Umbilical Cord Blood Collection and Separation for Haematopoietic Progenitor Cell Banking

Umbilical cord blood collection and separation for haematopoietic progenitor cell banking

JA Ademokun1, C Chapman2, J Dunn1, D Lander2, K Mair2, SJ Proctor1 and AM Dickinson1

1School of Clinical Sciences, Department of Haematology, University of Newcastle-upon-Tyne and 2National Blood Service, Newcastle-upon-Tyne, UK

Summary: wide may reduce the present delay between the initiation of a search for an unrelated donor and actual transplantation Cord blood transplantation has been proven to be a that occurs using existing unrelated marrow donor panels.3 suitable form of treatment for a variety of diseases in Separation and processing of cord blood samples in large childhood and more recently in an increasing number of numbers for storage in cord blood banks ideally needs to adult patients. Banks of cord blood cryopreserved after be partially automated to allow large numbers of samples HLA testing are required in order to provide various to be processed efficiently. A closed system reduces the risk HLA types for unrelated transplantation. To optimize of bacterial contamination after collection. The processing storage space cord blood needs to be stored as a separ- method must allow adequate recovery of nucleated cells ated product. Several early methods of cord blood sep- and progenitors to enable engraftment. Early attempts at aration resulted in a significant loss of progenitor cells. separating cord blood by density gradient techniques led to We used a separation procedure where the donation loss of mononuclear cells and to the suggestion that cord was separated by centrifugation into a buffy coat frac- blood should be stored unseparated.4,5 These donations, tion, a red cell fraction, and a plasma fraction. Twenty- cryopreserved as whole blood, have been transplanted suc- five samples, (mean initial volume 81 ml) were assessed. cessfully.1,2,5 However, for cord blood banks to be econ- Nucleated cells were recovered in the buffy coat frac- omical and efficient, volumes smaller than that of whole tion. Recoveries of nucleated cell count, total progeni- packs need to be stored. Furthermore, clinical problems tors and CD34-positive cells in the buffy coat were 90%, could arise with red cell lysis and some form of red cell 88% and 100%, respectively. The buffy fraction was depletion is necessary to reduce the risk of transfusion reac- tested for sterility by aerobic and anaerobic culture. tions in cases of ABO incompatibility between donor cord Using this closed bag system, volume reduction was ach- blood and recipient. A study using the density gradient ieved while maintaining sterility and retaining progeni- method to separate cord blood6 did not find a major loss tor cells in a final mean buffy coat volume of 44 ml. of progenitor cells, and a later study7 suggested a modifi- Red cell and plasma fractions were available for ABO cation of the common density gradient separation methods grouping, virology testing and cryopreservation. The as suitable for processing cord blood before cryopreserv- results show that cord blood can be effectively volume- ation. Falkenburg et al6 compared cell separation of cord reduced using simple and readily available blood bank- blood and bone marrow by red cell lysis, methylcellulose ing techniques. sedimentation and density gradients, and showed similar Keywords: cord blood collection; separation; banking recovery of nucleated cells from both cell sources provided separation was initiated within 8 h of collection. Other techniques investigated include 3% gelatin sedimentation,9 and this has been successfully applied in a clinical setting.10 Cord blood has become an acceptable source of haematopo- In our present study, we have used a triple bag system ietic marrow repopulating stem cells. A variety of malig- in which the cord blood is separated by centrifugation into nant and non-malignant haematological diseases have been three components: a buffy coat, a red cell and a plasma treated using transplanted cord blood but the total number fraction. This method of processing allowed for mainte- 1,2 of transplants performed remains small. The best use of nance of sterility, volume reduction and separation of red umbilical cord blood is likely to come when cord blood blood cells and some granulocytes from buffy coat progeni- banks are established with large numbers of HLA typed tor cells. samples stored ready for prompt use in unrelated recipients. This usage has recently been initiated by the New York group who have transplanted 25 patients with unrelated cord blood transplants.2 Expansion of this procedure world- Materials and methods

