INTRODUCTION
Allogeneic hematopoietic stem cell transplantation (HSCT) is the only treatment modality that provides a cure or produces long-term survival for a variety of pediatric diseases, including certain hematologic malignancies, bone marrow (BM) failure, hemoglobinopathies, immuno-deficiencies, and metabolic disorders [1, 2]. The development of BM, umbilical cord blood, and mobilized peripheral blood as graft sources for hematopoietic stem cell transplantation has occurred [3]. Despite the increased use of mobilized peripheral blood and umbilical cord blood, BM remains the primary graft source in pediatric patients [2]. HLA-matched siblings are considered the best donors for HSCT due to the reduced risk of transplant-related complications and improved clinical outcomes [1, 2, 4]. BM donation is a well-tolerated procedure with few complications, most of which are perioperative events, including anesthesia-related events, postoperative pain, and anemia [5, 6]. Stroncek et al. [7] reported a significant correlation between the marrow harvest procedure duration and anesthesia duration with donor pain and fatigue following BM donation; however, BM harvest procedure duration had the highest correlation with post-harvest pain and fatigue.
The National Marrow Donor Program (NMDP) recommends that the collected BM volume not exceed 20 mL/kg of donor body weight to minimize the risk associated with the BM harvesting procedure on the donor and to have a sufficient volume of collected BM; this volume was determined based on adult donors and is suitable for children as well [6]. Furey et al. [6] developed a potential algorithm for calculating the BM harvest volume. The proposed algorithm revealed that BM harvest volume required to achieve the desired CD34+ cell count (×106 cells/kg of recipient weight) depends on the donor age, and harvesting up to 20 mL/kg of donor weight may not be necessary for all pediatric donors. This may result in a smaller BM harvest volume without pediatric transplant outcome. This algorithm needs large-scale validation.
Age-related changes in human hematopoietic stem/progenitor cells have been reported [8], and Yabe et al. [9] confirmed that pediatric donors younger than 5 years provide a larger number of nucleated cells and CD34+ cells relative to their body weight than older pediatric BM donors, which may enable a smaller harvest volume. Collecting a smaller harvest volume would benefit donors by decreasing the anesthesia duration, reducing the risk of blood allotransfusion, decreasing post-procedure pain, and shortening the recovery period [6]. However, Bittencourt et al. [10] reported that a CD34+ cell dose of 3×106 cells per kg of recipient weight or more from a matched sibling donor improved hematopoietic recovery in pediatric and adult patients.
This retrospective study determined the optimal amount of BM for pediatric recipients from matched sibling donors. Additionally, we examined transplant survival, post-transplant infections in the first 100 days, and Graft-Versus-Host Disease (GVHD) free or relapse-free survival (GRFS) to CD34+ cell dose per kg of recipient weight. We validated the algorithm proposed by Furey et al. [6] to determine the BM volume for harvesting 5×106 CD34+ cells per kg based on recipient weight (mL/kg of donor weight) [(recipient weight×5.0)÷(donor weight×0.7), (recipient weight×5.0)÷ (donor weight×0.36), and (recipient weight×5.0)÷(donor weight×0.3) for corresponding donor age groups <6, 6–12, and >12 yr, respectively].
MATERIAL AND METHODS
Study design
In this retrospective study, we reviewed the medical records of all pediatric sibling donors (age at harvest <18 yr) who had undergone a single BM harvest between January 2007 and December 2017 at our center. The medical records of all consecutive pediatric transplant-naïve recipient-donor pairs were obtained from the computerized patient information management system. Data of the volume of BM harvested relative to donor weight, harvest-related hospitalization days and side effects, transplant characteristics, and outcome-related parameters for recipients were collected. The parents or legal guardians provided written informed consent as per institutional practice.
BM harvest procedure
Multiple punctures were used to harvest BM from the posterior iliac crest under general anesthesia. With each attempt, a volume of approximately 5 mL was aspirated [11, 12]. The target volume was 15–20 mL/kg of recipient weight and did not exceed 20 mL/kg of donor weight [13]. The BM product was processed in a stem cell laboratory, and flow cytometry was used to count CD34+ and CD3+ cells.
