Technical Reports A randomized phase III clinical trial of ... - Nature

Bone Marrow Transplantation (2003) 31, 1043–1051 & 2003 Nature Publishing Group All rights reserved 0268-3369/03 $25.00

www.nature.com/bmt

Technical Reports A randomized phase III clinical trial of autologous blood stem cell transplantation comparing cryopreservation using dimethylsulfoxide vs dimethylsulfoxide with hydroxyethylstarch SD Rowley, Z Feng, L Chen, L Holmberg, S Heimfeld, B MacLeod and WI Bensinger Clinical Research Division, Fred Hutchinson Cancer Research Center Seattle, WA, USA

Summary: Hematopoietic stem cells intended for autologous transplantation are usually cryopreserved in solutions containing 10% dimethylsulfoxide (DMSO, v/v) or 5% DMSO in combination with 6% hydroxyethylstarch (HES, w/v). We performed a single-blinded, randomized study comparing these cryoprotectant solutions for patients undergoing autologous peripheral blood stem cell (PBSC) transplantation. A total of 294 patients were evaluable; 148 received cells frozen with 10% DMSO and 146 received cells frozen in 5% DMSO/6% HES. Patients who received cells frozen with the combination cryoprotectant recovered their white blood cell count X1.0 109/l at a median of 10 days, one day faster than those who received PBSC frozen with DMSO alone (P ¼ 0.04). Time to achieve neutrophil counts of X0.5 109 and X1.0 109/l were similarly faster for the recipients of the cells frozen in the combination solution. This effect was more pronounced for patients who received quantities of CD34+ cells higher than the median for the population. Median time to discontinuation of antibiotic use was also one day faster for the recipients of cells cryopreserved with DMSO/HES (P ¼ 0.04). In contrast, median times to recovery of platelet count X20 109/l were equivalent for each group (10 days; P ¼ 0.99) and the median numbers of red cell and platelet transfusions did not differ. Bone Marrow Transplantation (2003) 31, 1043–1051. doi:10.1038/sj.bmt.1704030 Keywords: cryopreservation; dimethylsulfoxide; hematopoietic stem cell transplantation; hydroxyethylstarch

Cryopreservation allows for the long-term storage of hematopoietic stem cells (HSC) and is the preferred storage technique virtually for all components intended for autologous HSC transplantation. Although some unavoidable and undefined loss of HSC probably occurs with marrow or peripheral blood stem cell (PBSC) processing and cryopreservation, progressive loss over months to years

Correspondence: Dr SD Rowley, Hackensack University Medical Center, 20 Prospect Avenue, Suite 400, Hackensack, NJ 07601, USA Received 8 August 2002; accepted 31 December 2002

of proper storage is not obvious and may not occur if optimal storage conditions are followed. Cryopreservation allows the administration of multiple-day transplant conditioning regimens as well as elective storage for patients to be transplanted at a subsequent point in a course of treatment. That HSC can be successfully cryopreserved is evident from the rapid recovery of marrow function after marrowlethal conditioning regimens, including those used for allogeneic transplantation.1 Engraftment failure is generally not attributed to HSC cryopreservation. Most cryopreservation laboratories use a variation of the technique involving addition of a penetrating cryoprotectant such as dimethylsulfoxide (DMSO) and a source of plasma protein, cooling at a defined rate, and storage at cryogenic temperatures using mechanical or liquid nitrogen refrigerators.2,3 Variations include the final concentration of DMSO, the source and concentration of plasma protein, and the equipment used to cool and store the component. Some laboratories add hydroxyethylstarch (HES), an extracellular (nonpenetrating) cryoprotectant. DMSO is a colligative cryoprotectant that prevents dehydration injury to the cell by moderating the increasing concentration of nonpenetrating extracellular solutes during ice formation, and by decreasing the amount of water absorbed by the ice crystals at a defined temperature.4 Colligative effects depend upon the number of molecules, which is why most cryopreservation laboratories use high concentrations of DMSO (10% v/v contributing 1400 mOsm). Colligative properties do not explain the cryoprotection achieved by freezing cells in solutions of macromolecules such as HES. Solutions of high-molecular weight polymeric cryoprotectants contain relatively few particles and, moreover, do not freely penetrate the cell. These cryoprotectants may protect the cell by forming a viscous, glassy shell that retards the movement of water, thereby preventing progressive dehydration as water is incorporated into the growing extracellular ice crystals.5 HES is a polymeric substance containing chains of different molecular weights. Initially explored as a cryoprotectant for red blood cells, HES was also found to be an effective cryoprotectant for a variety of other cells.5,6 Macromolecular cryoprotectants may be used as single agents, but are generally used in combination with penetrating cryoprotectants. In one early study, the addition of polyvinyl-

Cryopreservation of blood stem cell SD Rowley et al

1044

pyrrolidone to glycerol or DMSO improved the cryopreservation of murine cells compared to the use of a penetrating agent alone.7 Stiff et al 8 froze human cells in a combination of 5% DMSO, 6% HES, and 4% human serum albumin, and reported improved progenitor cell survival (compared to 10% DMSO without serum protein) as determined using in vitro cultures. Those investigators subsequently successfully used this mixture to cryopreserve the marrow of 60 patients.9 Others have adopted this technique for the cryopreservation of PBSC.10 Only one clinical study comparing the cryopreservation of HSC with DMSO/HES or DMSO alone has been reported.11 That trial, limited to 12 patients in each arm, did not detect any difference in the kinetics of engraftment. The small number of patients enrolled, however, limited the power of the study to detect small differences in engraftment speed or transplant outcome. To better address this issue, we conducted a single-blinded, randomized study comparing the speed of hematological recovery and other secondary transplant outcomes for patients undergoing autologous PBSC transplantation using cells cryopreserved with DMSO and HES or with DMSO alone. The intent of this study was to determine if the use of DMSO and HES in combination could achieve more rapid recovery of peripheral blood cell counts after transplantation than the use of DMSO as the sole cryoprotectant.

