Hematopoietic stem cell transplantation (HSCT) has been successfully used as a treatment for a wide variety of both hematological and non-hematological diseases.1–8 Cryopreservation has been routinely employed to maintain the viability and proliferative capacity of peripheral blood stem cells (PBSC) prior to autologous stem cell transplantation (ASCT).9–11 This therapeutic strategy has been shown to be safe and is associated with a low occurrence of significant side effects related to infusion and graft failure.

The most common strategy for PBSC mobilization consists of administering granulocyte colony stimulating factor (G-CSF) with or without chemotherapy to stimulate cell migration to the peripheral blood.2,12,13 There are several mobilization protocols, however, despite being a well-established procedure, prediction of successful mobilization is challenging. Several factors have been associated with the outcomes of mobilization, such as age, diagnosis, and previous chemotherapy or radiotherapy. Insufficient mobilization is a major impediment for performing ASCT with poor mobilizers having delayed neutrophil and platelet recovery even with similar infused CD34+ cell doses.3,5,14,15

The enumeration of CD34+ cells is considered the gold standard to define the best time to perform the collection of hematopoietic stem cells (HSC) by apheresis.6,13,14,16–18 Pre-apheresis CD34+ cell enumeration is used to guide the procedure with the result being directly related to the success of mobilization.14,16 However, CD34+ cell enumeration is an expensive and time-consuming procedure that requires trained staff and highly-specialized laboratory facilities.6–8,14,16,17,19 Thus, the number of laboratories that perform CD34+ enumeration is limited, especially in developing countries. Some transplant institutions outsource this procedure, which leads to operational and logistical difficulties for collection.

To date, there is no model to successfully predict mobilization, and surrogate markers of CD34+ enumeration are limited. Sysmex has developed a hematological parameter that identifies a population of immature myeloid peripheral blood cells, called “hematopoietic progenitor cells” (HPC), that is based on size, density and resistance to lysis.2,6,16,18,20 However, this is not an available feature of blood count analyzers in some countries, including Brazil. The aim of this study was to identify characteristics that might influence mobilization efficacy and to create a model for the prediction of stem cell mobilization in ASCT candidates. This study analyzed factors associated with successful mobilization and, separately, factors associated with inadequate mobilization.

Although the pre-apheresis enumeration of CD34+ cells is known to be the gold standard to define the timing of HSC collection by apheresis, the results of the current study show that other clinical and laboratory data are associated with pre-apheresis CD34+ cell counts and can be useful during the collection process. Multiple myeloma diagnosis, only one mobilization cycle, and high nRBC, PBMNC and IG counts as well as low MCV were significantly associated with good mobilization. On the other hand, low body weight and two or more different chemotherapy regimens were significantly associated with poor mobilization.

G-CSF is a drug used to treat secondary neutropenia and acts to control hematopoiesis.12,15,23,24 According to the instructions provided by the manufacturer, G-CSF is a glycoprotein that regulates the production and release of functional neutrophils from the bone marrow. Furthermore, it induces secondary increases in circulating eosinophils and basophils. The IGs and nRBCs are precursor cells and we believe that the use of G-CSF could also stimulate the migration of these cells to the peripheral blood. The intense stimulation of cell production could generate smaller cells, which would explain the lower MCV in good mobilizers.

The size and internal complexity of HSCs are similar to those of monocytes and lymphocytes.25 Due to these characteristics, HSCs can be read as PBMNC in a automatic hematology counter, so an increase in the PBMNC count in good mobilizers is expected.

Initially, it was believed that the WBC count and the number of HSCs collected by apheresis were correlated. However, several studies have demonstrated the absence of any correlation between WBC and the enumeration of CD34+ cells in peripheral blood.7,13,14,16,17,26 The data of the present study corroborates previous studies showing an absence of an association between WBC and the pre-apheresis CD34+ cell count.

Several studies demonstrated a good correlation between the Sysmex HPC parameter and pre-apheresis CD34+ cell counts. However, reports also describe that this association may vary depending on the patient's disease and the mobilization regimen used.2,6,7,16–20 Despite the strong correlation found in the literature, the HPC parameter is not available on equipment sold in some countries, which makes its wide-spread use unfeasible. This study found other parameters associated with pre-apheresis CD34+ cell enumeration. These parameters could be used as an alternative to optimize HSC harvesting by apheresis.

Acquisition of new equipment involves complex logistics and high costs. Using data already available with current instrumentation is a way to optimize financial resources and improve the services offered to transplant centers. The results of this study are extremely relevant to laboratories and institutions that do not have a flow cytometer and rely on other institutions to define HSC collection by apheresis. Collection of HSC by apheresis and enumeration of pre-apheresis CD34+ cells are currently performed in different facilities separated by more than 40 km distance from our institution. The average time required to transport the sample, perform the CD34+ enumeration test, and release the result is 2.5 h. Our objective was to develop a model based on clinical and laboratory characteristics that predict successful mobilization to enhance PBSC collection in such institutions. The validation data show that the use of the model would rarely lead to starting an apheresis procedure in a patient with pre-apheresis CD34+ cell counts <10 µL. Although some good mobilizers are not identified by the model due to its low NPV, the only consequence for these patients would be to wait for the CD34+ count test by flow cytometry. The predictive accuracy of our model is also a step toward the development of accurate prognostic tests to identify individuals at risk of unsuccessful mobilization and to help select better therapeutic options.

The model may provide some benefits in the clinic. The first one is the optimization of the collection and the possibility of scheduling more than one collection per day. Starting a collection in the early morning would enable another collection in the afternoon, which would benefit patients waiting for an opportunity to undergo transplantation. In addition, it would allow CD34+ cell enumeration in the leukapheresis product on the same day in institutions that do not have night shifts. The enumeration of PBSC yield in apheresis procedures on the same day would bring several benefits to patients and transplant centers: (1) removal of the catheter on the same day, reducing the inconvenience caused and the risk of infections; (2) absence of the need to administer another dose of the mobilization regime preventively; (3) early release of a hospital bed that could be used for another patient.

The identification of characteristics associated with poor mobilization enables the adoption of additional clinical strategies for patients at high risk of unsuccessful mobilization. For example, if a patient is referred for ASCT and has already undergone several chemotherapy regimens, the transplant center could plan the mobilization with plerixafor. Other studies have identified other factors associated with poor mobilizers, such as age, prolonged chemotherapy and previous and extensive radiotherapy. Although researchers have tried to define characteristics associated with poor mobilization, as yet there is no consensus.2,3,19

Limitations of this study include its retrospective design and missing data on patients’ charts. Another limitation is the absence of patient height to calculate the body mass index (BMI); its use instead of weight would be more reliable. It is known that obesity can influence transplant outcomes and the best to calculate it would be through the BMI.27–29 Finally, there is a need to externally validate the prediction model in an independent population.

In conclusion, successful mobilization was more common in participants with higher weight, those undergoing their first mobilization cycle, and those with a diagnosis of multiple myeloma. Furthermore, high PBMNC, nRBC and IG counts as well as low MCV were associated with successful mobilization. A predictive model using these variables was established to identify successful mobilization. This model was validated in a subset of a study population and it identified participants who achieved successful apheresis with 76 % accuracy. These data can be used to help streamline and optimize the collection of HSC by apheresis.