Correspondence: Dr AM Dickinson, Department of Haematology, Royal Informed consent Victoria Infirmary, Queen Victoria Road, Newcastle-upon-Tyne, NE1 4LP, UK The collection of cord blood was performed in the depart- Received 21 October 1996; accepted 3 February 1997 ment of Obstetrics and Gynaecology at the Royal Victoria Cord blood collection and separation JA Ademokun et al 1024 Infirmary, Newcastle-upon-Tyne by experienced midwives counts in each collection (ie WBC minus nucleated red trained in cord blood collection. Nurse counsellors obtained cells). informed consent at a pre-natal clinic visit. Written informed consent was obtained from the mother for cord blood collection which could not be reserved for family use, and to hepatitis and HIV testing at the time of delivery and, for Immunophenotyping a maternal blood retest at 6 months after delivery. At the CD34 estimation was carried out on whole blood and the time of delivery the midwives also collected information buffy coat fraction after a method described by Sutherland on the type of delivery, weight of placenta, and weight, sex et al12 involving simultaneous staining with Becton Dickin- and condition of the infant. son, Oxford, UK (BD) monoclonal anti-CD45 (anti-HLe- 1) fluorescein isothiocyanate (FITC) and anti-CD34 (anti- HPCA-2) phycoerythrin (PE)-conjugated antibodies. BD anti-IgG1 PE was used as control. Flow cytometry was per- Collection of cord blood formed on a Becton Dickinson FACScan and analysis was More than 200 collections have now been made during the performed using Lysis II software. CD45-positive events third stage of labour, with the placenta in utero. At delivery, were gated into a region R1 on a fluorescence (FL) 1 vs the cord was clamped, disinfected and the blood collected forward scatter (FSC) dot plot to exclude nucleated red from the umbilical vein into either a single blood collection blood cells, platelets and debris. The CD45-positive events bag (Baxter Health Care, Newbury, UK) or into a triple were analyzed for CD34 positivity by assessing events with bag set (R8361; Baxter Health Care). The single blood bag low side scatter in region R2 on a FL2 vs side scatter (SSC) contains citrate/phosphate/dextrose-adenine (CPD-A) anti- dot plot. coagulant, and the triple bag comprises a collection bag with citrate/phosphate/dextrose (CPD) as anticoagulant, and two separate bags for red cell and plasma collection. Both Progenitor cell assays bags were modified from standard for our use by reducing the 63 ml of anticoagulant to 15 ml by removing 48 ml with Mononuclear cells (MNC) were obtained by density gradi- a sterile needle and syringe. ent centrifugation over Lymphoprep (Nycomed Pharma, The volume of the cord blood collected was calculated Oslo, Norway; 1.077 g/ml). The cells were washed, counted from the weight of each bag before and after the collection. in a Neubauer chamber and viability assessed by Trypan Samples were taken at this stage for a white blood cell blue exclusion. The cells were plated at 1 × 105/ml in 0.9% count (WC), CD34-positive cell count and progenitor cell methylcellulose (StemCell Technologies: Metachem Diag- assays. The cord blood was collected and when necessary nostics, Northampton, UK) containing Iscove’s modified (ie following overnight deliveries) stored at 4°C for 12 h Dulbecco’s medium (IMDM; Life Sciences International, prior to separation and processing. Aerobic and anaerobic Basingstoke, UK) with supplements including 20% fetal bacteriology culture was performed on all donations. calf serum, sodium bicarbonate, pyruvate, glutamine and growth factors (erythropoietin (Boehringer Mannheim, Lewes, UK) 1.5 units/ml, recombinant granulocyte colony- stimulating factor (rG-CSF) 1000 units/ml and stem cell Separation of cord blood and nucleated cell counts factor (SCF) at 20 ng/ml). Both rG-CSF and SCF were gifts from Amgen, Cambridge, UK. Plates were incubated at Twenty-five donations collected into triple bags were separ- ° ated by centrifugation at 4400 g for 18.5 min at 22°C using 37 C in humidified 5% carbon dioxide in 95% air. Granulo- a Heraeus Cryofuge 8000 (Heraeus Instruments Ltd, Brent- cyte–macrophage colonies (CFU-GM), haemoglobinised wood, UK). The method described for separation of the (BFU-e) colonies, and mixed erythroid/myeloid colonies cord blood is a modification of that originally described by (CFU-GEMM) containing greater than 40 cells were Ho¨gman et al11 for peripheral blood component preparation scored on day 14. In 17 MNC samples the total number of colonies (CFU-GM, CFU-GEMM and BFU-e) per 105 and storage. After centrifugation, the supernatant plasma + and sedimented red cells were expressed by the Baxter cells plated and percentage CD34 were assessed and Optipress system (Baxter Health Care) into an upper and correlated. The results are expressed as total colonies in lower plasma bag and red cell bag, respectively, leaving cord blood cells harvested based on the mononuclear the buffy layer in the original collection bag. The WBC, cell count. CD34 and progenitor estimations were repeated on the buffy coat fraction. A full blood count was performed on the red cell fraction and an aerobic and anaerobic bacteri- Bacteriology ology screen was performed on the buffy coat fraction. Erythrocyte, leucocyte and platelet counts were perfor- 0.5 ml from each cord blood collection and from the buffy med using an automated cell counter (Coulter STKS, Coul- coat fraction after