of CD34+ cells was achieved (generally, 45 106/kg patient weight, although the physician caring for the patient could prescribe collection of higher numbers) or mobilization was deemed a failure. Patients with limited numbers of cells collected after one or two apheresis procedures were managed with large volume apheresis procedures or with a temporary postponement of apheresis until adequate numbers of CD34+ cells were present in the peripheral blood. Two or more attempts at mobilization and collection of PBSC were possible and all cells were frozen using the same technique to which the patient was originally assigned. PBSC were collected using a COBE Spectra (COBE BCT, Inc., Engelwood, CO, USA) as previously described.13 The peripheral circuit was anticoagulated with a mixture of 5000 units of heparin in 500 ml ACD-A infused at a blood to anticoagulant ratio of 30 : 1. An additional 40 ml of this anticoagulant solution was placed in the collect bag at the start of the procedure. For most patients, 12 l of blood was processed during each apheresis procedure. For pediatric patients, the volume of blood processed was 3 the patient’s blood volume not to exceed 12 l. The processing of up to 6 the patient’s blood volume (large-volume apheresis, calculated from apheresis device nomograms based on patient height, weight and gender) could be prescribed by the attending physician for an individual patient.13

Methods

Cryopreservation solutions

Patients

Patients were randomly assigned to the use of one of two different cryopreservation solutions. For patients assigned to 10% DMSO, a fresh 2 cryoprotectant solution containing 20% (v/v) of pharmaceutical grade DMSO (Cryoserv, Research Industries, Salt Lake City, USA) and 8% (w/v) human serum albumin (HSA, Bayer Corp., West Haven, CT, USA) in Normosol-R, pH 7.4 (Abbott Laboratories, Abbott Park, IL, USA) was made each day of collection. For patients assigned to the combination of DMSO and HES, a fresh 2 cryoprotectant solution containing 10% (v/v) DMSO, 12% (w/v) HES, and 8% (w/v) human serum albumin (HSA) in Normosol-R, pH 7.4 was made each day of collection. Batches of a stock solution of 20% (w/v) HES solution were made by dissolution of HES powder (McGaw, Inc., Irvine, CA, USA) in Normosol-R, pH 7.4 using a volumetric flask to adjust for volume expansion with hydration of the HES powder and ensure a final stock concentration of 20%. The HES solution was packaged in 50 ml single-use vials and steam autoclaved. Each lot (10% of the bottles filled) was tested for sterility and pyrogen content before use.

This was a prospective, single-blinded, randomized study of two cryopreservation solutions. Patients were eligible for this study if PBSC were being collected for autologous transplantation and if the components were not to be processed except for cryopreservation as described below. No restrictions were placed on patient age, diagnosis, prior therapy, PBSC mobilization regimens, or transplant conditioning regimens. Patients eligible were undergoing their first dose-intense regimen with autologous PBSC rescue. This research study was approved by the Institutional Review Board of the Fred Hutchinson Cancer Research Center and written informed consent was obtained for all patients before enrollment and randomization.

Collection of PBSC by apheresis PBSC mobilization techniques were not prescribed by protocol and these patients were treated with a variety of chemotherapy and/or cytokine regimens. No attempt was made to stratify patients based on the PBSC mobilization regimen used. For patients treated with chemotherapy plus a hematopoietic cytokine, apheresis was generally initiated when the peripheral blood white cell count exceeded 1 109/l. For some patients (as prescribed by the individual patient’s physician), apheresis was initiated based on the level of CD34+ cells in the peripheral blood.12 For patients treated with a hematopoietic cytokine alone, apheresis was generally initiated on the fourth day of cytokine administration. Apheresis was continued daily until the target goal Bone Marrow Transplantation

Cryopreservation, storage, and reinfusion The volume of the PBSC component was reduced by centrifugation at 2300 g.14 Excess supernatant was expressed, the volume of the residual cell pellet adjusted with autologous plasma as necessary, and the cell pellet diluted with an equal volume of the designated 2 cryoprotectant solution. The cells were frozen in 30–70 ml aliquots in 250 ml freeze bags (Cryocyte Freezing Container;


Baxter Healthcare Corp, Round Lake, IL, USA). The first collection for each patient was frozen in at least two bags. Subsequent collections could be frozen in single bags without regard to the concentration of nucleated cells or platelets.14 Cells were cooled at 11C/min using a ratecontrolled cooling device (Cryomed, New Baltimore, MD, USA) to 401C with compensation for the heat of fusion, and then at 101C/min to 801C before storage in the vapor phase of nitrogen at 1861C or colder.15 Most components were cryopreserved within 4 h of collection; cells arriving late in the laboratory could be held at 41C overnight in a monitored refrigerator before concentration and cryopreservation. Cryopreserved cells were infused at least 24 h after the administration of the last dose of chemotherapy. All patients were hydrated for 2 h before and after infusion (total, 4 h) with D5W1/2NS containing 20 meq/l KCl at a rate of 125 ml/m2/h. Patients were medicated before infusion with diphenhydramine (or other antihistamine) and hydrocortisone. Administration of antiemetics before or during the cell infusion was left to the discretion of the staff caring for the patient. Cells were transported to the patient’s bedside in liquid nitrogen carriers and thawed by immersion in a water bath at 371C until complete dissolution of ice crystals. The cells were then infused at a rate of 10–20 ml/min through a platelet product administration set with an inline 170–210 mm filter into a patent central venous catheter.

