation of induced pluripotent stem cells haplobanks for regenerative medicine development and population studies to facilitate exploration of drug development through functional genomics.

SUMMARY

The potential of allogeneic UCB for HCT and novel cell-based therapies is undervalued and underutilized. The inventory of high-quality UCB units available from public cord blood banks (CBB) should be expanding rather than contracting in order to address ongoing healthcare inequities and to maintain a valuable source of cellular starting material for cell and gene therapies and regenerative medicine approaches. The expertise in Good Manufacturing Practice-grade manufacturing provided by CBB should be supported to effectively partner with groups developing UCB for novel cell-based therapies.

Cord blood CD8+ T-cell expansion following granulocyte transfusions eradicates refractory leukemia

The action of hematopoietic cell transplantation in controlling leukemia is principally mediated by donor T cells directed against residual recipient malignant cells. However, its utility is limited by graft-versus-host disease (GVHD), where alloreactivity is extended beyond leukemic and marrow cells. In a human/murine chimeric model, we previously showed that the preferential infiltration of cord blood (CB) CD8+ T cells eradicates an Epstein-Barr virus-driven lymphoblastoid tumor without causing xenogeneic GVHD. In the clinic, however, cord blood CD8+ T-cell reconstitution is significantly delayed, and the observation of such a robust antileukemia effect mediated by cord blood CD8+ T cells has not been reported. We describe an observation of very early T-cell expansion in 4 high-risk pediatric leukemia patients receiving third-party, pooled granulocytes after T cell-replete CB transplantation (CBT). The T-cell expansion was transient but robust, including expansion of CD8+ T cells, in contrast to the delayed CD8+ T-cell expansion ordinarily observed after T cell-replete CBT. The CD8+ T cells were polyclonal, rapidly switched to memory phenotype, and had the ability to mediate cytotoxicity. This phenomenon is reproducible, and each patient remains in long-term remission without GVHD. The results suggest that fetal-derived CB CD8+ T cells can be exploited to generate robust antileukemia effects without GVHD.

Impact of the CD4:CD8 ratio in bone marrow on stem cell mobilization and engraftment in autologous stem cell transplant patients

Bone marrow (BM) damage after previous chemotherapy, such as that involving alkylating agents, and radiation therapy alone cannot explain poor hematopoietic progenitor cell mobilization. We examined the T lymphocytes of BM in 67 autologous peripheral blood stem cell transplant (auto-PBSCT) patients with non-Hodgkin lymphoma (NHL) or multiple myeloma (MM) to establish whether the cellular phenotype predicts mobilization and engraftment between January 2000 and January 2020 at the Japanese Red Cross Society Wakayama Medical Center. The total number of mobilized CD34+ cells was <2 × 10⁶ /kg in 30 patients (group A) and ≥2 × 10⁶ /kg in 37 (group B). The median absolute number of CD3+CD4+ cells was lower in group A than in group B (P = .013), and the median absolute number of CD3+CD8+ cells was higher in group A than in group B (P = .016). A low CD4:CD8 ratio was observed in all patients in group A, whereas all patients in group B showed a normal CD4:CD8 ratio (P < .001). A strong correlation was found between the CD4:CD8 ratio and median total CD34+ cells yield (r = .723, P < .001). The present results showed that a lower CD4:CD8 ratio correlated with later neutrophil and platelet engraftment (r = .662, P = .007 and r = .571, P = .008, respectively). The present results indicate that the CD4:CD8 ratio in BM contributes to the prediction of mobilization and engraftment in auto-PBSCT patients.

Community acquired respiratory virus infections in adult patients undergoing umbilical cord blood transplantation

Characteristics and risk factors (RFs) of community-acquired respiratory virus (CARV) infections after umbilical cord blood transplantation (UCBT) are lacking. We retrospectively analyzed CARV infections in 216 single-unit myeloablative UCBT recipients. One-hundred and fourteen episodes of CARV infections were diagnosed in 62 (29%) patients. Upper respiratory tract disease (URTD) occurred in 61 (54%) whereas lower respiratory tract disease (LRTD) in 53 (46%). The 5-year cumulative incidence of CARV infection was 29%. RFs for developing CARV infections were: prednisone-based graft-versus-host disease (GVHD) prophylaxis and grade II–IV acute GVHD. RFs analysis of CARV progression to LRTD identified 2007–2009 period and absolute lymphocyte count (ALC) < 0.5 × 10⁹/L. ALC < 0.5 × 10⁹/L had a negative impact on day 60 mortality in both overall CARV and those with LRTD, whereas proven LRTD was associated with higher day 60 mortality. CARV infections had a negative effect on non-relapse mortality. Overall survival at day 60 after CARV detection was significantly lower in recipients with LRTD compared with URTD (74% vs. 93%, respectively). In conclusion, CARV infections after UCBT are frequent and may have a negative effect in the outcomes, in particular in the context of lymphocytopenia.