processing triple bag donations was ter Electronics Ltd, Luton, UK). In order to correct for the transferred to aerobic and anaerobic blood culture bottles presence in the cord blood of nucleated red cells, a manual Bactec 6A, and 7A respectively (Becton Dickinson). These white cell differential count was made on all samples. The bottles were then incubated at 37°C for 7 days. Positive results were expressed as corrected total nucleated cell cultures identified by measurement of CO2 production were Cord blood collection and separation JA Ademokun et al 1025 then gram stained and plated onto agar plates to confirm Triple bag separation of whole cord blood bacterial or fungal identity. The cord blood volume collected and used for the triple bag separation (n = 25) ranged from 44 ml to 133 ml (mean 81 ± 4.29 ml). The volume of cord blood collected, buffy Statistics coat volume recovered, along with numbers of nucleated cells and CD34-positive cells is shown in Table 1. The Results are expressed as mean ± standard error (s.e.). The initial nucleated cell counts ranged from 3.89 × 108 to correlation between the volume of cord collected and the 15.68 × 108 (mean 10.53 ± 0.62 × 108) per pack. The mean nucleated cell count was analyzed by linear regression nucleated cell count obtained from the buffy fraction was analysis. ␹2 analysis was used to analyse low volumes of 9.44 ± 0.61 × 108 range (3.1–16.02 × 108) with a mean cord collected and degree of bacterial contamination. nucleated cell recovery of 90% with the poorest recovery being 62%. The initial lymphocyte count obtained was 3.6 ± 0.82 × 108 (range 0.84–5.39 × 108) with a mean lymphocyte count on the buffy coat fraction of ± × 8 × 8 Results 3.21 0.73 10 (range 0.79–5.72 10 ; mean lymphocyte recovery 89%) (data not shown). The number of CD34-positive cells collected per unit was Cord blood collection assessment of sterility, and 3.04 ± 0.39 × 106 (range 0.57–6.69 × 106). The CD34-posi- nucleated cell counts tive cell count obtained from the buffy fraction ranged from The mean volume of cord blood collected from 201 0.55 × 106 to 7.9 × 106 (mean 3.05 ± 0.43 × 106) recovery donations was 56.6 ± 2 ml (range 20–187 ml). During this 100% (Table 1). period when midwives were being trained in collection The red cell fraction was assessed for WBC content and techniques, 10% of all samples were Ͻ30 ml. was shown to have 1% or less of the original total nucleated The mean total nucleated cell counts of all samples was cell count. The final separated buffy coat volume ranged 8.12 ± 0.34 × 108 and this result correlated with the volume from 28 to 60 ml (mean = 44 ml). collected (Figure 1) (r = 0.81, P = 0.001). The incidence of Nineteen cord blood units were assessed for progenitor bacterial contamination was 24 in 201 (11.9%) samples cell content before and after separation (Table 2). The mean tested which mainly consisted of skin flora; Staphylococcus total progenitor content (CFU-GM, BFU-e and CFU- albus (n = 5), Streptococci (n = 6), Bacteroides (n = 4), GEMM) based on the mononuclear cell count was Diphtheroid (n = 4), Klebsiella (n = 1), Pseudomonas 8.85 × 105 per collection before separation and 7.81 × 105 (n = 1), E. coli (n = 1) and Bacillus (n = 2). Although there after separation. The initial mean total CFU-GM count was was a trend towards low volumes of cord blood collected 3.31 ± 0.60 × 105 per collection based on the total mono- being contaminated, using ␹2 analysis there was no signifi- nuclear cell count. After separation the mean CFU-GM cant correlation (P = 0.17) between low volumes of cord from the buffy coat fraction was 2.82 ± 0.64 × 105 with a collected (Ͻ40 ml) and bacterial contamination (data not mean recovery of 85%. The lowest total CFU-GM content shown). recovered was 0.9 × 105 and the highest 12.3 × 105. Twenty-five samples were collected using the triple bag The initial mean BFU-e content was 4.48 ± 0.85 × 105 method and 1/25 (4%) demonstrated bacterial contami- per collection and post-separation the BFU-e content was nation. 4.09 ± 0.93 × 105 per collection; recovery 91.3%. Similarly CFU-GEMM recoveries following the separation procedure were excellent (pre 1.06 ± 0.34 × 105; post 0.9 ± 0.40 × 105; recovery 84.9%). In a parallel study using a MNC fraction following separation of whole cord blood (n = 17) over lymphoprep 40 (Nycomed) we assessed the units for total progenitor cell 8 content and percentage of CD34-positive cells. This analy- 10 35 × sis demonstrated a correlation of coefficient of 0.63 30 (P = 0.01) Figure 2. 25 20 15 10 Discussion 5 Nucleated cell count The increasing use of umbilical cord blood for bone mar- 0 0 20 40 60 80 100 120 140 160 180 200 row reconstitution after myeloablative chemotherapy means Cord blood volume (ml) that banks with large numbers of cryopreserved cord blood donations need to be established to meet increasing Figure 1 Relationship between cord blood volume and total nucleated 2,13–15 cell counts. The total nucleated cell count was correlated with volume demand. Our aim was to assess collection of cord collected in 201 donations. A positive correlation of r = 0.81 was found blood and volume reduction while maintaining sufficient between volume of sample and nucleated cell count (P = 0.001). progenitor cells to enable engraftment. A system was Cord blood collection and separation JA Ademokun et al 1026 Table 1 Cell counts in whole cord blood and buffy coat following triple bag separation