Definitions
We used a range of ±5 kg to determine whether the donor’s weight was the same, lower, or higher than that of the recipient. Overall survival (OS) was defined as survival with or without the primary disease. However, transplant-related mortality (TRM) was defined as death due to any cause other than relapse of malignant disease. Acute GVHD was defined according to the method described by Glucksberg et al. [14]. Chronic GVHD was defined in accordance with the National Institutes of Health guidelines [15]. The time to neutrophil engraftment is the first day of achieving an absolute neutrophil count ≥0.5×109/L for three consecutive days from the time of BM infusion. The time to platelet engraftment was the first day of platelet count >20×109/L without transfusion for 7 days after BM infusion. The CD34+ cell dose was ×106 cells/kg of recipient weight. A composite index of GRFS was defined as survival in the absence of grade III–IV acute GVHD, systemic therapy-requiring chronic GVHD, relapse of malignant disease, rejection of the graft in patients with a nonmalignant disorder, or death from any cause during the first year after allogeneic transplantation [16]. Event-free survival (EFS) was defined as survival in the absence of any of the following during follow-up: relapse of primary disease or rejection of the graft (primary and secondary), development of a new malignancy, or death from any cause throughout the course of follow-up.
Infection episodes in recipients
All infections in transplant recipients from day 0 to day 100 post-transplantation were reviewed. Bacteremia was defined as the presence of viable bacteria in the blood. Viremia was defined as a cytomegalovirus load >500 copies/mL, an Epstein-Barr virus load >1,000 copies/mL, or an adenovirus load >1,000 copies/mL based on PCR results [17]. Definitions of invasive fungal infections were performed as previously described by Satwani et al. [18].
Statistical analysis
Descriptive statistics for continuous variables were provided as medians and ranges, and categorical variables were presented as numbers and percentages. We used the chi-square or Fisher’s exact test to test the significance of the associations between categorical variables. Independent-sample Mann–Whitney U tests and Kruskal–Wallis tests were used to test the significance of the differences between two and more than two categories of continuous variables, respectively. Spearman’s correlation coefficient was used to calculate correlations between continuous variables. Binary logistic regression was used to test the effects of CD34+ cell dose per kg of recipient weight on GVHD, survival, and GRFS. SPSS Statistics for Windows (version 20.0, IBM Corp., Armonk, NY, USA) was used to analyze the data. P<0.05 was considered to be statistically significant.
RESULTS
Recipients’ characteristics
We included the data of 531 donor-recipient pairs in this study. The median age of the recipients was 4.4 years, and >50% of them were boys (N=298, 56.1%). Most recipients underwent transplantation for nonmalignant disorders. Table 1 shows demographic and transplant characteristics. The median CD34+ cell dose infused was 7.0×106 (1.3–17.7) per kg of recipient weight.
Donor characteristics and BM harvest volume
The study cohort comprised all pediatric (age at harvest <18 yr) sibling donors who underwent a single BM harvest and their recipients. Of the 531 donor-recipients pairs, 276 (52.0%) donors were boys, 146 (27.5%) were younger than 5 years old at harvest, 182 (34.3%) were between 5 and 10 years old, and the remaining 203 (38.2%) were 10 years old or older; the median donor age was 8.4 years (1.0–18.0), 66 (12.4%) donors weighed less than their recipients, 108 (20.3%) had equivalent weights, and 357 (67.2%) weighed more.
The median BM volume harvested per kg of donor weight was significantly higher for lower-weight donors than for equivalent or higher-weight donors (15.7 mL/kg vs. 14.6 mL/kg or 9.0 mL/kg, respectively, P<0.001). The median volume of BM harvested per kg of recipient weight was significantly higher in higher-weight donors than in equivalent-weight or lower-weight donors (22.3 mL/kg vs. 14.8 mL or 8.7 mL/kg, respectively, P<0.001).
Hemoglobin levels and Packed Red Blood Cells (PRBC) transfusion
The median hemoglobin (Hb) levels before harvest were significantly lower in lower-weight donors than in equivalent or higher-weight donors (115.0 vs. 119.0 or 125.0 g/L, respectively, P<0.001), and the same was true for post-harvest levels (lower weight: 97.0 vs. equivalent weight: 97.0 vs. higher weight: 110.0 g/L, P<0.001). Additionally, the median decrease in hemoglobin levels after harvest was significantly different among the three donor weight groups (lower weight: 20.0 vs. equivalent weight: 21.0 vs. higher weight: 16.0 g/L, P<0.001).
