Pre-engraftment syndrome after unrelated donor umbilical cord blood transplantation in patients with hematologic malignancies

Pre-engraftment syndrome (PES) after umbilical cord blood transplantation (CBT) remains poorly characterized, and the prognosis and appropriate management are unclear. Therefore, we retrospectively analyzed the incidence, risk factors, manifestations, and clinical outcomes of PES in CBT recipients, who had been treated for hematologic malignancies at our transplantation center. PES was defined as unexplained fever higher than 38.3°C that is not associated with documented infection and unresponsive to antimicrobial manipulations and/or unexplained erythematous skin rash occurring prior to neutrophil engraftment. A total of 81 patients (median 18 yr, range 3-48) received either myeloablative (n=72) or non-myeloablative (n=9) conditioning. Neutrophil engraftment was achieved in 69 of the 81 cases [86.2%, 95% confidence interval (CI)=78.9-94.1%], and the median time to more than 0.5 × 10(9) /L ANC was 19 d (range, 12-39). Fifty-one patients (63.0%) developed PES at a median of 7d (range 3-15) post-transplant: 46 patients had both rash and unexplained fever; one patient had unexplained fever alone; and four patients had rash only. Forty-seven patients (92.2%) received IV methylprednisolone (MP) at a median dose of 1 mg/kg (range 0.4-3). All patients treated with MP responded as evidenced by fever resolution combined with resolution of rash. All patients with PES had high serum levels of C-reactive protein (CRP), which were significantly reduced after effective steroid treatment. Univariate analysis identified myeloablative conditioning and younger age as significant risk factors for developing PES. Cumulative incidence of grade II-IV acute graft-versus-host disease (aGVHD) in the PES+ and PES- groups was 51.5% (95% CI=38.0-70.0%) and 17.0% (95% CI=6.9-41.7%), respectively. In a multivariate analysis, we found significantly increased risk of grade II-IV aGVHD among PES patients (P=0.041). However, PES was not associated with sustained donor engraftment, the day to neutrophil recovery, chronic graft-versus-host disease, transplant-related mortality at day 180, and overall survival. Despite of the inherent limitations of this small retrospective study, PES seemed to be common after CBT and associated with high incidence of aGVHD.     

Impact of a single human leucocyte antigen (HLA) allele mismatch on the outcome of unrelated bone marrow transplantation over two time periods. A retrospective analysis of 3003 patients from the HLA Working Group of the Japan Society for Blood and Marrow Transplantation

A previous Japanese study revealed that a human leucocyte antigen (HLA)‐A or ‐B allele mismatch was associated with higher overall mortality, whereas an HLA‐C or ‐DRB1 allele mismatch did not affect mortality after serologically matched unrelated bone marrow transplantation (BMT). This study reanalysed 3003 adult patients who underwent unrelated BMT from a serologically HLA‐A, ‐B, or ‐DR matched unrelated donor between 1993 and 2009 using the latest database, that included 1966 HLA‐matched unrelated BMT and 187, 31, 524, and 295 unrelated BMT with a single HLA‐A, ‐B, ‐C, or ‐DRB1 allele mismatch, respectively. As opposed to our previous findings, HLA‐C and ‐DRB1 mismatches had a significant negative impact [hazard ratio (HR) 1·35, P < 0·001, and HR 1·45, P < 0·001] on survival in the period 2000–2009. The negative impact of each single HLA allele mismatch was not significantly different among the HLA‐A, ‐B, ‐C, and ‐DRB1 mismatches (P = 0·79). An interaction test revealed that the effects of single HLA‐C and ‐DRB1 allele mismatches significantly differed over the two time periods (P = 0·032 and P = 0·0072, respectively). In conclusion, the impact of a single HLA allele mismatch changed over time. In the recent cohort, the negative impact of HLA‐DRB1 and ‐C mismatches became apparent.     

Impact of hematopoietic chimerism at day +14 on engraftment after unrelated donor umbilical cord blood transplantation for hematologic malignancies

Background

Cord blood transplant is a feasible treatment alternative for adult patients with hematologic malignancies lacking a suitable HLA-matched donor. However, the kinetics of myeloid recovery is slow, and primary graft failure cannot be detected easily early after transplantation. We investigated the impact of hematopoietic chimerism status from unselected marrow cells 14 days after transplantation on predicting engraftment after a cord blood transplant.