Cord blood units Initial volume of cord Separated buffy coat Initial Buffy coat Initial CD34 × 106b Buffy coat n = 25 blood collected volume NCC × 108a NCC × 108a CD34 × 106b

1 125 40 11.34 12.08 6.12 7.90 2 90 37 10.50 9.36 5.80 5.90 3 133 47 12.28 9.40 2.20 2.63 4 98 52 12.88 11.54 2.44 1.92 5 104 36 9.16 8.67 2.29 1.70 6 77 60 11.96 8.80 5.80 2.80 7 84 41 12.17 11.07 1.46 1.32 8 66 37 6.64 6.60 1.06 2.20 9 45 28 4.50 3.10 1.08 1.10 10 56 36 8.02 7.80 0.80 1.00 11 94 47 9.68 6.39 5.22 3.38 12 97 35 12.09 11.48 2.29 2.41 13 60 34 10.42 9.45 2.60 6.30 14 69 43 13.61 11.48 2.90 6.31 15 81 51 10.70 9.40 5.30 1.70 16 62 48 11.55 11.56 2.88 1.15 17 85 52 8.20 6.70 1.47 1.87 18 79 51 12.97 11.83 2.07 2.36 19 68 49 6.64 5.78 4.60 4.33 20 66 42 7.20 6.17 0.57 0.74 21 80 49 12.20 11.22 3.90 3.47 22 44 35 3.89 4.29 0.77 0.55 23 91 49 15.68 13.52 4.80 5.90 24 91 47 15.58 16.02 6.69 6.40 25 85 49 13.30 12.30 0.93 0.86 Mean ± s.e. 81 ± 4.29 44 ± 1.5 10.53 ± 0.62 9.44 ± 0.61 3.04 ± 0.39 3.05 ± 0.43

aNCC = total nucleated cell counts × 108 in each cord blood unit. bCD34 × 106 = total CD34-positive cells in each cord blood unit.

Table 2 Progenitor cell content in whole cord blood and buffy coat 600 following triple bag separation

MNC 500 5 Whole cord blood Buffy coat Recovery % 400

CFU-GM × 105 (n = 19) 3.31 ± 0.60 2.82 ± 0.64 85.2 300 BFU-e × 105 (n = 19) 4.48 ± 0.85 4.09 ± 0.93 91.3 CFU-GEMM × 105 (n = 19) 1.06 ± 0.34 0.9 ± 0.42 84.9 200 Total progenitors 8.85 ± 1.51 7.81 ± 1.7 88.2 100

