Influence of the degree of donor bone marrow hyperplasia on patient clinical outcomes after allogeneic hematopoietic stem cell transplantation

This study evaluated the influence of the degree of donor bone marrow (BM) hyperplasia on patient clinical outcomes after allogeneic hematopoietic stem cell transplantation (allo-HSCT). Twelve patients received allo-HSCT from hypoplastic BM donors between January 2010 and December 2017. Forty-eight patients whose donors demonstrated BM hyperplasia were selected using a propensity score matching method (1:4). Primary graft failure including poor graft function and graft rejection did not occur in two groups. In BM hypoplasia and hyperplasia groups, the cumulative incidence (CI) of neutrophil engraftment at day 28 (91.7% vs. 93.8%, P=0.75), platelet engraftment at day 150 (83.3% vs. 93.8%, P=0.48), the median time to myeloid engraftment (14 days vs. 14 days, P=0.85) and platelet engraftment (14 days vs. 14 days, P=0.85) were comparable. The 3-year progression-free survival, overall survival, CI of non-relapse mortality and relapse were 67.8% vs. 71.7% (P=0.98), 69.8% vs. 77.8% (P=0.69), 18.5% vs. 13.6% (P=0.66), and 10.2% vs. 10.4% (P=0.82), respectively. In multivariate analysis, donor BM hypoplasia did not affect patient clinical outcomes after allo-HSCT. If patients have no other suitable donor, a donor with BM hypoplasia can be used for patients receiving allo-HSCT if the donor Complete Blood Count and other examinations are normal.


INTRODUCTION
In recent decades, the number of allogeneic hematopoietic stem cell transplantations (allo-HSCTs) has steadily increased by up to 10% annually worldwide (Gratwohl et al., 2010;Pasquini and Wang, 2010;Baldomero et al., 2011). Furthermore, notable improvements in haploidentical hematopoietic stem cell transplantation (haplo-HSCT) now allow this technique to be widely performed when patients with hematologic malignancies lack suitable human leukocyte antigen (HLA)-matched sibling donors (MSDs) Huang, 2012, 2014;Wang et al., 2013Wang et al., , 2015Wang et al., , 2016aWang et al., , 2016bApperley et al., 2016;Xu et al., 2018;Yu et al., 2018;Yu et al., 2019). These developments are accompanied by a parallel increase in the number of donors involved in transplantation and considerable substantial changes in the donation process. Pre-transplant screening work-up of donors for allo-HSCT is essential to minimize risks to the recipient and to protect the donor. This work-up usually includes a medical history, physical examination and investigations, which in most centers include laboratory parameters, such as complete blood count (CBC), blood chemistry, and testing for transmissible diseases (Buckner et al., 1984). Chest Xrays and electrocardiograms are routinely carried out in some centers.
Although the World Marrow Donor Association (WMDA; http://www.worldmarrow.org, accessed on May 12, 2010) has provided recommendations about unrelated donor examinations (Halter et al., 2013;Lown et al., 2014), which did not include bone marrow (BM) aspirate, there is still uncertainty regarding BM aspiration of a related donor, including MSD and haploidentical donor (HID). Some centers perform BM aspirates as part of their donor work-up process. It is reported that BM aspirates, as part of routine donor assessment for allo-HSCT, can reveal the presence of occult hematological malignancies in otherwise asymptomatic individuals (Kiss et al., 2004). In the Peking University Institute of Hematology, BM aspirates are routinely performed as part of the pre-transplant assessment on all potential donors and examined for morphology. If the morphology of donor BM displays hypoplasia, second BM aspirates are performed, and a BM biopsy is also needed. The selection of the best donor in the setting of related haplo-HSCT has been conducted according to our previous study (Wang et al., 2014). However, a number of questions remain. If the only HLA-MSD available demonstrates BM hypoplasia but nor-mal CBC and other indicators, is it necessary to replace with an alternative donor? If the best haploidentical donor demonstrates BM hypoplasia but normal CBC and other indicators, should it be replaced by a second-choice donor? If there are no other donors available, can transplantation be carried out? All these situations remain challenges. However, no data about the influence of the degree of donor BM hyperplasia on patient clinical outcomes after allo-HSCT are available. Hence, we attempted to identify the association between donor BM hypoplasia and the clinical outcomes of patients after allo-HSCT.