Analysis of PBSC components CD34+ cell counts were obtained as previously described.21 In brief, a sample of 1 106 nucleated cells were stained with phycoerythrin (PE)-conjugated anti-CD34 stain (HPCA-2, Becton Dickinson, San Jose, CA, USA) and fluorescein isothiocyanate (FITC)-conjugated antibody to CD14 (Leu-M3, Becton Dickinson) added at the concentrations recommended by the manufacturer after lysis of red blood cells using ammonium chloride. Control samples consisted of the same cell quantity incubated with PE-conjugated IgG1 (Becton Dickinson) in conjunction

Table 1

1045

Chemotherapy regimens and post-transplant care These patients were treated with a variety of transplant conditioning regimens (28 different) appropriate to the disease being treated (Table 1). The seven most common regimens used were busulfan, 12 mg/kg in 12 doses over 3 days, followed by melphalan (Mel), 100 mg/m2 in two doses over 2 days, and then by thiotepa (TT), 500 mg/m2 in two doses over 2 days (BuMelTT, n ¼ 162);16 cyclophosphamide (Cy) 100 mg/kg, etoposide (VP16) 60 mg/kg, and 12 Gy total body irradiation (TBIVP16Cy, n ¼ 34);17 Mel 140 mg/m2 and TT 500 mg/m2 (MelTT, n ¼ 18); Mel 200 mg/m2 (Mel, n ¼ 16); Bu 16 mg/kg in 16 divided doses over 4 days (Bu, n ¼ 9); 131I-labeled CD19 antibody, VP16, Cy (I131VP16Cy, n ¼ 7);18 and Bu 16 mg/kg in 16 divided

Patient characteristics

Number of patients Age (years) Gender (M : F) Diagnosis Breast cancer Non-Hodgkin’s lymphoma Multiple myeloma Acute myelogenous leukemia Hodgkin’s disease Other Number of prior regimens Prior radiotherapy (Y : N) Number of components collected Number of components per patient CD34+ cells frozen ( 106/kg) a

with the CD14 stain. Viability testing using exclusion of 7aminoactinomycin D (7-AAD, Sigma Chemical, St Louis, MO, USA) was performed for all components stored overnight. The cells were analyzed using a FACScan flow cytometer (Becton Dickinson) and 100 000 events acquired. Listmode data were analyzed using Winlist software (Verity Software House, Inc., Topsham, ME, USA). ‘Bright’ CD34+ cells were defined with histogram analysis using simultaneous gates first established on nucleated cells (defined by forward and orthogonal light scatter characteristics), viability (defined by lack of 7-AAD staining), and CD14-negative cells (defined by low orthogonal light scatter and lack of CD14 staining). Initially, the percentage of viable CD34+ cells was calculated after subtraction of the number of cells showing nonspecific staining in the isotype control sample. Use of these control samples was subsequently eliminated after validation of the procedure showed this step did not improve the precision of CD34+ cell enumeration. The absolute number of CD34+ cells was obtained by multiplying the percentage of CD34+ cells by the cell count and volume of the specimen. Cell counts for the PBSC components were obtained using a Sysmex NE2500 (Sysmex Corporation of America, Long Grove, IL, USA). White cell differential counts were performed manually using Wright-Giemsa-stained specimens.

DMSO

DMSO/HES

148 48.0 (2–67)a 56 : 92

146 47.0 (1–65)a 61 : 85

53 37 19 5 8 26 2 (0–4)a 44 : 104 416 2 (1–13)a 10.4 (3.6–140.4)a

51 37 18 6 8 26 2 (0–9)a 43 : 103 422 2 (1–12)a 10.2 (2.9–111.2)a

P-value 0.35 0.57 1.0

0.78 0.94 0.67 0.86

Indicates median (range).

Bone Marrow Transplantation


1046

doses over 4 days and Cy 120 mg/kg i.v. over 2 days (BuCy, n ¼ 6). The remainder of the patients received a variety of dose-intense, myeloablative chemotherapy or radiationbased regimens appropriate to the disease being treated. All patients received routine post-transplant care including blood component support in accordance with hospitalspecific policies. Patients underwent transplantation at any of seven different hospitals in Seattle. Filgrastim was routinely administered after transplantation at one hospital, but was otherwise discouraged.