Total progenitors per 10 Total 0 0 0.5 1 1.5 2 2.5 3.0 3.5 assessed that reduced the risk of microbial contamination %CD34-positive cells in MNC fraction during processing. We evaluated the total nucleated cell Figure 2 Relationship between cord blood mononuclear cell CD34- counts, colony-forming units (CFU-GM) and absolute positive % and total progenitor cell numbers. The percentage of CD34- CD34-positive cell numbers before and after separation and positive cells in the mononuclear cell fraction of 17 cord blood samples volume reduction. was correlated with total progenitor cell numbers. A positive correlation = = The mean volume collected using the triple bag collec- of r 0.63 was found (P 0.01). tion system (81 ± 4.2 ml) was similar to that recently reported by others using a single collection bag.2,16 The collected.18 Our studies observed a correlation between nucleated cell counts (10.53 ± 0.63 × 108) and CD34-posi- CFU-GM and CD34-positive cell numbers which was tive cell counts (3.04 ± 0.39 × 106) were also similar to within the range reported by others for CB,19 but not as those reported from other centres worldwide.4,8,16,17 The high as one original paper by Fritsch et al20 which may be nucleated cell counts significantly correlated with the vol- due to differences in culture techniques. Kurtzberg et al2 ume of cord blood collected, and this result is in agreement have recently described results using cord blood for trans- with other studies suggesting that the CFU-GM and CD34- plantation into unrelated recipients and stated that the num- positive cell counts also correlate positively with volume ber of nucleated cells transfused per kilogram body weight Cord blood collection and separation JA Ademokun et al 1027 correlated with the rate of myeloid engraftment (P = 0.002) coat. We aim to use this triple bag system for collection of and that the nucleated cell dose may be a more important cord blood for banking in our region. indicator of engraftment than CD34-positive cell or CFU- GM numbers. Our results, which show similar nucleated cell recovery to that obtained from bone marrow processed by centrifug- Acknowledgements ation,21 demonstrate that cord blood can be separated without undue loss of nucleated cells by this method. Recovery We are indebted to the Department of Obstetrics and Gynaecology of up to 85% nucleated cells and 80% mononuclear cells and midwifery staff at the Royal Victoria Infirmary for their col- have been reported when bone marrow and peripheral laboration with this project and to Mrs M Graham for excellent blood stem cell harvests are processed this way.21 Apart secretarial assistance. We wish to thank the Tyneside Leukaemia from good cell recovery an important feature of this method Research Association and the Hayward Foundation for support. of cord blood separation is the ability to use a closed system for processing with by-products of the process, References namely red cells and plasma being available for ABO grouping, and with DMSO for cryopreservation respect- 1 Wagner JE, Kernan NA, Steinbuch M et al. Allogeneic sibling ively. umbilical-cord-blood transplantation in children with malig- The method of recovery of nucleated cells also conforms nant and non-malignant disease. Lancet 1995; 346: 214–219. to the guidelines recommended for collection, processing 2 Kurtzberg J, Laughlin M, Graham ML et al. Placental blood and storage of human bone marrow and peripheral blood as a source of hematopoietic stem cells for transplantation into stem cell transplantation.22 Volume reduction is important unrelated recipients. New Engl J Med 1996; 335: 157–166. to increase storage capacity to a cost effective level.17 The 3 Howard MR, Gore SM, Hows JM et al. A prospective study mean buffy volume of 44 ml would result in a final storage of factors determining the outcome of unrelated marrow donor volume of 84 ml if 20% DMSO was used as cryoprotectant searches: report from the International Marrow Unrelated Search or 52 ml if 50% DMSO was used as has been recently sug- and Transplant Study Working Group on behalf of collaborating gested by Rubinstein et al.17 centres. Bone Marrow Transplant 1994; 13: 389–396. 4 Broxmeyer HE, Douglas GW, Hangoc G. Human umbilical ± × 8 The mean nucleated cell count of 9.44 0.61 10 cord blood as a potential source of transplantable obtained in the buffy fraction is within the range used for stem/progenitor cells. Proc Natl Acad Sci USA 1989; 86: 1,23 2 successful engraftment in both related and unrelated 3828–3832. cord blood transplantation. Several centres have used separ- 5 Rubinstein P, Rosenfield RE, Adamson JW, Stevens CE. ation methods for cord blood prior to cryopreservation, sub- Stored placental blood for unrelated bone marrow reconsti- sequent thawing and transplantation.1,2,17 Good recoveries tution. Blood 1993; 81: 1679–1690. of progenitor cells as measured by colony-forming assays 6 Charboro P, Newton I, Schaal JP, Herve P. The separation of have been obtained following hydroxyethyl starch sedimen- human cord blood by density gradient does not induce a major tation,17 3% gelatin sedimentation9 or percoll density gradi- loss of progenitor cells. Bone Marrow Transplant 1992; 9: ents.6 The majority of these manipulations, except for the 109–110. 