Design and Methods

Seventy-one adult patients with hematologic malignancies undergoing single-unit unrelated donor cord blood transplantation after a myeloablative conditioning regimen were included in the study. All patients received conditioning regimens based on busulfan, thiotepa and antithymocyte globulin. Chimerism status was assessed analyzing short tandem repeat polymorphisms.

Results

The cumulative incidence of myeloid engraftment at 1 month was significantly lower in patients with mixed chimerism than in those with complete donor chimerism (55% vs. 94%; p<0.0001). For patients achieving myeloid recovery, the median time of engraftment was 16 days when donor chimerism at day + 14 was higher than 90%, compared with 24 days when donor chimerism was below this level (p<0.001). A donor chimerism level of 65% was found to be the best cut-off point for predicting primary graft failure, with a sensitivity of 97% and a specificity of 80%. The incidence of primary graft failure was 67% for patients with less than 65% donor chimerism at day +14 as compared to only 2% for those with more than 65% donor chimerism (p<0.001). Patients with mixed chimerism also had a lower cumulative incidence of platelet engraftment than those with complete chimerism (62% vs. 89%; p=0.01).

Conclusions

Donor-recipient chimerism status at day +14 predicts engraftment after a single-unit cord blood transplant in adults.     

Rapid T‐cell chimerism switch and memory T‐cell expansion are associated with pre‐engraftment immune reaction early after cord blood transplantation

Until recently, both the British Society for Haematology and American College of Chest Physicians recommended platelet monitoring in all surgical patients receiving prophylactic low molecular weight heparin (LMWH) for the early diagnosis of heparin‐induced thrombocytopenia (HIT). These guidelines were reversed in 2012 based upon an analysis considering resource expenditure, assay result timeframes, and complications relating to HIT treatment. However, there are no large studies reviewing lower limb arthroplasty patients on an individual basis to determine the incidence of HIT in this patient group. This study investigated 10 797 patients who underwent primary hip or knee arthroplasty with LMWH prophylaxis over a 5 years period. 32·6% of patients (n = 3515) had platelet counts recorded up to 14 d postoperatively with 13 patients (0·37%) developing thrombocytopenia. Platelet counts recovered spontaneously in five patients, and two patients had other identifiable causes. Only one of the remaining six patients developed thrombosis indicating an incidence of HIT‐related thrombosis of 0·03%. The potential for identifying HIT with platelet monitoring in patients receiving LMWH prophylaxis is low and therefore routine monitoring for HIT is not justified.     

Occurrence of graft‐versus‐host disease increases mortality after umbilical cord blood transplantation for acute myeloid leukaemia: a report from Eurocord and the ALWP of the EBMT

Background

The efficacy of umbilical cord blood transplantation (UCBT) as treatment for acute myeloid leukaemia (AML) relies on immune‐mediated graft‐versus‐leukaemia effects. Previous studies have suggested a strong association between graft‐versus‐host disease (GVHD) occurrence and graft‐versus‐leukaemia effects after allogeneic hematopoietic cell transplantation.

Methods

Here, we evaluated the kinetics of relapse rate in correlation with GVHD occurrence after UCBT. The kinetics of relapse rate over time in correlation to GVHD occurrence were assessed by calculating the relapse rate per patient‐year within sequential 90‐day intervals. The impact of GVHD on relapse and mortality was further studied in multivariate Cox models handling GVHD as a time‐dependent covariate.

Results

The study included data from 1068 patients given single (n  = 567) or double (n  = 501) UCBT. The proportion of patients with grade II, III and IV acute GVHD was 20%, 7% and 4%, respectively. At 2 years, the cumulative incidence of chronic GVHD was 42%, the cumulative incidence of relapse was 32%, and overall survival was 32% as well. Relapse rates declined gradually over time during the first 30 months after transplantation. There was a possible suggestion that grade II–IV acute (HR = 0.8, P  = 0.1) and chronic (HR = 0.65, P  = 0.1) GVHD decreased relapse risk. However, grade II–IV acute GVHD significantly increased early (the first 18 months after UCBT) mortality (HR = 1.3, P  = 0.02), whilst chronic GVHD increased each early (HR = 2.7, P  < 0.001) and late (HR = 4.9, P  < 0.001) mortality after UCBT.