Patient characteristics
Study design and algorithm of patients after allo-HSCT are shown in Figure 1. A total of 60 patients were analyzed: 12 in the BM hypoplasia group and 48 in the BM hyperplasia group. The basic characteristics of the patients are shown in Table 1. The median follow-up was 30.5 (range, 6.0-88.5) months. The median age was 23 years (range, 14-57 years). Fifty patients were diagnosed with malignant disease (83.3%), and 10 were diagnosed with nonmalignant disorders (aplastic anemia, 16.7%). Patients with acute leukemia were all in remission. Twenty patients (33.3%) received grafts from MSDs, whereas 40 (66.7%) patients received grafts from HIDs. No significant differences were

Engraftment
Primary graft failure including primary poor graft function (PGF) and graft rejection did not occur in either the BM hypoplasia group or the BM hyperplasia group. Secondary PGF occurred in only one patient in the BM hyperplasia group and none in the BM hypoplasia group. Donor neutrophil engraftment was achieved in all patients. In the BM hypoplasia group and the BM hyperplasia group, there was no difference in the cumulative incidence (CI) of neutrophil engraftment at day 28 (91.7% vs. 93.8%, P=0.75; Figure  2A). The median time to neutrophil engraftment was 14 days (range, 10-22) and 14 days (range, 10-25) in the BM hypoplasia and hyperplasia groups, respectively (P=0.85). Incomplete platelet recovery before dying of nonrelapse mortality (NRM) was observed in four patients, one in the BM hypoplasia group and three in the BM hyperplasia group. The CI of platelet engraftment at day 150 was not significantly different (83.3% vs. 93.8%, P=0.48; Figure  2B). The median time to platelet engraftment was 14 days (range, 8-130) in the BM hypoplasia group and 14 days (range, 8-87) in the BM hyperplasia group (P=0.85). The use of donors with BM hyperplasia did not affect neutrophil and platelet engraftment of patients in either univariate or multivariate analyses (Table 2). A multivariate analysis of our results indicated that HID was associated with a trend toward inferior platelet engraftment (P=0.081) ( Table 2).

Graft-versus-host disease (GVHD)
The overall CI of grade 2-4 acute GVHD (aGVHD) at day 100 (25% and 22.9%, P=0.87) did not differ between the BM hypoplasia group and the BM hyperplasia group ( Figure 2C). Eleven patients in the BM hypoplasia group and 45 patients in the BM hyperplasia group survived for more than 100 days after allo-HSCT and were eligible for chronic GVHD (cGVHD) evaluation. The 3-year CIs of cGVHD (45.5% vs. 40.9%, P=0.72) were similar between the two groups ( Figure  2D). In multivariate analysis, donors with BM hypoplasia did not affect grade 2-4 aGVHD and cGVHD, while grade 2-4 aGVHD and cGVHD were associated with HID (P=0.043 and P=0.04, respectively) ( Table 2).

Relapse
One patient (8.3%) in the BM hypoplasia group relapsed at nine months after allo-HSCT and died of relapse. Six patients (12.5%) in the BM hyperplasia group showed relapse after allo-HSCT, five of whom died of relapse. The median time to relapse after allo-HSCT was 5 months (range, 4.3-31 months). The 3-year relapse rates for patients in the BM hypoplasia group and the BM hyperplasia group were 10.2% and 10.4%, respectively (P=0.82; Figure 3A). None of the variables, including BM hypoplasia, affected the incidence of relapse in the multivariate analysis (Table 2).

NRM
No patient experienced early NRM (within day +28) before engraftment. Until the last follow-up, two patients (16.7%) in the BM hypoplasia group and six patients (12.5%) in the BM hyperplasia group died of NRM at a median of 3.9 months (range, 1.7-23.7) after allo-HSCT. In the BM hypoplasia group, one patient died at 2 months after allo-HSCT from hepatic venular occlusive disease, and one patient died at 11.8 months after allo-HSCT from lung infection. In the BM hyperplasia group, three patients died at 2.3, 4.8, and 11.8 months after allo-HSCT from lung infection; two died at 1.7 and 23.7 months after allo-HSCT from GVHD, and one died at 3 months after allo-HSCT from a central nervous system infection. The CI of the 3-year NRM was comparable between the two groups (18.5% and 13.6%, P=0.66) ( Figure  3B). Multivariate analysis showed that none of the risk factors evaluated, including the degree of BM hyperplasia of the donor, predicted NRM (Table 2).

Progression-free survival (PFS) and overall survival (OS)
Fourteen patients died in the entire cohort: 8 (57.1%) died of NRM, and 6 (42.9%) died of relapse.  Figure  3C and D). The use of donors with BM hypoplasia did not affect PFS or OS in the multivariate analysis (Table 2).