Statistical analysis The intent of this study was to demonstrate the difference on engraftment kinetics for two different cryopreservation solutions. We proposed that a 2 day difference in time to platelet recovery would be considered significant; based on the engraftment experience at the time of study design with previous patients who received HSC frozen in DMSO alone, for a two-sided test with 0.10 statistical significance and 95% power to detect this difference, an estimated sample size of 150 per arm was determined. A full-interim analysis upon enrollment of 150 evaluable patients was added to the study design with the expectation that the study would be terminated if the interim results indicated very significant difference in engraftment kinetics, namely if a two-sided P-value for a test of the null hypothesis was o0.017. The interim analysis determined that, even with full patient accrual, a statistically significant difference in the kinetics of platelet recovery would not be found. The interim analysis for granulocyte recovery demonstrated a significant difference, but initial study design did not set a level of significance for early stopping of the study based on granulocyte engraftment. Therefore, the study was allowed to proceed to completion. Patients were randomized to have cells frozen with one or the other cryoprotectant solution at the time of enrollment and before cryopreservation of the first PBSC product. Patient enrollment was stratified based on diagnosis (breast cancer, non-Hodgkin’s lymphoma, Hodgkin’s Disease, multiple myeloma, acute myelogenous leukemia, vs other diagnoses), number of previous chemotherapy regimens (0–2 vs 42, excluding the PBSC mobilization regimen or use of corticosteroids alone), and previous treatment with involved field radiotherapy (yes vs no). The intent of stratification was to achieve parity for the number of CD34+ cells collected for each group of patients using characteristics previously described to predict the quantity of CD34+ cells collected.19 Patients, apheresis center personnel, and medical staff caring for the patient were blinded to the cryoprotectant used. Primary end points of engraftment times were defined as the first of three sequential days on which the absolute neutrophil count (ANC, defined as polymorphonuclear cells and band forms) exceeded 0.5 109/l and the first of 7 sequential days without platelet transfusion that the platelet count exceeded 20 109/l. Shorter periods (2–6 days) of observation could also be used for the definition of engraftment, but only if the target count doubled during the period of time of observation. Patients for whom data meeting the above criteria were lacking were censored as Bone Marrow Transplantation

not reaching this end point on the day before the first day of possible recovery. Secondary engraftment end points included time to a total leukocyte count (WBC) exceeding 1.0 109/l, ANC exceeding 0.1 109 or 1.0 109/l, time to a platelet count exceeding 50 109/l, and time to last platelet transfusion. Secondary transplant outcomes included the numbers of days to hospital discharge and to discontinuation of broadspectrum antibiotics. Blood component usage between the day of infusion and day of discharge from the transplant service (including both inpatient and outpatient days of service) was quantified. Patient characteristics are summarized using medians and ranges unless otherwise described. Wilcoxon rank sum test was used to compare the location of continuous measurement between the two groups. w2 test was used to compare the distribution of categorical measurement between groups. For the time-toevent outcomes with censoring, weighted log rank test was used to compare the distribution of time to event between the two groups. The weight used was the Kaplan–Meier estimate and the test was equivalent to the Peto modification of Gehan–Wilcoxon test. The effect of continuous variable on time-to-event outcome was analyzed by Cox proportional hazards regression model. Cox regression was also used to analyze the effect of certain variable after adjusting for other variables. All tests were done at twosided 0.05 level. All statistical analyses were done using S-plus 2000 statistical analysis software.

Results Patient enrollment A total of 1306 components were cryopreserved for 436 patients. Of these, 352 patients met protocol criteria and are available for analysis. To date, 294 patients have been transplanted and are the subject of this report (Table 1); 11 patients underwent transplantation (three at hospitals not part of this study) but records of transplant course were not available for review; 47 patients have not yet undergone transplantation. Of the 84 patients ineligible for evaluation, 60 died before transplantation could be performed (many of these patients not eligible for analysis had inadequate numbers of cells collected for transplantation and did not undergo high-dose therapy), 11 received components that were processed before or after freezing in violation of the protocol, and 11 had cells cryopreserved using techniques that differed from those required for this study or received marrow cells in addition to PBSC. One patient withdrew consent before cryopreservation of PBSC and one was not eligible because of administration of a previous high-dose regimen with autologous PBSC transplantation. The distribution of the stratification factors (patient diagnosis, number of prior chemotherapy regimens, previous treatment with involved field radiotherapy) did not differ between the two study groups (Table 1). By study design, all patients whose transplant data were eligible for analysis had adequate numbers of CD34+ cells collected to be able to proceed to transplantation. The total number of CD34+ cells/kg collected and the number of apheresis


procedures required to achieve these amounts also did not differ for the two study groups (Table 1).

Transplantation Transplant conditioning regimens used are shown in Table 2. In general, the conditioning regimen used was determined by patient diagnosis and the distribution of regimens used did not differ for the two groups of patients studied. A small number of patients in each group received cytokine support during the period of aplasia and the proportions of patients who received post-transplant growth factors in each group were not different. The quantity of CD34+ cells infused differs from the quantity collected because some products could be reserved for future use if the target dose of CD34+ cells prescribed by the patient’s physician for transplantation was exceeded and the physician elected to maintain cells in storage for future use. Again, the median number of CD34+ cells, and the median number of nucleated cells (Table 2) infused did not differ for the two patient groups. Infusion-related toxicity experienced by these patients was generally mild and transient as has been reported by others (data not shown). Two patients who received components cryopreserved with DMSO alone experienced serious neurological toxicity (seizure, transient ischemic attack) as has been previously observed with the infusion of cryopreserved cells.20 None of the recipients who received components frozen in DMSO/HES experienced similar serious infusionrelated toxicities. None of the products frozen with either technique was lost to clumping after thawing.

Engraftment One patient undergoing treatment for a diagnosis of myelofibrosis whose cells were cryopreserved with DMSO/HES failed to achieve the end points of engraftment and received a second infusion of cells frozen in the same solution on day 22 of transplantation. This patient achieved an ANC of 0.1 109/l on day 28 after initial transplantation (data for this patient were, therefore, Table 2