Conclusions

The occurrence of grade II–IV acute or chronic GVHD each increases overall mortality after UCBT for AML mitigating the possible graft‐versus‐leukemia effect of GVHD.

     

Influence of human leucocyte antigen disparity and graft lymphocytes on allogeneic engraftment and survival after umbilical cord blood transplant in adults

The dose of graft-nucleated cells and CD34(+) haematopoietic progenitor cells are predictors of allogeneic engraftment and survival in umbilical cord blood (UCB) recipients. In this single institution prospective phase II trial, flow cytometric analyses of CD34(+) progenitor and lymphocyte populations in unmodified single unit human leucocyte antigen (HLA)-disparate UCB grafts infused into 31 consecutive adults (median age 41 years, range 20-64) receiving myeloablative conditioning were compared with clinical outcomes. Median infused UCB graft-nucleated cells and CD34(+) dose was 2.2 x 10(7)/kg and 1.2 x 10(5)/kg respectively. Day to absolute neutrophil count >/=0.5 x 10(9)/l with full donor chimerism averaged 27 d (range 12-41). Univariate analyses demonstrated that UCB graft-infused cell doses of CD34(+) (P = 0.015), CD3(+) (P = 0.024) and CD34(+)HLADR(+)CD38(+) progenitors (P = 0.043) correlated with neutrophil engraftment. This same analysis did not demonstrate a correlation between CD34(+) (P = 0.11), CD3(+) (P = 0.28) or CD34(+)HLADR(+)CD38(+) (P = 0.108) cell dose and event-free survival (EFS). High-resolution matching for HLA-class II (DRB1) resulted in improved EFS (P = 0.02) and decreased risk for acute graft-versus-host disease (GVHD) (P = 0.004). Early mortality (prior to post-transplant day +28) occurred in three patients, while 26 patients achieved myeloid engraftment. These results suggest that UCB graft matching at DRB1 is an important risk factor for acute GVHD and survival, while higher UCB graft cell doses of CD34(+), committed CD34(+) progenitors and CD3(+) T cells favourably influence UCB allogeneic engraftment.     

Risk factors affecting outcome of unrelated cord blood transplantation for children with familial haemophagocytic lymphohistiocytosis

Allogeneic haematopoietic stem cell transplantation is still the only available curative option for Familial Haemophagocytic Lymphohistiocytosis (FHLH). Most studies report outcomes after bone marrow or peripheral blood stem cell transplantation. We analysed the outcomes of 118 children with FHLH undergoing single-unit umbilical cord blood transplantation performed from 1996 to 2014. Myeloablative conditioning regimen was given to 90% of the patients, and was mostly busulfan-based (n = 81, 76%), including anti-thymocyte globulin or alemtuzumab (n = 102, 86%). The cumulative incidence of Day 60 neutrophil engraftment was 85%; and that of non-relapse mortality and acute graft-versus-host disease (GvHD) was 21% and 33% at 100 days, respectively. The 6-year cumulative incidence of chronic GvHD was 17% and the 6-year probability of overall survival was 55%. In multivariate analysis, children receiving a graft with a total nucleated cell dose greater than 9·9 × 10⁷/kg had a better overall survival (hazard ratio [HR]: 0·49, 95% CI: 0·27–0·88, P = 0·02). Degree of human leucocyte antigen (HLA) matching was associated with improved disease-free survival (5/6 vs. 6/6 HR: 2·11, 95% confidence interval [CI]: 1·01–4·4, P = 0·05 and ≤4/6 vs. 6/6, HR: 2·82, CI: 1·27–6·23, P = 0·01). Umbilical cord blood transplantation with a high cell dose and good HLA match is a suitable alternative option to haematopoietic stem cell transplantation in children with FHLH who lack a HLA-matched donor.     