Prospective Randomized Study Comparing Myeloablative Unrelated Umbilical Cord Blood Transplantation versus HLA-Haploidentical Related Stem Cell Transplantation for Adults with Hematologic Malignancies

In this prospective randomized study, we compared the outcomes of single-unit umbilical cord blood transplantation (UCBT) and unmanipulated haploidentical stem cell transplantation (haplo-SCT) with post-transplantation cyclophosphamide (PTCy) in adults with hematologic malignancies. All patients received a myeloablative conditioning (MAC) regimen consisting of thiotepa, busulfan, and fludarabine, with antithymocyte globulin (ATG) added for UCBT recipients. Nineteen patients were randomized to UCBT and the other 26 to haplo-HSCT. Four patients (15%) allocated to the haplo-HSCT arm lacked a suitable donor and were crossed over to the UCBT arm. Finally, 23 underwent UCBT and 22 underwent haplo-HSCT. The cumulative incidence of neutrophil recovery was 87% at a median of 19 days (range, 13 to 24 days) in the UCBT arm versus 100% at a median of 17 days (range, 13 to 25 days) in the haplo-SCT arm (P = .04). Platelet recovery was 70% at a median of 40 days (range, 18 to 129 days) in the UCBT arm versus 86% at a median of 24 days (range, 12 to 127 days) in the haplo-HCT arm (P = .02). Rates of acute graft-versus-host disease (GVHD) grade II-IV or grade III-IV, overall chronic GVHD, and extensive chronic GVHD in the UCBT and Haplo-SCT arms were 43% versus 36% (P = .8), 9% versus 9% (P = 1), 66% versus 43% (P = .04), and 41% versus 23% (P = .2), respectively. Two-year nonrelapse mortality and relapse in the 2 arms were 52% versus 23% (P = .06) and 17% versus 23% (P = .5), respectively. Two-year disease-free survival, overall survival, and GVHD/relapse-free survival in the 2 arms were 30% versus 54% (P = .2), 35% versus 59% (P = .1), and 17% versus 40% (P = .04), respectively. Our data show that in the context of an MAC regimen, haplo-SCT with PTCy provides improved outcomes compared with ATG-containing single-unit UCBT.

Partial T Cell-Depleted Peripheral Blood Stem Cell Transplantation from HLA-Identical Sibling Donors for Patients with Severe Aplastic Anemia

We analyzed the outcomes of 26 consecutive patients with acquired severe aplastic anemia (SAA) undergoing peripheral blood stem cell transplantation (PBSCT) with partial ex vivo T cell depletion with a targeted T cell dose from HLA-identical sibling donors. The median patient age was 37 years (range, 3 to 63 years). Four patients with uncontrolled pneumonia at the time of transplantation died, on days +1, +2, +21, and +26. All evaluable patients engrafted, with a median time to neutrophil recovery of 11 days (range, 10 to 14 days) and a median time to platelet recovery of 19 days (range, 8 to 53 days). Two patients had transient grade I acute graft-versus-host disease (GVHD) with skin involvement, but no patients developed grade II-IV acute GVHD. Two patients had mild skin chronic GVHD, and 1 patient had moderate chronic GVHD with ocular involvement. No relapse was observed after a median follow-up of 114 months (range, 4 to 233 months). The overall cumulative incidence of TRM at 10 years was 19%, whereas it was 5% for those with a Karnofsky Performance Status (KPS) score >60 at the time of transplantation. Disease-free survival, overall survival, and GVHD and relapse-free survival at 10 years were 81%, 81%, and 80%, respectively, for all patients and 95%, 95%, and 90%, respectively, for patients with a KPS score >60 at transplantation. Our data indicate that PBSCT with partial ex vivo T cell-depleted targeted cell dose grafts from an HLA-identical sibling donor is a feasible, safe, and effective approach to reduce GVHD and cure patients with SAA.