Follow-up of donors with BM hypoplasia
Until the last follow-up on June 30, 2018, the median followup of donors was 36.7 months (range, 7.2-85 months). None of the donors with BM hypoplasia developed any hematological diseases. The CBC of all donors with BM hypoplasia continues to be normal. Donation of stem cells did not cause any obvious adverse reactions to these donors.

DISCUSSION
To our knowledge, this study represents the first formal report on the influence of the degree of BM hyperplasia of donors on patient clinical outcomes after allo-HSCT. Although retrospective, its strengths include proposing a new and practical clinical problem. Furthermore, a propensity score matching method in R was used. When the only HLAmatched sibling donors or the best haploidentical donor demonstrated BM hypoplasia with normal CBC and other indicators, it remains a challenge to know whether an alternative donor or a second-choice donor should be used. Our results demonstrate that a donor with BM hypoplasia did not affect the outcome of the patient, provided that CBC and other indicators were normal, and it was not a contraindication for transplantation. Future work should follow up and verify the current conclusion by extending the follow-up time and increase the number of cases.
The proportion of donors with BM hypoplasia in our study was 0.7% (28/4203). This study represents the first formal report on the proportion of donors with BM hypoplasia. Donors with BM hypoplasia are a rare but serious problem for allo-HSCT, especially when patients have no other sui- A multivariate analysis of our results indicated that HID was associated with a trend towards inferior platelet engraftment (P=0.081). This may be related to the higher incidence of prolonged isolated thrombocytopenia in HID allo-HSCT than in MSD allo-HSCT (Chang and Huang, 2012;Wang et al., 2014;Zhang et al., 2015;Xu et al., 2016;Fu et al., 2018). The CI of grade 2-4 aGVHD and cGVHD in the hypoplasia group was not significantly different from that of the hyperplasia group. HID was associated with higher grade 2-4 aGVHD and cGVHD, which is consistent with our recently updated results (Wang et al., 2013(Wang et al., , 2016a(Wang et al., , 2016bXu et al., 2016). Nevertheless, the rates of GVHD in the HID group were quite acceptable in our previous study, partly because of the relatively low incidence in the MSD cohort. In our report, the rate of NRM and relapse in the hypoplasia group was also not significantly higher than the rate in the hyperplasia group, which was similar to that reported in our previous studies (Wang et al., 2013(Wang et al., , 2016a(Wang et al., , 2016bXu et al., 2016). In addition, the patient survival of the two groups was comparable and similar to that identified in our previous study but higher than that observed in other previous reports (Aversa et al., 2005Ciceri et al., 2008) because of a more favorable relapse rate and a much lower NRM. However, no factor was discovered in the risk analysis of NRM, relapse, PFS and OS, and this lack of finding might be related to the small sample size of the study.
Although the WMDA recommends minimum standards for the assessment of unrelated donor medical suitability, it does not include BM aspirates (WMDA; http://www. worldmarrow.org, accessed on May 12, 2010; Boo et al., 2011;Halter et al., 2013;Lown et al., 2014). For related donor (including MSD and HID) allo-HSCT, the necessity of donor BM aspiration is also uncertainty. In general, little information is available as to whether it is necessary for donors to undergo BM aspirates. It was reported that three transplants had to be canceled because of the presence of occult hematological malignancies through BM aspirates that were undetectable by routine blood testing in potential donors (Kiss et al., 2004). This report suggested that it is reasonable for older donors and donors with potential familial malignancies to undergo more intense screening investigations, including BM aspiration, to identify occult hematological malignancies prior to stem cell donation. Since our results show that the use of donors with hypoplastic BM does not affect the clinical outcomes of patients after allo-HSCT, the next question is whether the donor needs to perform BM aspirates as long as his or her CBC and other examinations are normal. We note that two of the three donors with occult hematological malignancies detected through BM aspiration in the report by Kiss et al. (2004) were older than 60 years. However, the age of the donors in our present study was younger than the age in their report, with only one of 12 donors older than 60 years. We recommend that young donors do not have to undergo BM aspiration as long as CBC and other indicators are normal and there is no history of familial malignancies or inherited hematological diseases. However, whether older donors (≥60 years) need BM aspiration requires further investigation.
Our study has several limitations. First, this is a retrospective and single-center study. Therefore, certain biases exist, although we made adjustments in the multivariate analysis to try to account for these. Second, the relatively small number of patients in different subgroups limited the power of the analysis. Third, insufficient data on immune reconstitution limited our ability to incorporate these important variables into the analysis.
In summary, our study is the first report on the influence of the degree of donor BM hyperplasia on patient clinical outcomes after allo-HSCT. Our results demonstrate that donors with BM hypoplasia but with normal CBC and other indicators did not affect the clinical outcomes of patients after allo-HSCT and was not a contraindication for transplantation. In this case, if patients with no other suitable donor, a donor with BM hypoplasia can be used for patients receiving allo-HSCT. Although retrospective and with limited sample size, we have addressed this clinical problem and present the preliminary observations and discussion. Further multicenter and prospective studies are warranted to confirm this conclusion.

