High CD34+ cell dose promotes faster platelet recovery after autologous stem cell transplantation for acute myeloid leukemia

Biology of Blood and Marrow Transplantation 9:643-648 (2003) 䊚 2003 American Society for Blood and Marrow Transplantation 1083-8791/03/0910-0006$30.00/0 doi:10.1016/S1083-8791(03)00232-5

High CD34ⴙ Cell Dose Promotes Faster Platelet Recovery after Autologous Stem Cell Transplantation for Acute Myeloid Leukemia Nathan Gunn,1 Lloyd Damon,2 Paul Varosy,1 Willis Navarro,2 Tom Martin,2 Curt Ries,2 Charles Linker2 1 Department of Medicine, 2Adult Leukemia and Bone Marrow Transplant Program, Division of Hematology/Oncology, University of California, San Francisco, California

Correspondence and reprint requests: Charles Linker, MD, Rm. A502, 400 Parnassus Ave., San Francisco, CA 94143 (e-mail: [email protected]). Received September 4, 2002; accepted June 29, 2003

ABSTRACT We studied platelet engraftment in 58 patients with acute myeloid leukemia in first remission treated with autologous stem cell transplantation (ASCT) to determine whether CD34ⴙ cell doses >10 ⴛ 106/kg were associated with faster platelet engraftment. We compared engraftment rates in patients receiving CD34ⴙ doses between 5 and 10 ⴛ 106/kg (standard-dose ASCT) with those receiving doses >10 ⴛ 106/kg (high-dose [HD] ASCT). We also studied neutrophil engraftment rates and platelet and red blood cell transfusion requirements. In multivariate adjusted models, the rate of platelet recovery to >20,000/␮L was 4-fold greater among subjects who received HD-ASCT (hazard ratio [HR], 4.1; confidence interval [CI], 1.8-9.2; P ⴝ .001), with median recovery times of 14 versus 28 days. The rate of platelet recovery to >50,000/␮L was 2-fold greater (HR, 2.1; CI, 1.3-5.9; P ⴝ .01), with median recovery times of 19 versus 46 days. Faster platelet recovery resulted in the need for fewer platelet transfusions among the subjects who received HD-ASCT (mean transfusions, 3.7 versus 9.8; P ⴝ .005). Although not statistically significant, neutrophil recovery data in the adjusted model suggested a similar effect in the HD-ASCT group, with faster engraftment times at absolute neutrophil counts >500/␮L (median, 9.2 versus 12 days; HR, 1.6; CI, 0.69-3.5; P ⴝ .29) and absolute neutrophil counts >1000/␮L (median, 9.5 versus 12 days; HR, 1.3; CI, 0.56-2.8; P ⴝ .58). Subjects who received HD-ASCT required fewer red blood cell transfusions (4.0 versus 9.8 units; P ⴝ .01). Our findings suggest that CD34ⴙ cell doses >10 ⴛ 106/kg CD34ⴙ result in faster engraftment and fewer red blood cell and platelet transfusions. © 2003 American Society for Blood and Marrow Transplantation

KEY WORDS Stem cell

●

CD34⫹

●

Autologous transplantation

INTRODUCTION The use of peripheral blood stem cells (PBSC) rather than bone marrow has resulted in significantly faster hematopoietic engraftment during autologous stem cell transplantation (ASCT) [1-3]. Faster engraftment has translated into less morbidity and mortality and reduced use of resources such as hospitalization, antibiotics, and transfusions [2-9]. With ASCT, the dose of progenitor cells, as assayed by the CD34 antigen, has been shown to correlate with the rate of neutrophil and platelet engraftment and the need for blood product transfusions [1,2,4-6,10-13]. One large retrospective study demonstrated that a

BB&MT

CD34⫹ cell dose ⬎2.5 ⫻ 106/kg was adequate for engraftment [14], whereas others suggested that a dose ⬎5 ⫻ 106/kg was associated with improved time to platelet recovery [8,15,16]. This latter dose has often been believed to be optimal [1,17], and it has been proposed that doses above this threshold would not provide further benefit. However, this question has not been definitively studied. We studied the use of ASCT as treatment for acute myeloid leukemia (AML) in first remission [18]. During the early years of this study, the ability of PBSC to promote durable engraftment had not been established in AML, and prior studies of bone marrow transplantation showed that patients with AML en643

N. Gunn et al.

grafted more slowly than patients with other diagnoses. Because of this, there was concern that durable engraftment from PBSC in patients with AML might also be compromised, either because of the intensity of prior therapy or because of abnormalities in bone marrow biology related to the underlying disease. Therefore, we initially harvested backup bone marrow and attempted to collect a CD34⫹ cell dose ⬎10 ⫻ 106/kg. After we documented consistent and sustained engraftment with PBSC, we subsequently removed the requirement for backup bone marrow and decreased the target CD34⫹ cell dose to 5 ⫻ 106/kg. This change in protocol design, with all other aspects of the treatment held constant, gave us the opportunity to evaluate whether a change in CD34⫹ cell dose affected the speed of platelet engraftment.