Invasive fungal disease in patients undergoing umbilical cord blood transplantation after myeloablative conditioning regimen

OBJECTIVE

Characteristics and risk factors (RFs) of invasive fungal disease (IFD) have been little studied in the setting of umbilical cord blood transplantation (UCBT).

METHOD

We retrospectively included 205 single-unit myeloablative UCBT recipients with a median follow-up of 64 months.

RESULTS

Fifty-six episodes of IFD were observed in 48 patients (23%) at a median time of 123 days after stem cell infusion. Invasive mold disease (IMD) occurred in 42 cases, 38 of them (90%) caused by invasive aspergillosis whereas invasive yeast disease (IYD) occurred in 14 cases, most of them due to candidemia (n = 12, 86%). The 5-year cumulative incidence of IFD, IMDs, and IYDs was 24% 19%, and 7%, respectively. In multivariate analysis, three RFs for IMDs were identified: age >30 years (HR 3.5, P = 0.017), acute grade II-IV graft-versus-host disease (HR 2.3, P = 0.011), and ≥1 previous transplant (HR 3.1, P = 0.012). The probability of IMDs was 2.5%, 14%, and 33% for recipients with none, 1, or 2-3 RFs, respectively (P < 0.001). Among IFD, IMDs had a negative effect on non-relapse mortality in multivariate analysis (HR 1.6, P = 0.039). IMDs showed a negative impact on overall survival (HR 1.59, P = 0.018).

CONCLUSION

Invasive mold disease were very common and serious complication after UCBT.

Single umbilical cord blood with or without CD34+ cells from a third-party donor in adults with leukemia

We retrospectively compared the clinical outcomes of adults with acute leukemia who received single-unit umbilical cord blood (UCB) transplantation (sUCBT) (n = 135) or stem cell transplant using coinfusion of a UCB graft with CD34⁺ cells from a third-party donor (Haplo-Cord) (n = 72) at different institutions within the Grupo Español de Trasplante Hematopoyético. In multivariable analysis, patients in the Haplo-Cord group showed more rapid neutrophil (hazard ratio [HR], 2.3; 95% confidence interval [CI], 1.5-3.3; P < .001) and platelet recovery (HR, 1.6; 95% CI, 1.2-2.3; P = .015) and lower incidence of chronic graft-versus-host disease (GVHD) (relative risk, 0.5; 95% CI, 0.3-0.8; P = .01). Nonrelapse mortality, relapse, disease-free survival (DFS), and GVHD/relapse-free survival were similar in the 2 groups. Regarding disease-specific outcomes, DFS in both acute myeloid leukemia (AML) and acute lymphoblastic leukemia patients was not significantly different; however, a significantly higher relapse rate was found in patients with AML treated with Haplo-Cord (HR, 2.3; 95% CI, 1-5.4; P = .04). Our study confirms that Haplo-Cord was an effective strategy to accelerate neutrophil and platelet recovery and shows that, in the context of specific treatment platforms, sUCBT and Haplo-Cord offer similar long-term outcomes.

GvHD prophylaxis after single-unit reduced intensity conditioning cord blood transplantation in adults with acute leukemia

To investigate better GVHD prophylaxis in reduced intensity conditioning umbilical cord blood transplantation (RIC-UCBT), we compared transplant outcomes after UCBT among GvHD prophylaxes using the registry data. We selected patients transplanted for AML or ALL with a calcineurin inhibitor and methotrexate (MTX)/mycophenolate mofetil (MMF) combination. A total of 748 first RIC-UCBT between 2000 and 2012 (MTX+ group, 446, MMF+ group, 302) were included. The cumulative incidence of neutrophil and platelet counts higher than 50 000/μL was significantly better in the MMF+ group (relative risk (RR), 1.55; P<0.001: RR, 1.34; P=0.003, respectively). In multivariate analyses, the risk of grade II-IV and III-IV acute GvHD was significantly higher in the MMF+ group than in the MTX+ group (RR, 1.75; P<0.001: RR, 1.97; P=0.004, respectively). In disease-specific analyses of AML, the risk of relapse of high-risk disease was significantly lower in the MMF+ group (RR, 0.69; P=0.009), whereas no significant difference was observed in the risk of relapse-free and overall survival in high-risk disease. In patients with standard-risk disease, no significant differences were noted in the risk of relapse or survival between the MTX+ and MMF+ groups. Collectively, these results suggest that MMF-containing prophylaxis may be preferable in RIC-UCBT, particularly for high-risk disease.