censored on day 27), but expired on day 31 before achieving other engraftment end points. Although three patients died of transplant-related complications on days 9–10 before achieving any end point of engraftment, one more died after achieving granulocyte end points of engraftment but before achieving a platelet count X20 109/l, and two more before achieving a platelet count X50 109/l, the rest of the patients in both groups (including the nine others that died of transplant-related complications) experienced prompt granulocyte and platelet engraftment (Table 3) and no other patient in either group required infusion of additional cells for delayed, incomplete, or failed engraftment. Patients who received PBSC products frozen in DMSO/ HES achieved a statistically significant faster WBC recovery X1.0 109/l (a median of 1 day faster) than patients who received products frozen in DMSO alone (Table 3). Similarly, the median time to achieve an ANC X1.0 109/l showed a trend to 1 day faster. The time to achieve an ANC of X0.5 109/l was significantly different in favor of patients whose products were frozen with DMSO/HES (Figure 1a, Table 3), although the median time to achieve this end point of engraftment did not differ between the two groups. The faster engraftment also resulted in a decrease in the number of days broadspectrum antibiotics were administered. In contrast, no significant differences in any of the end points of platelet recovery (Figure 1b, and Table 3) were found. Furthermore, the numbers of units of red blood cells or platelets infused during the initial transplant course did not differ for these two groups of patients (Table 3). Patients who received more than the median number of CD34+ cells/kg were more likely to benefit from receiving cells cryopreserved in DMSO/HES (Table 4). No difference was found in median times to achieving end points of platelet engraftment for either recipients who received above or below the median number of CD34+ cells. The use of post-transplant cytokines was evenly distributed across the groups. A total of 79 patients (27%) received at least one dose of filgrastim before achieving WBC, granulocyte, or platelet

1047

Conditioning regimens and cell infusions

Conditioning regimen BuMelTT TBIVP16Cy MelTT Mel Bu I131VP16Cy BuCy Other Post-transplant cytokine use (Y : N) CD34+ cells infused ( 106/kg) TNC infused ( 108/kg) Volume of products infused

DMSO (n=148a)

DMSO/HES (n=146a)

78 17 11 10 5 1 3 23 37 : 111 6.4 (3.6–82.0)b 7.6 (0.9–59.8)b 109 (30–782)b

84 17 7 6 4 6 3 18 42 : 104 6.5 (1.9–55.6)b 7.3 (1.1–76.2)b 97 (30–843)b

P-value 0.48

0.43 0.91 0.69 0.80

a

Actual number of subjects used to summarize the data for each variable may differ from the total number of patients as a result of missing values in that variable. b Indicates median (range). See Methods for descriptions of conditioning regimens used. TNC=total nucleated cell count in the PBSC product.



1048 Table 3

Engraftment kinetics and transplant outcome DMSO (n=148)

DMSO/HES (n=146)

P-value

11 10 11 12

10 10 11 11

0.04 0.09 0.05 0.07

10 13 9 17 14

10 13 9 16 13

0.99 0.88 0.75 0.54 0.04

4 (0–60) 14 (0–202) 7

4 (0–20) 14 (0–276) 8

0.98 0.57 0.98

a

Granulocyte engraftment WBC X1.0 109/l ANC X0.1 109/l ANC X0.5 109/l ANC X1.0 109/l Platelet engraftmenta Platelet X20 109/l Platelet X50 109/l Last transfusion Time to initial hospital dischargea (excluding patients who died) Time to antibiotic stoppagea Transfusionsb Red blood cell (Units) Platelets (Units) Transplant-related mortalityb a

Median times to achieve the designated outcome for patients classified by cryoprotectant used. P-value is based on weighted log rank test, which compares the two estimated recovery curves, not the median time to achieve the endpoint of engraftment. b Median numbers of units of blood products transfused (range) and the actual number of patients who expired during the first 100 days after transplantation (P-value is based on w2 analysis).

(Table 5). However, in none of these analyses did the recipients of products frozen in DMSO alone achieve the end point of engraftment faster than the recipients of products frozen with DMSO/HES. As was previously described for the entire population, the kinetics of platelet recovery did not differ when these subpopulations were analyzed.

Discussion

(a) Probability of achieving an ANC count X0.5 109/l for patients stratified by cryopreservation technique. p ¼ 0.05 (Gehan–Wilcoxon). (b) Probability of achieving a platelet count X20 109/l for patients stratified by cryopreservation technique. p ¼ 0.99 (Gehan–Wilcoxon).

Figure 1

engraftment and were classified as receiving cytokine support. Cytokine administration was associated with faster WBC and granulocyte recovery but slower platelet recovery (Table 5). When separated into groups according to cryopreservation solution used that received or did not receive post-transplant filgrastim, the differences in granulocyte engraftment kinetics failed to reach significance Bone Marrow Transplantation

Reproducible cryopreservation of HSC permits the routine long-term storage of these cells for autologous or allogeneic use. The survival of cells through the steps of cryopreservation, storage, thawing, and infusion is reliably achieved despite differing sources of HSC (eg, PBSC vs marrow vs umbilical cord blood), diagnoses and prior medical treatment of patient or donor, and minor variations in techniques used to collect and process the cells. However, the techniques currently used by clinical HSC cryopreservation laboratories are not based on in vivo studies of HSC engraftment potential, and the optimal concentrations of cryoprotectants, plasma proteins, and other additives; and the optimal rates of cooling for these various cryopreservation solutions have not been defined in clinical studies using engraftment kinetics or success as an end point. Quantification of the losses of HSC, if any, that occur during the freezing and thawing of HSC have not been studied for human HSC products. Using a murine model of transplantation, van Putten21 studied the number of cells required to rescue 50% of mice from a dose of TBI and reported 52–66% losses of engraftment potential for murine bone marrow cells frozen in DMSO alone. The dose of cells frozen in DMSO required to rescue 50% of animals was 4.3 105; the dose required to rescue the same proportion of mice receiving fresh cells was 1.5 105. Similar damage to the engrafting potential of human HSC during cryopreservation and thawing may also occur. We performed a large randomized study of two cryopreservation solutions commonly used by HSC proces-


