Risk Factors for Poor Autologous Peripheral blood Stem Cell Mobilization among Lymphoproliferative Disease Patients

Original article: Risk Factors for Poor Autologous Peripheral blood Stem Cell Mobilization among Lymphoproliferative Disease Patients Mohd Nazri Hassan1, Azlan Husin2, Rapiaah Mustaffa3, Rosline Hassan4, Mohd Ismail Ibrahim5, Abu Dzarr Abdullah6, Noor Haslina Mohd Noor7 Abstract: Objective: Autologous peripheral blood haemopoietic stem cell (PBSC) transplantation is a standard therapeutic option for eligible patients with lymphoproliferative disease (LPD). The prerequisite for autologous PBSC transplantation is the successful stem cell mobilization. This study is aimed to determine the factors associated with poor PBSC mobilization in LPD patient at our center. Materials and methods: This retrospective record review involved 39 multiple myeloma (MM) and 92 of lymphoma patients who had undergone PBSC mobilization from January 2009 until December 2016. Patients were mobilized with combination chemotherapy and granulocyte colony stimulating factor. Factors affecting mobilization including patient’s, disease and treatment characteristics werestudied. Results: Majority of patients were Malay (93.9%) with the mean age at mobilization of 41.4 years. The mean of CD34+ cell dosage was 9.6x106 cells/kg. Successful and poor mobilization was found to be 90.8% and 9.2% respectively. Multivariate analysis showed that the significant risk factors for poor mobilization were age of ≥ 60 years (adjusted OR=38.43, p=0.005) and PB CD34+ cell count, <20 cells/uL (adjusted OR=132.69, p<0.001). Conclusion:PB CD34+ cell count and age ≥ 60 years were the main risk factors for poor PBSC mobilization. Thus, alternative strategies of mobilization is needed to reduce risk of poor mobilization in a such group of patient.


Introduction:
Salvage chemotherapy followed by autologous peripheral blood haematopoietic stem transplantation (APBSCT) is globally accepted as the standard of care for lymphoproliferative disorder (LPD) namely multiple myeloma (MM), non-Hodgkin lymphoma (NHL) and relapsed Hodgkin lymphoma (HL) patients 1,2

. Hospital Universiti Sains Malaysia (USM)
is a referral transplantation center for APBSCT treatment of MM and lymphoma patients for the East Coast states of Malaysia since 2009. Peripheral blood stem cell (PBSC) is now the preferred source of haematopoietic stem cell (HSC) replacing bone marrow because of relative ease of collection, avoidance of hospital admission or general anaesthesia, faster haematopoietic engraftment and lesser transplant related morbidity 3 . Adequate stem cell mobilization is essential for a successful APBSCT. Poor mobilization leads to a stem cell collection failure and this result in increase resource utilization in terms of increase use of growth factors, mobilization reattempts using other mobilzer agent, hospitalizations, transfusions requirement, and antibiotics for neutropenic fever 4,5 .
A minimum CD34+ cell dosage of 2x10 6 cells/kg are required to achieve haematopoietic engraftment 6,7 while some studies suggest that higher dosage, more than 5x10 6 cells/kg are associated with faster neutrophil and platelet engraftment 3,6 . Factors affecting PBSC mobilization had been studied, and beside the PB CD34+ cell count, no other factor has been shown to conclusively predict stem cell mobilizability 1,8 . Furthermore, factors unique to a particular population, may still be a significant determinant affecting stem cell mobilization, and should be studied.
Thus, the aim of this study was to determine the risk factors affecting poor PBSC mobilization in LPD patient at our center. Hopefully, the results obtained may provide a useful guidance to predict which patient is at risk for unsuccessful PBSC mobilization. Hence additional intervention can be incorporated early on to reduce of collection failure.