Precision in donor selection: Identifying ideal stem-cell donors through their T cells

HLA-identical siblings have always been considered ideal donors for allogeneic hematopoietic stem-cell transplantation (allo-HSCT) in the treatment of hematologic cancers. Recent data suggest that we should rethink this paradigm. In "High Graft CD8+ Cell Dose Predicts Improved Survival and Enables Better Donor Selection in Allogeneic Stem-Cell Transplantation With Reduced-Intensity Conditioning," we identified a group of stem-cell donors whose grafts contain an optimal composition of T-cells, leading to a dramatic decrease in disease relapse risk and an improvement in overall survival following allo-HSCT. To demonstrate this, we analyzed the outcomes of 200 patients with hematologic malignancies who underwent allo-HSCT after reduced-intensity conditioning (RIC). The analysis focused on T-cell content of peripheral blood stem-cell grafts. We found that higher graft CD8+ T-cell dose (CD8hi), a trait found only in grafts collected from young donors, was associated with improved survival due to a reduction in the risk for cancer relapse without a significant increase in graft-versus-host disease (GVHD). Though not all young donors mobilized CD8hi grafts, we found that a low CD4:CD8 ratio in the peripheral blood could identify these ideal donors prior to transplant. The likelihood of finding CD8hi donors correlated inversely with age, and elderly RIC transplant recipients had a low chance of receiving an ideal graft from their similarly aged siblings. Here, we examine these findings and their implications on choosing donors according to age and relatedness. We also explore biological mechanisms that determine the CD4:CD8 ratio in healthy donors.

Cord Blood Units with High CD3(+) Cell Counts Predict Early Lymphocyte Recovery After In Vivo T Cell-Depleted Single Cord Blood Transplantation

Although high absolute lymphocyte count (ALC) early after transplantation is a simple surrogate for immune reconstitution, few studies to date have established the predictive factors for ALC after umbilical cord blood transplantation (UCBT). We retrospectively studied the factors associated with early lymphocyte recovery and the impact of the ALC on day +42 (ALC42) of ≥300 × 10(6)/L on outcomes in 210 consecutive pediatric and adult patients (112 males; median age, 15 years; range, 0.3 to 60 years; interquartile range, 4 to 36 years) who underwent myeloablative in vivo T cell-depleted single UCBT between 2005 and 2014 for malignant and nonmalignant disorders. In a logistic multivariate regression model, factors favoring a higher ALC42 were higher infused CD3(+) cell dose (odds ratio [OR], 2.7; 95% CI, 1.4 to 5.2; P = .004), lower antithymocyte globulin dose (OR, 2.3; 95% CI, 1.2 to 4.5; P = .01), and better HLA match (OR, 2.1; 95% CI, 1.1 to 4.1; P = .03). In multivariate analysis, lower ALC42 was associated with higher nonrelapse mortality (hazard ratio [HR], 1.76; 95% CI, 1.34 to 2.32; P = .001), whereas a higher ALC42 was associated with better disease-free survival (HR, 2.03; 95% CI, 1.15 to 3.6; P < .001) and overall survival (HR, 2.03; 95% CI, 1.17 to 3.6; P < .001). Our study suggests that the selection of better HLA-matched cord blood units containing higher CD3(+) cell counts and the use of conditioning regimens with lower ATG doses could improve immune reconstitution after UCBT.