Toxoplasmosis in cord blood transplantation recipients

Toxoplasmosis is a devastating opportunistic infection that can affect immunocompromised patients such as cord blood transplantation (CBT) recipients. The clinical characteristics of 4 toxoplasmosis CBT patients treated at our institution are reviewed, together with 5 cases collected from the literature. The rate of toxoplasmosis in our hospital was 6% in CBT recipients and 0.2% in other types of allogeneic hematopoietic stem cell transplantation (P < 0.001). Five patients (56%) presented disseminated toxoplasmosis and 4 patients (44%) had localized infection in the central nervous system. In 5 of the 9 patients considered (56%), cytomegalovirus viral replication had been detected before the clinical onset of toxoplasmosis. Seven patients (78%) had previously developed graft‐versus‐host disease. All patients who exhibited disseminated disease died due to Toxoplasma infection. Pre‐transplant serology was positive in 1 patient, negative in 3 patients, and not performed in another. Only 1 of these 5 patients with disseminated disease had received Toxoplasma prophylaxis with cotrimoxazole. It could be concluded that mortality in CBT patients with disseminated toxoplasmosis is unacceptably high. The negative results of serology in the majority of these cases, and its unspecific clinical presentation, makes diagnosis exceedingly difficult. Better diagnostic tests and prophylaxis strategy are needed in CBT recipients.     

A comparison of immune reconstitution and graft-versus-host disease following myeloablative conditioning versus reduced toxicity conditioning and umbilical cord blood transplantation in paediatric recipients

Immune reconstitution appears to be delayed following myeloablative conditioning (MAC) and umbilical cord blood transplantation (UCBT) in paediatric recipients. Although reduced toxicity conditioning (RTC) versus MAC prior to allogeneic stem cell transplantation is associated with decreased transplant-related mortality, the effects of RTC versus MAC prior to UCBT on immune reconstitution and risk of graft-versus-host disease (GVHD) are unknown. In 88 consecutive paediatric recipients of UCBT, we assessed immune cell recovery and immunoglobulin reconstitution at days +100, 180 and 365 and analysed risk factors associated with acute and chronic GVHD. Immune cell subset recovery, immunoglobulin reconstitution, and the incidence of opportunistic infections did not differ significantly between MAC versus RTC groups. In a Cox model, MAC versus RTC recipients had significantly higher risk of grade II-IV acute GVHD [Hazard Ratio (HR) 6·1, P = 0·002] as did recipients of 4/6 vs. 5-6/6 HLA-matched UCBT (HR 3·1, P = 0·03), who also had significantly increased risk of chronic GVHD (HR 18·5, P = 0·04). In multivariate analyses, MAC versus RTC was furthermore associated with significantly increased transplant-related (Odds Ratio 26·8, P = 0·008) and overall mortality (HR = 4·1, P = 0·0001). The use of adoptive cellular immunotherapy to accelerate immune reconstitution and prevent and treat opportunistic infections and malignant relapse following UCBT warrants further investigation.     

Positive impact of chronic graft‐versus‐host disease on the outcome of patients with de novo myelodysplastic syndrome after allogeneic hematopoietic cell transplantation: a single‐center analysis of 115 patients

To evaluate the impact of graft‐versus‐host disease (GVHD) and prognostic factors for patients with myelodysplastic syndrome (MDS) after allogeneic hematopoietic cell transplantation (allo‐HCT), we retrospectively reviewed 115 patients with MDS or acute myeloid leukemia with multilineage dysplasia (AML‐MLD) after allo‐HCT at our center. Eighty one patients received reduced‐intensity conditioning (RIC) regimens, whereas 34 received myeloablative conditioning regimens. Although the RIC group was significantly older and included more patients with poor cytogenetic risk, no difference in 4‐yr overall survival (OS) was seen between the two groups. In a multivariate analysis, covariates associated with a worse OS were the French‐American‐British stage of refractory anemia excess blasts in transformation/AML‐MLD at peak, poor cytogenetic risk, bone marrow blasts of 20% or higher at HCT and the absence of chronic GVHD (cGVHD). By using semi‐landmark analyses, we found that the presence of cGVHD significantly improved OS in high‐risk patients or the RIC group. However, there was no difference in OS between those with and without cGVHD among low‐risk MDS patients. These findings suggest that the graft‐versus‐leukemia effect may be more beneficial in high‐risk patients who do not receive intensive preparative regimens.     