Study design and characteristic of patients
This was a retrospective record review involved of all lymphoma and MM patients who had undergone APBSC mobilization from January 2009 until December 2016 in Stem Cell Transplantation Unit, Department of Internal Medicine, Hospital USM. A total of 132 LPD patients underwent PBSC mobilization during 8 years of period. One lymphoma patient was excluded from this study because of procedural failure resulting from equipment failure. The factors that might affect PBSC mobilization including patient's characteristics, disease and treatment status and haematological parameters during collection from 131 patients had been investigated. The patients who were enrolled in this study include 39 with MM, 55 with non-Hodgkin lymphoma (NHL) and 37 with Hodgkin lymphoma (HL). For the purpose of the analysis, the disease was categorized into early stage (stage I and II for MM and lymphoma) and advanced stage (stage III for MM and stage III/IV for lymphoma). The disease staging for MM was done based on International staging system (stage I, II, III) 9 . While for lymphoma, Lugano classification was applied (stage I, II, III, IV) 10 .
A standard approach in our institution is that, patients with lymphoma that achieved partial response (PR) following anthracycline based chemotherapy; cyclophosphamide-doxorubicin-vincristineprednisone with or without rituximab (CHOP, R-CHOP) for NHL; or adriamycin-bleomycinvinblastine-dacarbazine (ABVD) for HL will undergo stem cell collection. The decision regarding APBSCT will be decided later on case to case basis. On the other hand, for MM a minimum achievement of PR with marrow plasma cells less than 5% following at least four cycles of thalidomide with or without velcade based chemotherapy regime was a prerequisite for stem cell collection. In addition, there were circumstances where stem cell collections were performed not confined to the above criteria.
Full blood count was evaluated daily with Sysmex XE-5000 haematology analyzer and daily peripheral blood (PB) CD34+ cells counts enumeration was started when total white blood cell count (WBC) had risen from nadir to more than 1.0x10 9 /L. Leukapheresis for PBSC collection was carried out when PB CD34+ cell counts were at least 20 cells/ uL. Since patients were enrolled over an extended period and different apheresis machines models were used at different period, collection protocol varied between one patient from another in accordance to model used. Leukapheresis procedures were performed by using Spectra Optia (Terumo BCT, Lakewood, CO USA) or Com.Tec (Fresenius, Lake Zurich, III) blood cell separator. A total of 2.5 to 3 times of the calculated patient's total blood volume was processed daily until targeted CD34+ dose of at least 3x10 6 cell/kg or 6x10 6 cell/kg were collected for lymphoma or MM respectively. Venous access was obtained via a double lumen catheter placed in the femoral vein.

CD34+ cell enumeration
Peripheral blood CD34+ cells count and leukapheresis product CD34+ dose was determined by using single-platform flow cytometric method using BD TruCOUNT TM Stem Cell Enumeration kit with CellQuest TM Pro program in Becton Dickson FACSCalibur TM flow cytometry and based on International Society of Haematotherapy and Graft Engineering (ISHAGE) gating strategy 11 .

Definition of outcome of mobilization
Successful mobilization is defined based on the collection of CD34+ dose of ≥2x10 6 cells/ kg by leukapheresis after a single mobilization procedure 12,13 .
Poor or unsuccessful mobilization is defined based on peak PB CD34+ cells count of <20 cells/ uL 14 or collection of CD34+ dose of <2x10 6 cells/ kg by leukapheresis after a single mobilization procedure 12,13 .

Statistical analysis
The data was analysed using SPSS (statistical package for the social) software version 22.0. The descriptive results were expressed as percentage, mean and standard deviation. Pearson Chi square (categorical variable) and independent T-test (numerical variable) were used to compare the independent variables between two groups, MM and lymphoma. The bivariate Pearson Correlation was used to evaluate correlation between PB CD34+ cell count and harvested CD34+ cell dosage. Simple (SLR) and multiple (MLR) logistic regression analysis were used for statistical analysis of potential risk factors that might influence poor PBSCs mobilization. From the result of SLR, clinically important independent variables with p value of <0.25 were included in the MLR analysis. The p value of <0.05 was considered significant.
Ethical clearance: The written consent to participate for APBSCT was obtained from all patients prior to initiation of PBSC mobilization treatment. This study was approved by Human Research Ethics Committee, Universiti Sains Malaysia with protocol number of USM/JEPeM/140362.