METHODS Patients

Between May 1993 and November 1999, we treated 67 patients with AML in first remission in a standardized protocol at the University of California San Francisco. Eligible patients were initially between the ages of 16 and 69 years. All patients were in their first remission with de novo AML and were eligible if they were in complete remission for ⬎30 days and ⬍6 months. Complete remission was defined by meeting all of the following criteria: normal bone marrow morphology with ⬍5% blasts, resolution of previously abnormal cytogenetics, no evidence of extramedullary leukemia, and an absolute neutrophil count (ANC) ⱖ500/␮L and platelet count ⱖ140,000/␮L. Additionally, organ function needed to be adequate, as demonstrated by the presence of the following: bilirubin ⬍1.5 mg/dL, alkaline phosphatase and aspartate transaminase less than twice the upper limit of normal, creatinine ⬍2.0 mg/dL, cardiac ejection fraction ⱖ40%, and a carbon monoxide diffusion capacity ⬎50%. Exclusion criteria included prior myelodysplasia, myeloproliferative disease, or chemotherapy-related leukemia. Patients who had received prior postremission high-dose cytarabine were ineligible, but patients may have received less intensive postremission chemotherapy. All patients gave written, informed consent in accord with the University of California San Francisco Committee on Human Research. Of the 67 patients originally enrolled in our prospective cohort analysis, 2 patients were excluded because of incomplete records, and 7 were excluded because of early relapse of AML. The remaining 58 patients were retrospectively divided into high-dose (HD) and standard-dose (SD) ASCT recipients depending on whether their CD34⫹ cell dose was 5 to 9.9 ⫻ 106/kg or ⱖ10 ⫻ 106/kg. The predictor variable in this study was CD34⫹ 644

cell dose during ASCT. The primary outcome was time to platelet engraftment above threshold values of 20,000/␮L, 50,000/␮L, and 100,000/␮L. Secondary outcomes included time to neutrophil engraftment above ANC ⬎500/␮L and ANC ⬎1000/␮L and total platelet and red blood cell (RBC) transfusion requirements. The time to platelet recovery was defined as the first day of 2 contiguous platelet values above the threshold value not dependent on platelet transfusion; subsequent platelet measurements could not decrease below the threshold value. Day 0 was defined as the day of transplantation. Treatment

Patients received therapy in 2 sequential steps. The first step was consolidation chemotherapy with cytarabine 2000 mg/m2 intravenously (IV) every 12 hours ⫻ 8 doses on days 1 to 4 plus etoposide 40 mg/kg by continuous IV infusion over 96 hours on days 1 to 4. Granulocyte colony-stimulating factor (G-CSF) 5 ␮g/kg/d subcutaneously was administered from day 14 until completion of stem cell collection. The G-CSF dose could be escalated to 10 ␮g/kg/d if stem cell collection was proceeding slowly. Stem cell collection was initiated when the white blood cell count was ⱖ10,000/␮L. From 1993 to November 1996, the target harvest dose of CD34⫹ cells was ⱖ10 ⫻ 106/kg. Thereafter, the target dose was reduced to ⱖ5 ⫻ 106/kg. Stem cells were collected by apheresis; 18 L of blood was processed daily. A buffy coat was prepared by centrifugation and mixed in M199 media with 5% autologous plasma and 10% dimethyl sulfoxide to achieve a final cell concentration of 2.5 ⫻ 108 cells per milliliter. The stem cell product was frozen in a controlled-rate freezer and stored in liquid-phase liquid nitrogen. After a rest period of at least 4 weeks (median, 6 weeks) after consolidation, patients were treated with ASCT. The preparative regimen was busulfan 16 mg/kg orally (total dose) administered on days ⫺7 to ⫺4 plus etoposide 60 mg/kg IV over 10 hours on day ⫺3. Unmanipulated stem cells were infused on day 0. G-CSF 5 ␮g/kg/d subcutaneously was given from day 0 until ANC was ⱖ1500/␮L for 2 days or ⱖ5000/␮L for 1 day. Platelet transfusions were given if platelets were ⬍15,000/␮L in the presence of mucositis or fever or ⬍10,000/␮L in asymptomatic patients. Patients received RBC transfusions for a hematocrit ⬍25%. CD34 Assay