Given the indication for PRBC transfusion, including symptomatic anemia (hypotension not responding to fluid boluses, tachycardia) or Hb below 80 g/L, 47 (8.9%) donors required allogeneic PRBC transfusions post-harvest. The need for PRBC transfusion for donors was significantly associated with their age group at harvest (P<0.001); donors younger than 5 years old had the highest association with the need for PRBC transfusion (N=30, 20.5%), followed by donors between 5 and 10 years (N=13, 7.1%), and then those ≥10 years old (N=4, 2.0%). The median BM harvest volume relative to donor weight (kg) among donors requiring PRBC transfusion was 17.0 mL/kg (8.1–21.8 mL/kg), while that among those donors who did not require PRBC transfusion was 10.2 mL/kg (1.4–21.0, P<0.001).
For post-harvest PRBC-transfused donors, the median pre- and post-harvest hemoglobin concentrations were 112.0 g/L (91–143 g/L) and 89.0 g/L (75.0–128.0 g/L), respectively, with a median decrease in the hemoglobin level of 26.0 g/L (2.0–62.0 g/L). The median pre-harvest hemoglobin concentration, post-harvest hemoglobin concentration, and decrease in the hemoglobin level were 124 g/L (91.0–169.0 g/L), 106.0 g/L (66.0–152.0 g/L) and 18.0 g/L (0.0–51.0 g/L), respectively, for donors who did not require post-harvest PRBC transfusion (P<0.001, <0.001, and 0.007, respectively). The need for PRBC transfusion was also significantly associated with donor weight at harvest (P< 0.001); 17 (25.8%) lower-weight donors, 23 (21.3%) equivalent-weight donors, and 7 (2.0%) higher-weight donors required PRBC transfusion.
CD34+ cell yield (×10<sup>6</sup> cells per kg of recipient weight) relative to donor age
The CD34+ cell yield per kilogram of recipient weight was significantly higher in younger donors. In particular, the median CD34+ cell yield was 6.4×106 (1.2–33.8×106) in donors younger than 5 years old at harvest, while the yields were 4.7×106 (0.3–28.5×106) for those aged 5–10 years and 2.1×106 (0.3–11.3×106) for those older than 10 years of age (P<0.001).
CD34+ cell yield (×10<sup>6</sup> cells per kg of recipient weight) relative to harvest volume
Harvested BM was able to satisfy the minimum CD34+ cell dose of ≥3.0×106 per kg of recipient weight [19] in 506 (95.3%) paired recipients. In this subgroup of 506 donor-recipient pairs, the median CD34+ cell yield (×106 per kg of recipient weight) was higher for the youngest donor group (8.0 vs. 7.5 vs. 7.0, P=0.350) and donors who weighed more than the corresponding recipient (lower weight: 5.1 vs. equivalent weight: 6.4 vs. higher weight: 8.1, P<0.001, Table 2). A weak negative correlation between the BM harvest volume (mL/kg) per donor weight and CD34+ cell yield ×106 per kg of recipient weight was observed (Spearman’s rho=-0.154, P<0.001). Fig. 1 shows a scatterplot depicting ANC recovery with respect to the total volume harvested per donor weight and CD34+ cell dose ×106 per kg of recipient weight.
Hematopoietic cell recovery in recipients
The median time for neutrophil engraftment was 16 days (10–37), and platelet recovery was 29 days (6–148 days). No correlation was found between the infused CD34+ cell dose (×106 per kg of recipient weight) and platelet recovery time (Spearman’s rho=0.02, P=0.686). However, the infused CD34+ cell dose (×106 per kg of recipient weight) was negatively correlated with ANC recovery time (Spearman’s rho=-0.105, P=0.024). However, the malignant and nonmalignant subgroups differed in transplant analysis. The median time for neutrophil engraftment for recipients with the malignant disease was 17 days (10–49 days) compared with 15 days (10–51 days) for recipients with a nonmalignant disorder (P<0.001). The median time to platelet engraftment for recipients with the malignant disease was 26 days (15–110 days) versus 31 days (6–148 days) for recipients with the nonmalignant disorders (P=0.002).