Results:
Majority of patients were Malay (93.9%) ethnic reflecting population distribution. Disease wise, lymphoma (70.2%) was the main diagnosisandthe mean age of all patients at mobilization, 41.4 years. Lymphoma patient underwent PBSC mobilization at younger age compare to MM patient and was statistically significant (p<0.001). The mean of collected CD34+ dosage for all patients was 9.6x10 6 cells/kg and MM patients had significant higher CD34+ dosage compare to lymphoma patients(p=0.003). The details of all patients characteristics and comparison between MM and lymphoma patients are shown in Table 1 and Table  2 respectively.

Poor mobilizer
Majority of patients (90.8%) had successful mobilization (CD34+ ≥2x10 6 cells/kg) and only 12 (9.2%) of patients was documented to have poor mobilization where half of them did not qualify for leukapheresis since PB CD34+ cell count was very low, <10cells/µL at the peak of GCSF stimulation. Nine of poor mobilizer were NHL (majority were diffuse large B cell lymphoma), and only two of them with HL and 1 with MM. The details of poor mobilizer patients are shown in Table 3.

Risk factors of poor mobilization
We found that there was significant strong positive correlation between PB CD34+ cell count and harvested CD34+ cell dosage (r = 0.774, p<0.001) (Figure 1). Univariate analysis documented that there were significant differences betweenpoor mobilization and age at mobilization (>60/≤60 years; p=0.029), PB CD34+ cells count (<20/≥20 cells/ uL; p<0.001), previous chemotherapy regime and cycle(≥3/<3 line; p=0.010 ans ≥8/<8 cycles; p=0.047 respectively) and premobilization platelet count (p=0.049) by. However from the multivariate analysis, the only factors that were mantained significantly associated with poor PBSC mobilization was age at mobilization (adjusted OR=38.43, p=0.005) and PB CD34+ cell count (adjusted OR=132.69, p<0.001). Patients who underwent PBSC mobilization at the age ≥ 60 years have increased odds of having poor mobilization by 38.4 times than patients who are <60 years. Meanwhile patients with PB CD34+ of <20 cells/uL have increased odds of having poor mobilization by 132.7 times than patients with PB CD34+ of ≥20 cells/uL.
Factors such as gender, weight, blood group, type of diagnosis, stage and status of disease, bone marrow infiltration, haematological parameter, duration of GCSF given, number of chemotherapy cycles and regimes, history of radiotherapy and duration of diagnosis to mobilization were not significant as risk factors of poor mobilization (Table 4).      The model reasonably fits well. Model assumption are met. There are no interaction and multicollinearity problems.