Comparison of Outcomes of 8/8 and 7/8 Allele-Matched Unrelated Bone Marrow Transplantation and Single-Unit Cord Blood Transplantation in Adults with Acute Leukemia

To investigate an up-to-date alternative donor selection strategy, we compared the transplantation outcomes of 8/8 and 7/8 allele-matched unrelated bone marrow transplantation (UBMT) with those of umbilical cord blood transplantation (UCBT) and redefined the role of UCBT with extended analysis. Using Cox and competing risk regression analyses, we analyzed the transplantation outcomes in adult patients with acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML). A total of 2472 first myeloablative transplantations between 2000 and 2010 were included (8/8 UBMT, 1001; 7/8 UBMT, 656; UCBT, 815). For acute and chronic graft-versus-host disease (GVHD) and nonrelapse mortality (NRM), we applied the combined analyses including both AML and ALL data. In the multivariate analyses, severe acute GVHD and NRM after UCBT were comparable with 8/8 UBMT, whereas those after 7/8 UBMT were significantly higher. The incidence of extensive chronic GVHD was significantly lower with UCBT compared with after 8/8 and 7/8 UBMT. With adjusted analyses for AML, UCBT and 8/8 UBMT showed similar overall survival (OS), whereas 7/8 UBMT showed inferior OS. For ALL, we found no significant difference in OS among the 3 groups. Cord blood may be the first choice alternative to 8/8 UBMT for both AML and ALL.

High Graft CD8 Cell Dose Predicts Improved Survival and Enables Better Donor Selection in Allogeneic Stem-Cell Transplantation With Reduced-Intensity Conditioning

PURPOSE

To characterize the impact of graft T-cell composition on outcomes of reduced-intensity conditioned (RIC) allogeneic hematopoietic stem-cell transplantation (alloHSCT) in adults with hematologic malignancies.

PATIENTS AND METHODS

We evaluated associations between graft T-cell doses and outcomes in 200 patients who underwent RIC alloHSCT with a peripheral blood stem-cell graft. We then studied 21 alloHSCT donors to identify predictors of optimal graft T-cell content.

RESULTS

Higher CD8 cell doses were associated with a lower risk for relapse (adjusted hazard ratio [aHR], 0.43; P = .009) and improved relapse-free survival (aHR, 0.50; P = .006) and overall survival (aHR, 0.57; P = .04) without a significant increase in graft-versus-host disease or nonrelapse mortality. A cutoff level of 0.72 × 10(8) CD8 cells per kilogram optimally segregated patients receiving CD8(hi) and CD8(lo) grafts with differing overall survival (P = .007). Donor age inversely correlated with graft CD8 dose. Consequently, older donors were unlikely to provide a CD8(hi) graft, whereas approximately half of younger donors provided CD8(hi) grafts. Compared with recipients of older sibling donor grafts (consistently containing CD8(lo) doses), survival was significantly better for recipients of younger unrelated donor grafts with CD8(hi) doses (P = .03), but not for recipients of younger unrelated donor CD8(lo) grafts (P = .28). In addition, graft CD8 content could be predicted by measuring the proportion of CD8 cells in a screening blood sample from stem-cell donors.

CONCLUSION

Higher graft CD8 dose, which was restricted to young donors, predicted better survival in patients undergoing RIC alloHSCT.

Bloodstream infections in adult patients undergoing cord blood transplantation from unrelated donors after myeloablative conditioning regimen

The incidence, epidemiology, and risk factors of bloodstream infection (BSI) and their impact on transplant outcomes after umbilical cord blood transplantation (UCBT) are not well defined. Between May 1997 and December 2012, 202 isolates in 189 episodes of BSI were registered in 134 of 241 patients who underwent single-unit myeloablative UCBT. Cumulative incidence (CI) of developing at least 1 episode of BSI was 21%, 29%, 34%, 42%, and 52% at days +7, +14, +30, +100, and +365, respectively. The median time of onset for the first BSI episode was day +10 (range, day -7 to +1217). Early BSI before day 7 was associated with increased nonrelapse mortality (relative risk [RR], 1.5; 95% confidence interval [CI], 1.1 to 2.3; P = .04), whereas BSI before day 14 was an independent adverse risk factor for neutrophil recovery (RR, .6; 95% CI, .5 to .9; P = .002). A higher CD8(+) cell dose of the graft was the only variable independently associated with reduced risk of BSI (RR, .1; 95% CI, .02 to .7; P = .02). The gram-negative rod (GNR) to gram-positive bacteria ratio was .9 before day +30 and 1.6 thereafter (P = .03). Escherichia coli (31%) and Pseudomonas sp. (28%) were the most frequently isolated among GNR. The overall crude mortality rate was 12% at day 7 and was higher for GNR (18%) compared with gram-positive bacteria (7%) (P = .03). These findings emphasize the importance of preventing bacterial infections during conditioning and the very early post-UCBT period.