Predominant or complete recipient T-cell chimerism following alemtuzumab-based allogeneic transplantation is reversed by donor lymphocytes and not associated with graft failure

The clinical significance of mixed chimerism following allogeneic haematopoietic stem cell transplantation (HSCT) remains controversial. Its relevance and incidence are probably influenced by the conditioning regimen and incorporation of T‐cell depletion. The presence of recipient chimerism levels >40–50% following T‐cell replete reduced intensity transplantation correlates with a high risk of graft rejection, regardless of donor‐lymphocyte infusions, but it is unclear whether this finding translates to T‐cell depleted transplants. We conducted a retrospective single‐institution analysis of patients receiving alemtuzumab‐based HSCT. 27/152 (18%) evaluable cases had predominantly recipient T‐cell chimerism at 3 months or beyond. By contrast, coincident chimerism in the granulocyte lineage was predominantly of donor origin (median 100%) in all but one patient. Donor lymphocyte infusion effectively converted predominantly recipient T‐cell chimerism to ful donor chimerism in all evaluable cases including three cases with no detectable donor T cells. The only graft failure occurred in the patient with predominantly recipient myeloid chimerism in whom rejection occurred rapidly before donor lymphocytes could be administered. We conclude that predominant or complete recipient T‐cell chimerism following alemtuzumab‐based regimens does not have the same clinical implications as that following T‐cell replete transplants and can be effectively converted with donor lymphocytes without the need for lympho‐depleting agents or re‐conditioning.     

New ways to separate Graft‐versus‐Host Disease and Graft‐versus‐Tumour effects after allogeneic haematopoietic stem cell transplantation

A major challenge to transplant immunologists and physicians remains the separation of harmful graft‐versus‐host disease (GvHD) and beneficial graft‐versus‐tumour (GvT) effects after allogeneic haematopoietic stem cell transplantation. Recent advances in our understanding of the allogeneic immune response provide potential new opportunities to achieve this goal. Three potential new approaches that capitalize on this new knowledge are considered in depth; the manipulation of organ‐specific cytokines and other pro‐inflammatory signals, the selective manipulation of donor effector T cell migration, and the development of cell‐mediated immunosuppressive strategies using donor‐derived regulatory T cells. These new approaches could provide strategies for local control of allogeneic immune responses, a new paradigm to separate GvHD and GvT effects. Although these strategies are currently in their infancy and have challenges to successful translation to clinical practice, all have exciting potential for the future.     

Novel adopted immunotherapy for mixed chimerism after unrelated cord blood transplantation in Omenn syndrome

Omenn syndrome is a variant form of severe combined immunodeficiency. It is fatal unless treated by allogeneic stem cell transplantation (SCT), which is the only curative approach. However, both treatment-related complications and graft rejection are major obstacles to treatment success. This report describes a case with Omenn syndrome who developed mixed chimerism after unrelated cord blood transplantation (UCBT). This case was successfully treated by altering the patient's immunosuppression and donor lymphocyte infusion (DLI) with donor cord blood-derived activated CD4+ T cells ex vivo expanded from the cord blood cell residues in an infused bag. This novel development to expand CD4+ T-lymphocytes from the donor cord blood unit for the use of DLI would serve as a useful method to overcome the risk of graft rejection in SCT for primary immunodeficiency disorders with residual cell-mediated immunity without compounding graft-vs.-host disease, especially in the UCBT setting.     

Post‐transplant bendamustine reduces GvHD while preserving GvL in experimental haploidentical bone marrow transplantation

Advances in haploidentical bone marrow transplantation (h‐BMT) have drastically broadened the treatment options for patients requiring BMT. The possibility of significantly reducing the complications resulting from graft‐versus‐host disease (GvHD) with the administration of post‐transplant cyclophosphamide (PT‐CY) has substantially improved the efficacy and applicability of T cell‐replete h‐BMT. However, higher frequency of disease recurrence remains a major challenge in h‐BMT with PT‐CY. There is a critical need to identify novel strategies to prevent GvHD while sparing the graft‐versus‐leukaemia (GvL) effect in h‐BMT. To this end, we evaluated the impact of bendamustine (BEN), given post‐transplant, on GvHD and GvL using clinically relevant murine h‐BMT models. We provide results indicating that post‐transplant bendamustine (PT‐BEN) alleviates GvHD, significantly improving survival, while preserving engraftment and GvL effects. We further document that PT‐BEN can mitigate GvHD even in the absence of Treg. Our results also indicate that PT‐BEN is less myelosuppressive than PT‐CY, significantly increasing the number and proportion of CD11b⁺Gr‐1^hi cells, while decreasing lymphoid cells. In vitro we observed that BEN enhances the suppressive function of myeloid‐derived suppressor cells (MDSCs) while impairing the proliferation of T‐ and B‐cells. These results advocate for the consideration of PT‐BEN as a new therapeutic platform for clinical implementation in h‐BMT.     