Among all recipients, ANC engraftment by day +28 was observed in 450 (84.7%), whereas 72 (13.6%) did not achieve this milestone, and late ANC engraftment (beyond day +28) was recorded in 9 (1.7%). The median CD34+ cell dose infused (×106 per kg of recipient weight) was higher in the recipients who achieved ANC engraftment by day +28 than in those who achieved engraftment after day +28 [6.8×106, (1.5–17.22) vs. 6.5×106, (1.3–11.0), P=0.239]. However, the infused CD34+ cell dose (×106 per kg of recipient weight) was significantly lower in those who achieved ANC engraftment by day +28 than in those who did not achieve this milestone [6.8×106, (1.3–17.2) vs. 9.0×106, (3.0–17.7), P< 0.001]. We classified ANC engraftment according to the CD34+ cell dose infused (×106 per kg of recipient weight): <4×106 (N=58), 4 to <7×106 (N=203), 7 to <10×106 (N=167), and 10×106 and above (N=103). The highest decline in the ANC engraftment rate of 14.9% relative to the CD34+ cell dose (×106 per kg of recipient weight) infused was observed in the 7 to <10×106 category, which ranged from 84.45 to 71.8%. A lower cell dose correlated with ANC engraftment (P<0.001; Fig. 2).
Infections during the first 100 days post-infusion
Infections during the first 100 days after transplantation were recorded in 168 (31.6 %) recipients. Of the 247 isolates, 156 (63.2%) were bacterial, 51 (20.6%) were viral, 38 (15.4%) were fungal, and 2 (0.8%) were parasitic. No significant association between the CD34+ cell dose infused (×106 per kg of recipient weight) and the incidence of infection was found (odds ratio, 0.982; 95% CI, 0.939–1.03; P=0.443).
Survival and GRFS
With a median follow-up time of 64.3 months (95% CI, 58.5–70.0 mo) and 100 mortality events, the 3- and 5-year cumulative OS probabilities for our cohort were 81.7±1.7% (events=95) and 80.6±1.8% (events=99), respectively. The 1-year TRM was 15.6% (N=82). The infused CD34+ cell (×106 per kg of recipient weight) and TNC (per kg of recipient weight) doses were not associated with overall mortality (P=0.589 and P=0.991, respectively). The GRFS rate was 73.8% (139 events) during the follow-up period. Last follow-up EFS was 73.8% (119 events). The infused CD34+ cell dose (×106 per kg of recipient weight) was not significantly associated with GRFS (P=0.829). The incidence of overall acute GVHD (grades I–IV) was 23.9% (N=127), of which grade III–IV events accounted for 24.4% (N=31). Additionally, the CD34+ infused cell dose (×106 per kg of recipient weight) was not found to be associated with the incidence of acute GVHD, severe acute GVHD (grades III–IV only), or chronic GVHD (P=0.541, 0.645, and 0.833, respectively). Based on chimerism studies, our recipients day +100 engraftment rate was 97.7% (N=508); 9 (1.7%) had primary graft failure, and 3 (0.6%) had secondary graft failure. We could not evaluate 11 (2.1%) recipient’s engraftment status on day +100 because they died early.
Comparison of the harvest volume with the NMDP Guidelines
The BM volume harvested from the sibling donors at our institution was within ±2.0 mL/kg of donor weight, as specified in the NMDP guidelines of up to 20 mL/kg of donor weight. However, compared with the optimum harvest volume calculated according to the formula proposed by Furey et al. [6], our harvest volume was equivalent (within ±2.0 mL/kg of donor weight) in 149 (28.1%) donors, lower in 43 (8.1%) donors, and higher in the remaining 339 (63.8%) donors. Upon subdividing the donors harvested higher than that proposed by Furey et al. [6] into three categories and then analyzing for ANC engraftment, we observed that subsequent increases above ±2.0 mL/kg of equivalent harvest volume, when infused in recipient pairs, did not produce more favorable results in terms of overall ANC recovery (P=0.001; Fig. 3, Table 3) and non-primed donors (P=0.017; Table 3).