Cells were analyzed by flow cytometry (FACScan; Becton Dickinson, Franklin Lakes, NJ), and 75,000 CD45⫹ events were collected for analysis by using the International Society for Hematotherapy and Graft Engineering protocol [19]. The total leukocyte count was performed by using a standard hematology ana-

High CD34ⴙ Dose

Table 1. Patient Characteristics: n (%), Mean ⫾ SD, or Median (Range) HD-ASCT* (N ⴝ 48)

Characteristic Baseline Age Male sex Cytogenetics Favorable Intermediate Poor Unknown Induction phase Regimen† High-dose cytarabine Standard-dose cytarabine All-trans retinoic acid Consolidation phase Days to platelets >20000/␮L >50000/␮L >100000/␮L Days to start of stem cell collection Number of aphereses Days from consolidation to ASCT Harvest Before November 1996 After November 1996 Transplantation phase CD34ⴙ transplant dose ⫹

SD-ASCT* (N ⴝ 10)

42 ⴞ 13 20 (42)

37 ⴞ 9.8 3 (30)

9 (19) 28 (58) 1 (2) 10 (21)

0 (0) 9 (90) 0 (0) 1 (10)

30 (63) 12 (25) 6 (13)

7 (70) 3 (30) 0 (0)

26 ⴞ 8.8 30 ⴞ 11 46 ⴞ 54 25 ⴞ 3.8 2 ⴞ 1.7 47 ⴞ 15

31 ⴞ 9.6 34 ⴞ 11 40 ⴞ 22 26 ⴞ 6.1 2 ⴞ 2.9 50 ⴞ 27

P Value

.27 .49 .27

.50

25 21

3 7

17 (10-164) ⫹

.12 .33 .78 .31 .43 .59 .16

6.6 (4.2-9.6)

.001

*SD-ASCT: CD34 5-9.9 ⫻ 10 /kg; HD-ASCT: CD34 ⱖ 10 ⫻ 10 /kg. †High-dose cytarabine regimens included daunorubicin. Standard-dose cytarabine regimens included anthracycline. All-trans retinoic acid regimens included either idarubicin or daunorubicin. 6

6

lyzer (Cell Dyn 3500 or 4000; Abbott Laboratories, Chicago, IL).

College Station, TX). Two-sided P values ⬍.05 were considered statistically significant.

Statistical Analysis

RESULTS

We compared the baseline characteristics of subjects who received HD-ASCT and SD-ASCT by using unpaired t tests, Mann-Whitney tests, and ␹2 tests where appropriate. We compared the median times to platelet recovery by using t tests or MannWhitney tests, depending on the observed distributions of the data. We estimated the rates of platelet recovery in the 2 groups by using the method of Kaplan and Meier, and we tested for differences between the 2 groups with the log-rank test [20,21]. We used Cox proportional hazards models [20,22] to determine the association of stem cell dose with platelet recovery, adjusting for potential confounders, including age, sex, cytogenetics, all-trans retinoic acid–versus cytarabine-based induction regimen, platelet recovery times during consolidation, and whether PBSC were harvested after November 1996. We tested for deviations from the assumption of proportional hazards during follow-up [23,24], and we found the assumption to be valid. All analyses were performed with Stata 7.0 (Stata Corp.,

Patients

BB&MT

HD-ASCT patients received a median CD34⫹ dose of 17 ⫻ 106/kg (range, 10-164 ⫻ 106/kg), whereas SD-ASCT received a median CD34⫹ dose of 6.6 ⫻ 106/kg (range, 4.2-9.6 ⫻ 106/kg; P ⫽ .001). We identified no statistically significant differences between the 2 groups in age, sex, cytogenetics (favorable, intermediate, poor, and unknown), induction chemotherapy regimen, consolidation-phase platelet recovery times, time from consolidation to harvest, time from consolidation to transplantation, or whether harvest occurred after November 1996 (when the CD34⫹ harvest goal was decreased from ⬎10 ⫻ 106/kg to ⬎5 ⫻ 106/kg; Table 1). There were no clinically apparent retinal, neurovascular, or gastrointestinal bleeding events in either group. There were no significant differences between groups in the development or duration of neutropenic fevers or the amount of prophylactic antibiotics used for these periods. Time to discharge was not significantly different: 26 ⫾ 6.9 days in the HD-ASCT 645