7 Harris DT, Schumacher MJ, Rychlik S et al. Collection, separ- recent processing procedures described by Rubinstein et 17 ation and cryopreservation of umbilical cord blood for use in al, do not involve a closed bag system of separation. The transplantation. Bone Marrow Transplant 1994; 13: 135–143. risk of bacterial contamination during processing can be 8 Falkenburg JHF, Van Luxemberg-Heijs SAP, Zijlmans JM et reduced using the triple bag system described and our al. Separation, enrichment, and characterization of human results show that this method allows good nucleated cell hematopoietic progenitor cells from umbilical cord blood. Ann and CD34 recovery despite the small volume of cord blood Haematol 1993; 67; 231–236. collections. Initial studies demonstrate that even larger vol- 9 Nagler A, Peacock M, Tantoco M et al. Separation of hemato- umes (Ͼ100 ml) of cord blood can be separated into much poietic progenitor cells from human umbilical cord blood. J smaller units which could be cryopreserved into one bag Hematother 1993; 2: 243–245. unit for storage and will allow for increased efficiency of 10 Pahwa RN, Fleischer A, Than S, Good RA. Successful haematopoietic reconstitution with transplantation of erythrocyte cord blood processing, essential for cord blood banking on depleted allogeneic human umbilical cord blood cells in a a large scale. A large number of samples can be processed child with leukemia. Proc Natl Acad Sci USA 1994; 10: without significant modification of normal blood banking 4485–4488. procedures. 11 Ho¨gman CF, Erikjsson L, Hedlund K, Wallvik J. The bottom The incidence of microbial contamination of harvested and top system: a new technique for blood component prep- bone marrow and peripheral blood has recently been aration and storage. Vox Sang 1988; 55: 211–217. reported.24 A total of 85/3910 samples sent for culture grew 12 Sutherland RD, Keating A, Nayar R et al. Sensitive detection micro-organisms (2.2%); the predominant organisms and enumeration of CD34+ cells in peripheral and cord blood detected were skin flora (89%). Using the triple bag method by flow cytometry. Exp Hematol 1994; 22: 1003–1010. of collection our current incidence of bacterial contami- 13 Gluckman E, Broxmeyer HE, Auerbach AD et al. Haematopo- ietic reconstitution in a patient with Fanconi’s anaemia by nation is 3–4% which compares favourably with other 18 means of umbilical cord blood from an HLA identical sibling. centres. New Engl J Med 1989; 321: 1174–1178. In conclusion, we are continuing to assess this system of 14 Wagner JE, Kernan NA, Broxmeyer HE. Transplantation of cord blood separation and further work is in progress on umbilical cord blood in 50 patients: analysis of the registry the cryopreservation thawing and washing17 of the buffy data. Blood 1994; 84: 395a (Abstr.). Cord blood collection and separation JA Ademokun et al 1028 15 Kurtzberg J, Laughlin M, Casey J. Umbilical cord blood: an 20 Fritsch G, Buchinger P, Printz D. Use of flow cytometric alternative source of hemopoietic stem cells for bone marrow CD34 analysis to quantify hematopoietic progenitor cells. reconstitution in unrelated donor transplantation. Exp Hematol Leuk Lymphoma 1993; 10: 443–451. 1995; 23: 709 (Abstr.). 21 Rowley SD. Hematopoietic stem cell processing and cryopre- 16 Kogler G, Callejas J, Hakenberg P et al. Hematopoietic trans- servation. J Clin Apheresis 1992; 7: 132–134. plant potential of unrelated cord blood: critical issues. J Hema- 22 Voak D, Cann R, Finney RD et al. Guidelines for the collec- tother 1996; 5: 105–116. tion, processing and storage of human bone marrow and per- 17 Rubinstein P, Dobrila L, Rosenfield RE et al. Processing and ipheral stem cells for transplantation. Transfusion Med 1994; cryopreservation of placental/umbilical cord blood for unre- 4: 165–172. lated marrow reconstitution. Proc Natl Acad Sci USA 1995; 23 Wagner JE, Broxmeyer HE, Byrd RL et al. Transplantation 92: 10119–10122. of umbilical cord blood after myeloablative therapy: analysis 18 Bertolini F, Lazzari L, Lauri E et al. Comparative study of of engraftment. Blood 1992; 79: 1874–1881. different procedures for the collection and banking of umbili- 24 Prince HM, Page SR, Keating A et al. Microbial contami- cal cord blood. J Hematother 1995; 4: 29–36. nation of harvested bone marrow and peripheral blood. Bone 19 Kinniburgh D, Russell NH. Comparative study of CD34-posi- Marrow Transplant 1995; 15: 87–91. tive cells and subpopulations in human umbilical cord blood and bone marrow. Bone Marrow Transplant 1993; 12: 489– 494.