Discussion:
The incidence of poor mobilization in MM and/ or lymphoma patients who were mobilized with combination of GCSF and chemotherapy varied between 4% and 47% 4, [14][15][16][17][18][19] Our result showed only 9.2% of patients were poor mobilizer keeping within the reported range. Disease characteristic alone cannot predict mobilization outcome, as within high risk group there will be a group of patients who will successfully mobilize with standard approaches and there are patients within low risk group who mobilized poorly 6 . There are many reports on patients who failed to mobilized sufficient numbers of PBSC, however because of the retrospective nature of the study, heterogeneity of studied population, low numbers of patients, use of different mobilization regimes and lack of uniform criteria of poor mobilization led to difficulty in deriving a conclusion 8,20 .
Thus, there are no specific well accepted factors to identify potential poor mobilized patient.The most consistent findings for the predictive factors of stem cell mobilization is the PB CD34+ cell count 8,21,22 . Few study had proposed the predicted poor mobilizer included patient who failed a previous mobilization 1,22,23 , diagnosis of NHL 1,15,24 , increasing age 12,20,23 , advanced disease at the time of mobilization 18,23 , BM involvement 23,25,26 , thrombocytopenia before mobilization 12,26-28 , history of radiotherapy 23,25,27 , multiple chemotherapy regime 18,24,27 and increased weight 15,20,26 . Our study only demostrated that low PB CD34+ cell count and advanced age were the most important risk factor for poor PBSC mobilization.
Our findings showed agreement with the all the previous study that PB CD34+ cell count is the main important factors for predicting of CD34+ cell dosage 8,20,21 . There was significant strong positive correlation between PB CD34+ cell count and CD34+ cell dosage. The level of PB CD34+ lower than 20 cells/uL shown to be the risk factor for poor mobilization. Thus, identification of patient with low PB CD34+ cell counts may allow for the initial mobilization attemps with the alternative mobilize agent, thereby reducing the high failure rate and resource utilization. Alternative strategies of mobilization such as using other mobilizer agent like plerixafor (mozobil), may be considered, which had been reported to be very efficient for mobilization of poorly mobilized patients 6,25,29 . The possible defects of low PB CD34+ mobilization is due insufficient number of HSC in the bone marrow due to HSC intrinsic factors, low number or defective niches or inadequate number or response of BM effector/ supporter cells 30 .
Beside CD34+ dosage, we also found that advanced age (≥60 year old) was shown to be most important risk factor for poor mobilization as reported by previous study in which younger patients had better CD34+ cell dosage 13,20 and increasing age adversely affected CD34+ cell dosage 1,12,15 . However, some of the study have failed to show independent effect of patient age on mobilization 21,31 . The influence of patient age on autologous stem cell mobilization is unclear. However it is reasonable hypothesis that as people age, the functional capacity of the early stem cell compartment may decrease and furthermore, chemotherapy may have differential effects on stem cell compartment and its functional capacity 31 . Thus, transplant team should be aware of the increased risk of a poor mobilization in advanced age patients and alternative remobilization methods should be considered from the beginning.
Our result showed that number of previous chemotherapy and premobilization platelet level had significant association with poor mobilization at univariate analysis, but it was not significant at multivariate analysis and it is concordance to previous study that demostrated there was no significant impact between CD34+ dosage with number of previous chemotherapy 15,26 or platelet level 21,32 . However, many studies reported that premobilization platelet counts and numbers of chemotherapy have been significantly associated with the total CD34+ cell dosage where thrombocytopenia 20,27,28 and increased number of previous chemotherapy were risk factor for poor mobilization 1,18,24,27 . Theoritically, repetitives cycles of chemotherapy lead to rapid exhaustion of HSC self renewal and reconstitution potential and may also damage niches for HSC and BM macrophages effector cells 30 and resulted to poor PBSC mobilization. However, our result did not support this theory.
Although some studies had reported that the diagnosis of lymphoma and advanced disease as risk factor for poor mobilization, but there are a few studies did not demonstrate significant result 4,21 which is consistent with our finding. Although our result showed that the diagnosis and stage of disease were not a risk factor for poor mobilization, but we found that 11 out of 12 poor mobilized patients were lymphoma and majority were NHL (9 patients). Majority of them also presented with advanced stage (stage III or IV).
Our study also did not show that the history of previous radiotherapy and increased weight as risk factor for poor mobilization although a few studies had reported their significant impact on the ability to mobilize and associated with poor CD34+ yield in APBSC collection 18,20,27 . A lot of other factors that had been investigated by previous studies such as gender, stage of disease, premobilization haemoglobin level and premobilization WBC count and similar results had been demonstrated with our results 14,15,21 . We demonstrated that those factors was not a risk factor for poor APBSC mobilization.
This study has limitations. Some of the independent variable could not be studied due to retrospective record review method. Furthermore, our result showed wide range of of 95% confidential interval value in MLR analysis indicates a poor distribution between the two comparing groups due to small sample size. Thus the findings need to be inferred with cautious since it might not be representative of the reference population.

Conclusions:
Patients for stem cell collection should be carefully screened for the presence of risk factors so that poor mobilizers can be identified early and suitable interventions can be incorporated to reduce rate of mobilization failure. Although various factors had been reported, we found that PB CD34+ count less than 20 cell/µL and age more than 60 years old were associated with poor mobilization outcome in patients with LPD planned for APBSCT.