Lymphocyte subset reconstitution after unrelated cord blood or bone marrow transplantation in children

We report the post‐transplant lymphocyte subset recovery of 226 children treated with Unrelated Cord Blood transplant (UCBT) (n = 112) or Unrelated Bone Marrow Transplant (UBMT) (n = 114) for malignant or non‐malignant diseases. Absolute numbers of natural killer (NK), B and T cells were monitored by flow cytometry up to 5 years post‐transplant. Immunological endpoints were: time to achieve a CD3⁺ cell count >0·5 and 1·5 × 10⁹/l, CD4⁺ > 0·2 and 0·5 × 10⁹/l, CD8⁺ > 0·25 × 10⁹/l, CD19⁺ > 0·2 × 10⁹/l, NK > 0·1 × 10⁹/l. These endpoints were analysed through the use of cumulative incidence curves in the context of competing risks. CD8⁺ T cell recovery was delayed after UCBT with a median time to reach CD8⁺ T cells > 0·25 × 10⁹/l of 7·7 months whereas it was 2·8 months in UBMT (P < 0·001). B cell recovery was better in UCBT, with a median time to reach CD19⁺ cells > 0·2 × 10⁹/l of 3·2 months in UCBT and 6·4 months in UBMT (P = 0·03). Median time for CD4⁺ T cell and NK cell recovery was similar in UCBT and UBMT. CD4⁺ T cells recovery was negatively correlated to age (better reconstitution in younger patients, P = 0·002). CD8⁺ T cells recovery was shorter in recipients with a positive cytomegalovirus serology (P = 0·001).     

Follow‐up of post‐transplant minimal residual disease and chimerism in childhood lymphoblastic leukaemia: 90 d to react

Relapse after transplantation is a major cause of treatment failure in paediatric acute lymphoblastic leukaemia (ALL). Here, we report the findings of a prospective national study designed to investigate the feasibility of immune intervention in children in first or subsequent remission following myeloablative conditioning. This study included 133 children who received a transplant for ALL between 2005 and 2008. Minimal Residual Disease (MRD) based on T cell receptor/immunoglobulin gene rearrangements was measured on days ?30, 30, 90 and 150 post‐transplantation. Ciclosporin treatment was rapidly discontinued and donor lymphocyte infusions (DLI) were programmed for patients with a pre‐ or post‐transplant MRD status ≥10^?3. Only nine patients received DLI. Pre‐ and post‐transplant MRD status, and the duration of ciclosporin were independently associated with 5‐year overall survival (OS), which was 62·07% for the whole cohort. OS was substantially higher in patients cleared of MRD than in those with persistent MRD (52·3% vs. 14·3%, respectively). Only pre‐transplant MRD status (Hazard Ratio 2·57, P = 0·04) and duration of ciclosporin treatment (P < 0·001) were independently associated with relapse. The kinetics of chimerism were not useful for predicting relapse, whereas MRD monitoring up to 90 d post‐transplantation was a valuable prognostic tool to guide therapeutic intervention.     

Pre‐transplant comorbidity burden and post‐transplant chronic graft‐versus‐host disease

The Haematopoietic Cell Transplantation‐Comorbidity Index (HCT‐CI) was designed as a predictor of non‐relapse mortality after HCT. Chronic graft‐versus‐host disease (GVHD) contributes to mortality after HCT. Here, we investigated whether the HCT‐CI could predict development of chronic GVHD or post‐chronic GVHD mortality. We retrospectively analysed data from 2909 patients treated with allogeneic HCT for malignant and non‐malignant haematological conditions at four institutions. In Cox regression models adjusted for potential confounders, increasing HCT‐CI was not statistically significantly associated with the development of chronic GVHD [hazard ratio (HR) = 1·02, P = 0·34]. Yet, the index was associated with an increased risk of non‐relapse mortality (HR = 1·29, P < 0·0001) as well as overall mortality (HR = 1·25, P < 0·001) following the development of chronic GVHD. The association between HCT‐CI and post‐chronic GVHD mortality was similar regardless of donor type or stem cell source. HCT‐CI scores could be incorporated in the design of clinical trials for treatment of chronic GVHD.