1049 Table 4

Engraftment kinetics for patients stratified by CD34+ cell dose Patients below the mediana

Outcome

Patients above the mediana

DMSO (n=75)

DMSO/HES (n=72)

P-valueb

DMSO (n=73)

DMSO/HES (n=74)

P-valueb

11 10 11 13

11 10 11 12

0.71 0.86 0.87 0.87

11 10 11 12

10 9 10 11

0.02 0.03 0.01 0.02

11 14 9

12 15 9

0.25 0.10 0.67

9 12 8

9 12 8

0.36 0.28 0.75

Granulocyte engraftment WBC X1.0 109/l ANC X0.1 109/l ANC X0.5 109/l ANC X1.0 109/l Platelet engraftment Platelet X20 109/l Platelet X50 109/l Last transfusion a

Median times to achieve the designated outcome for patients classified by cryoprotectant used and stratified by quartile of CD34+ cell dose/kg (first and second quartiles combined vs third and fourth quartiles combined). The median doses of CD34+ cells/kg infused are shown in Table 2. b P-value is based on weighted log rank test, which compares the two estimated recovery curves, not the median time to achieve the endpoint of engraftment.

Table 5

Effects of cytokine use on hematological recovery

Outcome

WBC X1.0 109/l ANC X0.5 109/l ANC X1.0 109/l Platelet X20 109/l Day to antibiotic stop

Effect of use for entire populationa

Patients treated without cytokinea

Patients treated with cytokinea

None (n=215)

Used (n=79)

P-value

DMSO (n=111b)

DMSO/HES (n=104b)

P-value

DMSO (n=35b)

DMSO/HES (n=42b)

P-value

11 12 13 10 14

10 9 10 11 13

o0.001 o0.001 o0.001 0.04 o0.001

11 12 13 10 14

11 11 13 10 14

0.18 0.18 0.22 0.78 0.15

10 9 10 11 15

9 9 10 11 12

0.24 0.51 0.62 0.94 0.24

a

Median number of days to achieve the designated outcome for patients classified by whether post-transplant cytokines were used and stratified by cryopreservation solution used. P-value is based on weighted log rank test, which compares the two estimated recovery curves, not the median time to achieve the end point of engraftment. b Actual number of subjects used to for each outcome may differ a bit because of some missing values.

sing laboratories. The differences between these solutions were in the concentration of DMSO and the presence or lack of HES. Both solutions used human serum albumin instead of the more variable autologous plasma used by many laboratories. We observed a small but statistically significant improvement in the kinetics of granulocyte recovery for recipients of HSC cryopreserved in the combination solution. A similar benefit was found for the time to discontinuation of broad-spectrum antibiotics, a variable expected to be closely associated with recovery of granulocyte counts. No difference in day of hospital discharge was found, but, at present, hospital stays are more likely to be prolonged because of mucositis or other nonhematopoietic toxicities resulting from the conditioning regimen than for management of neutropenia alone. This difference in time to granulocyte recovery was found for patients who received a dose of CD34+ cells/kg above the median for the population as a whole. That a difference in granulocyte engraftment was not found when patient populations were divided by the administration of filgrastim probably reflects a loss of power (adequate numbers of patients in each group) for statistical significance. The differences in granulocyte engraftment for the recipients of these two different cryopreservation techniques were not found for platelet engraftment, including when the population was stratified above or below the median dose of CD34+ cells/ kg. This divergence in result suggests a lineage-specific

effect upon granulocyte precursor cells. One possible explanation is that cells committed to the granulocyte lineage are more sensitive to possible toxic effects of the DMSO-only cryopreservation procedure than cells committed to platelet differentiation, and that the addition of HES provides better cryoprotection of these granulocyte precursors. Alternatively, it is possible that this is a membrane-specific effect and the combination cryoprotectant solution achieves better protection of plasma membrane proteins such as cytokine receptors required for the expansion and/or maturation of granulocyte populations. In support of this second hypothesis, Gilmore22 reported relative resistance of cryopreserved cells to giant cell tumor (GCT)-conditioned medium compared to other solutions used to stimulate the growth and maturation of cells in in vitro culture of myeloid progenitors (CFU-GM). We similarly reported relative resistant of CFU-GM to filgrastim compared to other cytokines after freezing and thawing, supporting this hypothesis of a selective damage to or loss of the filgrastim receptor or other membrane structure during the cryopreservation and thawing process.23 Thus, it is feasible that changes in the cryopreservation technique could have lineage-specific effects that would be detected in clinical outcomes. We did not perform subset analysis of the CD34+ cell populations for patients above and below the median of CD34+ cells, so we cannot determine if the differences in engraftment kinetics are a result of improved cryopreservation of CD34+ cells overall Bone Marrow Transplantation