DISCUSSION
This is among the most extensive retrospective single-center studies to date, reporting and analyzing the relationship between BM harvest volume and infused CD34+ cell doses (×106 per kg of recipient weight) in the pediatric population. In this study, several important observations were made. A CD34+ cell count of >7×106 per kg of recipient weight was not associated with improved hematopoietic recovery. Harvesting Large volumes may result in an unnecessarily high CD34+ cell count, and the volume of BM harvested can be safely reduced for younger donors. The proposed Furey et al. [6] algorithm may provide guidelines for determining the BM harvest volume required to obtain the desired CD34+ cell dose (×106 per kg of recipient weight), resulting in a safe reduction in the harvested volume in pediatric donors.
BM donation is a safe and well-tolerated procedure. In a prospective study conducted by the European Group for Blood and Marrow Transplantation Pediatric Diseases Working Party, the risk of blood transfusion after BM harvest was associated with donors younger than 4 years and a BM harvest volume exceeding 20 mL/kg [2]. In our study, the need for allogeneic blood transfusions was significantly associated with age at harvest, especially in donors younger than 5 years old and donors with disproportionately low weight relative to the recipient’s weight. This may lead to careful consideration of the maximum volume of BM that can be harvested from young children.
The impact of infused CD34+ cell dose (×106 per kg of recipient weight) on clinical outcomes in children undergoing allogeneic HSCT, mainly with unrelated donors, has been reported. PBSCT and UD BMT showed that infusion of a higher CD34+ cell dose (×106 per kg of recipient weight) is associated with improved patient outcomes [20, 21]. A few studies have reported the outcome of CD34+ cell doses from a matched sibling donor, and a recommended dose of ≥3×106 CD34+ cells per kg of recipient weight was associated with hematopoietic reconstitution [6, 10]. We found that the infused CD34+ cell dose (×106 per kg of recipient weight) was not associated with recipient OS, EFS, or GRFS. We did not observe any correlation between the infused CD34+ cell dose (×106 per kg of recipient weight) and platelet recovery time or the incidence of infections in the first 100 days. Our data showed that a CD34+ cell dose of >7×106 per kg recipient weight did not improve hematopoietic recovery in children with an HLA-matched sibling donor. As previously reported [9, 19, 22], our data confirm that a CD34+ cell dose (×106 per kg of recipient weight) is significantly related to donor age; donors younger than 5 years yielded higher concentrations of CD34+ cells, suggesting that the harvested volume can be reduced safely without a negative impact on the CD34+ cell count [6, 9].
Increasing collected BM volume boosts CD34+ cell yield. However, increasing the collected volume can increase anesthesia time, puncture sites, and blood loss [23]. Our data showed that in BM harvested at volumes up to 20 mL/kg of donor weight, CD34+ cell doses ≥3×106 per kg of recipient weight were collected from 506 (95.3%) donors.
Few studies demonstrated a positive correlation between the total harvested BM and TNC number [12, 24]. However, we found no positive correlations between the total BM volume and TNC number or CD34+ cell count (×106 per kg of recipient weight), suggesting that the collected BM volume can be decreased in pediatric donors, especially in those below 5 years.
Using our data, we validated the algorithm proposed by Furey et al. [6] regarding the optimal dose of CD34+ cells (×106 per kg of recipient weight) and BM volume. We demonstrated that incremental increases in harvested volume did not provide more favorable results in ANC recovery, further emphasizing that BM harvesting at volumes up to 20 mL/kg of a child donor weight might not be necessary for optimal clinical outcomes in every pediatric recipient.
In conclusion, this is the largest retrospective study to report and analyze the relationship between marrow harvest volume and CD34+ cell dose (×106 per kg of recipient weight) in the pediatric population from a single transplant center. Unmanipulated CD34+ cell dose >7×106 per kg of recipient weight did not improve hematopoietic recovery. It was not correlated with platelet recovery time, infectious toxicity, GVHD incidence, mortality, or GRFS. Small BM can safely be harvested in younger donors. The proposed algorithm could determine the BM harvest volume needed to obtain the desired CD34+ cell dose, reducing the harvested volume safely in pediatric donors. A larger observational study is needed to determine the ideal minimum CD34+ cells dose for pediatric sibling allogeneic HSCT.