Patients
As shown in Figure 1 ,4,203 donors,including HIDs and MSDs,including BM aspirate,at Peking University Institute of Hematology between January 2010 and December 2017. We reviewed the data of all patients (n=117) whose donors demonstrated morphologic hypoplasia on the first BM aspirates. All 117 donors received second BM aspirates with morphology and BM biopsy. We identified 28 donors and confirmed BM hypoplasia through two BM morphology tests and one BM biopsy; the donor's CBC and other examinations were normal, and the donor had no history of familial malignancies or inherited hematological diseases. The proportion of donors with BM hypoplasia was 0.7% (28/ 4,203). Sixteen patients changed donors since they had other suitable donors. Finally, 12 patients had to receive allo-HSCT using donors with hypoplasia BM because there was no other suitable donor. For each BM hypoplasia case, a set of control cases (1:4) was selected using propensity score matching method in R from subjects whose donors had confirmed BM hyperplasia and received allo-HSCT according to the following criteria: (1) diagnosis, (2) age (±5 years), and (3) year of transplantation (±2 years). Finally, 48 patients were selected as the matched control group. The final study cohort comprised 60 patients. Informed consent was obtained from all donors and recipients. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the ethics committee of Peking University People's Hospital.
All patients in the HID group and the majority of patients in the ISD group received granulocyte colony-stimulating factor (G-CSF)-mobilized, fresh, and unmanipulated BM cells plus G-CSF-primed peripheral blood (PB) stem cells. A minority of patients in the ISD group received only G-CSFmobilized PB stem cells because the donors declined to donate BM. Additionally, all patients received cyclosporine, mycophenolate mofetil, and short-term methotrexate for GVHD prophylaxis (Wang et al., 2013(Wang et al., , 2016a(Wang et al., , 2016b.

Definitions and evaluation
BM morphologic hypoplasia was defined as the degree of hyperplasia grade IV and V. BM biopsy hypoplasia was defined as the hematopoietic volume of BM less than 40%. Engraftment, graft failure, NRM, relapse, PFS, and OS were defined as previously described (Wang et al., 2013(Wang et al., , 2016a(Wang et al., , 2016b. Definition and grading of aGVHD was according to the modified Keystone criteria (Przepiorka et al., 1995) and cGVHD was defined and graded according to the National Institutes of Health criteria (Filipovich et al., 2005). A patient experiencing relapse was defined on the basis of histologic criteria.

End points
The primary study end point was the incidence of engraftment. Secondary end points were the CIs of grades 2-4 aGVHD, cGVHD, NRM, relapse, PFS and OS.

Statistical analysis
The data were up-to-date as of June 30, 2018, and represent a minimum of 6 months of follow-up. Propensity score matching method in R was guided step by step (Lanza et al., 2013). Comparisons of patient characteristics between the two groups were performed using the Mann-Whitney U test for continuous variables and the Chi-square test for categorical data. The CIs of relapse, NRM, engraftment, and GVHD were estimated using competing risks. Relapse and NRM were competing risks for each other. Death was a competing risk for engraftment, aGVHD and cGVHD. PFS and OS were estimated using the Kaplan-Meier estimator. Multivariate analysis for transplantation outcomes was performed using Cox regression models. The variables tested included the degree of BM hyperplasia of the donor (BM hypoplasia vs. BM hyperplasia), patient age (≤18 years vs. >18 years), diagnosis (aplastic anemia versus malignant disease), donor-recipient sex match (female-male vs. other), donor-recipient relation (mother donor vs. other), donor source (HID vs. MSD), conditioning regimen (chemotherapy based vs. TBI based) and graft source (BM+PB vs. PB). The Cox models always retained the degree of BM hyperplasia of the donor regardless of the P value, even if its coefficients were not significant. Other covariates were selected by backward elimination and were retained in the Cox models if the P value was <0.1 in the univariate analysis. A P<0.05 was considered statistically significant. All P values were obtained using two-sided hypothesis tests. The data analyses were conducted primarily using Statistical Package for the Social Sciences (SPSS) software (SPSS Inc., Chicago, IL, USA), whereas the competing risk analysis was performed using R software (version 3.5.2; https://www.r-project.org/).

Compliance and ethics
The author(s) declare that they have no conflict of interest.