1050

or of a specific CD34+ subpopulation or non-CD34+ accessory cells that might be found in products of patients who mobilize higher numbers of CD34+ cells. Although this study shows that both cryopreservation solutions achieve similar engraftment kinetics, the lineagespecific findings of this study encourage further exploration of cryoprotectants used for the freezing and storage of HSC. The cryoprotectant properties of glycerol were described in 1949 and those of DMSO 10 years later.24,25 The optimal concentration of either colligative cryoprotectant for the cryoprotection of HSC has not been determined in models of hematopoietic stem cell engraftment, although concentrations of DMSO as low as 5% have been used successfully for HCT.26 In in vitro models, the optimal concentration of DMSO used alone appears to be about 10%. Ragab et al 27 studied the survival of HSC from human donors after freezing with various concentrations of DMSO. They found a significant increase in progenitor cell recovery when the concentration of DMSO was increased from 7.5 to 10%. No further improvement was found with an increase to 12.5%. This improvement resulted from an increase in nucleated cell recovery from 17.2 to 35.6 and 32.1%, respectively, after washing in a preparation for cell culture. Cell recovery before the postthaw wash did not differ, nor did the numbers of colonies per 105 cells plated, and these findings could be an artefact of the laboratory testing. The loss of cells during the wash steps that was found in that study may or may not reflect events occurring during the direct intravenous infusion commonly used clinically. Limited numbers of in vitro studies have been performed for the combination of DMSO and HES. Stiff et al8 reported better recovery of myeloid colony cells for marrow samples frozen in DMSO with HES and HSA compared to cells frozen in DMSO without exogenous protein. Donaldson et al 28 studied the recovery of CD34+ cells cryopreserved in varying concentrations of DMSO and HES and reported an increase in CD34+ cell recovery from 12.2710.0% (mean7s.d.) to 85.4728.4% as the concentration of DMSO was increased from 2.5 to 5.0%, but no further improvement with further increase in concentration to 10% (HES concentration was kept constant at 4% (w/v)). Varying the concentration of HES in the presence of 5% DMSO had no effect on CD34+ cell recovery. Variation in cooling rates was not a component of this study. Different cells have different optimal cooling rates, and the optimal cooling rate is also dependent on the type and concentration of cryoprotectant used. In general, the higher the concentration of a colligative cryoprotectant, the slower the optimal cooling rate.29 In contrast, optimal rates for cooling when using extracellular cryoprotectants are generally more rapid, and the cells tolerate a broader range of cooling rates. The cooling rates in this study were rigidly controlled by use of a limited cryopreservation volume and a computer-controlled device to achieve a cooling rate of 11C/min. We cannot state that this rate of cooling is optimal for cells frozen in the DMSO/HES solution. Although some publications suggest that ‘uncontrolled’ freezing is appropriate or required for cryopreservation using HES,11 the cooling of an object is a physical process controlled by the volume of the object being cooled, the


heat transfer characteristics of the object, and the difference between the temperature of the object and its chamber.30,31 Therefore, simple immersion of the usual volume of HSC into a 801C refrigerator as initially reported by Stiff et al reproducibly achieves a cooling rate of about 1–31C/min and the mechanically controlled cooling technique used for this study does not negate the results of this study or the widespread adoption of these results (even if a different cooling velocity is found in a further study to be optimal for one or the other solution). In this large series of patients, we found that the administration of post-transplant cytokines and higher CD34+ cell dose infused also predicted faster granulocyte recovery. These observations, and the observation that post-transplant filgrastim could delay platelet recovery, have been reported by others.19,32,33 The clinical relevance of the small differences in engraftment kinetics found in this large randomized study of these two solutions as used for these patients is probably minimal, and the data presented here demonstrate that the combination of 5% DMSO and 6% HES as well as the use of 10% DMSO alone are both effective cryopreservation techniques. It is possible that the use of the combination solution may improve the kinetics of granulocyte engraftment and could decrease the cost of supportive care for the autologous PBSC recipient. This, plus the potential for decreased toxicity achieved by the 50% reduction in the amount of DMSO infused (DMSO infusion toxicity is dose related),34 supports the adoption of this combination cryoprotectant solution for clinical use, especially if commercially available HES solutions become available. More importantly than the validation of these two cryopreservation techniques, however, this study illustrates the potential benefit to be gained by further investigation of HSC cryopreservation techniques. The differences in granulocyte engraftment achieved by use of DMSO in combination with HES found in this study indicate the additional changes to HSC cryopreservation techniques may further improve engraftment kinetics after infusion of frozen HSC.

Acknowledgements Funding for this project was provided in part by Grants number CA18029 and CA15704 from the National Cancer Institute, Bethesda, MD, and the Jose Carreras Foundation Against Leukemia.

References 1 Stockschlader M, Kruger W, Kroschke G et al. Use of cryopreserved bone marrow in allogeneic bone marrow transplantation. Bone Marrow Transplant 1995; 15: 569–572. 2 Areman EM, Sacher RA, Deeg HJ. Processing and storage of human bone marrow: a survey of current practices in North America. Bone Marrow Transplant 1990; 6: 203–209. 3 Elliot C, McCarthy D. A survey of methods of processing and storage of bone marrow and blood stem cells in the EBMT. Bone Marrow Transplant 1994; 14: 419–423. 4 Karow AM, Webb WR. Tissue freezing. A theory for injury and survival. Cryobiology 1965; 2: 99–108.


1051 5 Takahashi T, Hirsh A, Erbe E, Williams RJ. Mechanism of cryoprotection by extracellular polymeric solutes. Biophys J 1988; 54: 509–518. 6 Ashwood-Smith MJ, Warby C, Connor KW, Becker G. Lowtemperature preservation of mammalian cells in tissue culture with polyvinylpyrrolidone (PVP), dextrans, and hydroxyethyl starch (HES). Cryobiology 1972; 9: 441–449. 7 van Putten LM. Monkey and mouse bone marrow preservation and the choice of technique for human application. Proceedings of the 11th Congress International Society of Blood Transfusion, Sidney, 1968, 797–801. 8 Stiff PJ, Murgo AJ, Zaroulis CG et al. Unfractionated human marrow cell cryopreservation using dimethylsulfoxide and hydroxyethyl starch. Cryobiology 1983; 20: 17–24. 9 Stiff PJ, Koester AR, Weidner MK et al. Autologous bone marrow transplantation using unfractionated cells cryopreserved in dimethylsulfoxide and hydroxyethyl starch without controlled-rate freezing. Blood 1987; 70: 974–978. 10 Makino S, Harada M, Akashi K et al. A simplified method for cryopreservation of peripheral blood stem cells at 801C without rate-controlled freezing. Bone Marrow Transplant 1991; 8: 239–244. 11 Takaue Y, Abe T, Kawano Y et al. Comparative analysis of engraftment after cryopreservation of peripheral blood stem cell autografts by controlled- versus uncontrolled-rate methods. Bone Marrow Transplant 1994; 13: 801–804. 12 Yu J, Leisenring W, Rowley SD et al. The predictive value of white cell or CD34+ cell count in peripheral blood for timing apheresis and maximizing yield. Transfusion 1999; 39: 442–450. 13 Rowley SD, Yu J, Gooley T et al. Trafficking of CD34+ cells into the peripheral circulation during collection of peripheral blood stem cells by apheresis. Bone Marrow Transplant 2001; 28: 649–656. 14 Rowley SD, Bensinger WI, Gooley TA, Buckner CD. Effect of cell concentration on bone marrow and peripheral blood stem cell cryopreservation. Blood 1994; 83: 2731–2736. 15 Rowley SD, Byrne DV. Low-temperature storage of bone marrow in nitrogen vapor-phase refrigerators: decreased temperature gradients with an aluminum racking system. Transfusion 1992; 32: 750–754. 16 Bensinger WI, Schiffman KS, Holmberg L et al. High-dose busulfan, melphalan, thiotepa and peripheral blood stem cell infusion for the treatment of metastatic breast cancer. Bone Marrow Transplant 1997; 19: 1183–1189. 17 Brunvand MW, Bensinger WI, Soll E et al. High-dose fractionated total-body irradiation, etoposide and cyclophosphamide for treatment of malignant lymphoma: comparison of autologous bone marrow and peripheral blood stem cells. Bone Marrow Transplant 1996; 18: 131–141. 18 Press OW, Eary JF, Gooley T et al. A phase I/II trial of iodine131-tositumomab (anti-CD20), etoposide, cyclophosphamide, and autologous stem cell transplantation for relapsed B-cell lymphomas. Blood 2000; 96: 2934–2942. 19 Bensinger WI, Appelbaum FR, Rowley S et al. Factors that influence collection and engraftment of autologous peripheralblood stem cells. J Clin Oncol 1995; 13: 2547–2555.

20 Rowley SD, MacLeod B, Heimfeld S et al. Severe central nervous system toxicity associated with the infusion of cryopreserved PBSC components. Cytotherapy 1999; 1: 311–317. 21 Van Putten LM. The effectiveness of different freeze storage techniques for mouse-bone-marrow cell-suspensions. Ann NY Acad Sci 1964; 114: 695–700. 22 Gilmore MJML. GCT-conditioned medium: an unsuitable stimulus for monitoring granulocyte–macrophage colonyforming cells in cryopreserved bone marrow. Cryobiology 1983; 20: 106–110. 23 Rowley SD, Hattenburg C. Cryopreservation of hematopoietic progenitor cells. Exp Hematol 1990; 18: 678(Abstract). 24 Polge C, Smith AU, Parkes AS. Revival of spermatozoa after vitrification and dehydration at low temperatures. Nature 1949; 164: 666. 25 Lovelock JE, Bishop MWH. Prevention of freezing damage to living cells by dimethylsulphoxide. Nature 1959; 183: 1394–1395. 26 Galmes A, Besalduch J, Bargay J et al. Cryopreservation of hematopoietic progenitor cells with 5-percent dimethyl sulfoxide at 80 degrees C without rate-controlled freezing. Transfusion 1996; 36: 794–797. 27 Ragab AH, Gilkerson E, Myers M. Factors in the cryopreservation of bone marrow cells from children with acute lymphocytic leukemia. Cryobiology 1977; 14: 125–134. 28 Donaldson C, Armitage WJ, Denning-Kendall PA, Nicol AJ, Bradley BA, Howes JM. Optimal cryopreservation of human umbilical cord blood. Bone Marrow Transplant 1996; 18: 725–731. 29 Leibo SP, Farrant J, Mazur P et al. Effects of freezing on marrow stem cell suspensions: interactions of cooling and warming rates in the presence of PVP, sucrose or glycerol. Cryobiology 1970; 6: 315–332. 30 Clark J, Pati A, McCarthy D. Successful cryopreservation of human bone marrow does not require a controlled-rate freezer. Bone Marrow Transplant 1991; 7: 121–125. 31 Rowley SD. Techniques of bone marrow and stem cell cryopreservation and storage. In: Sacher R, AuBuchon J (eds). Marrow Transplantation: Practical and Technical Aspects of Stem Cell Reconstitution. American Association of Blood Banks: Bethesda, 1992, pp 105–127. 32 Bensinger W, Singer J, Appelbaum FR et al. Autologous transplantation with peripheral blood mononuclear cells collected after administration of recombinant granulocyte stimulating factor. Blood 1993; 81: 3158–3163. 33 McQuaker IG, Hunter AE, Pacey S et al. Low-dose filgrastim significantly enhances neutrophil recovery following autologous peripheral-blood stem-cell transplantation in patients with lymphoproliferative disorders: evidence for clinical and economic benefit. J Clin Oncol 1997; 15: 451–457. 34 Davis JM, Rowley SD, Braine HG, Piantadosi S, Santos GW. Clinical toxicity of cryopreserved bone marrow graft infusion. Blood 1990; 75: 781–786.