N. Gunn et al.

Table 2. CD34⫹ Cell Dose and Engraftment Kinetics Unadjusted Median Days to Recovery after ASCT: Median (Range) or Mean ⴞ SD Variable Platelet recovery >20000/␮L >50000/␮L >100000/␮L ANC recovery >500/␮L >1000/␮L Platelet transfusions RBC transfusions

Multivariate-Adjusted*

HD-ASCT†

SD-ASCT†

Hazard Ratio (95% CI)

P Value

Hazard Ratio (95% CI)

P Value

14 (0-90) 19 (8-193) 29 (9-448)

28 (15-190) 46 (21-190) 129 (28-330)

3.5 (1.7-7.4) 2.5 (1.3-5.1) 1.5 (0.7-3.2)

.001 .01 .30

4.1 (1.8-9.2) 2.1 (1.3-5.9) 1.3 (0.55-2.9)

.001 .01 .59

9.2 ⴞ 1.0 9.5 ⴞ 1.2 3.7 ⴞ 2.5 4.0 ⴞ 2.9

12 ⴞ 5.2 12 ⴞ 5.6 9.8 ⴞ 3.3 9.8 ⴞ 15

2.1 (1.0-4.5) 1.8 (0.84-3.8)

.048 .14 .005 .01

1.6 (0.69-3.5) 1.3 (0.56-2.8)

.29 .58

CI indicates confidence interval. *Hazard ratios were adjusted for age, sex, cytogenetics profile, induction chemotherapeutic regimen, consolidation-phase platelet recovery times, time from consolidation to harvest, and whether cells were harvested after November 1996. †SD-ASCT: CD34⫹ 5-9.9 ⫻ 106/kg; HD-ASCT: CD34⫹ ⱖ10 ⫻ 106/kg.

group versus 28 ⫾ 7.5 days in the SD-ASCT group (P ⫽ .39). Platelet and Neutrophil Engraftment

In unadjusted models, HD-ASCT was associated with ⬎3-fold faster platelet engraftment at ⱖ20,000/␮L and ⬎2-fold faster engraftment at ⱖ50,000/␮L (Table 2; Figure 1). Our multivariate adjusted model accounted for the following factors: age, sex, cytogenetic risk group, induction chemotherapy regimen, consolidation-phase platelet recovery time, time between consolidation and harvest, and whether harvest occurred after November 1996. In the adjusted model, HD-ASCT was associated with a 4-fold faster rate of platelet recovery to ⱖ20,000/␮L and a 2-fold faster rate of platelet recovery to ⱖ50,000/␮L, with median engraftment times of 14 versus 28 days (P ⫽ .001) and 19 versus 46 days (P ⫽ .01) days, respectively (Table 2). There was suggestion of faster recovery to platelets ⱖ100,000/␮L, with median recovery times of 29 versus 129 days, but there was no statistically significant difference between the groups in either the adjusted or unadjusted models (Table 2). In the unadjusted model, HD-ASCT was associated with a statistically significant 2-fold increase in the neutrophil engraftment rate at ANC ⬎500/␮L, with median times of 9.2 versus 12 days (P ⫽ .048). Although this was not statistically significant, in the adjusted model, HD-ASCT was associated with a trend toward faster neutrophil engraftment at ANC ⬎500/␮L and ANC ⬎1000/␮L (Table 2). Patients who received HD-ASCT required fewer supportive blood product transfusions. Mean platelet transfusions were lower in the HD-ASCT group, at 3.7 ⫾ 2.5 versus 9.8 ⫾ 3.3 (P ⫽ .005). Mean RBC transfusion requirements were lower in the HDASCT group, at 4.0 ⫾ 2.9 versus 9.8 ⫾ 15 (P ⫽ .01). 646

There was no statistically significant difference in event-free survival (EFS) between patients who received HD-ASCT and SD-ASCT, with 6-year EFS rates of 61% versus 56%, respectively. When patients with acute promyelocytic leukemia are excluded (all of whom were in the HD-ASCT group), the EFS is 51% versus 56% (P ⫽ .74). DISCUSSION In this study, we found that subjects with AML who received HD-ASCT (CD34⫹ cell dose ⱖ10 ⫻ 106/kg) had faster engraftment of platelets than patients who received SD-ASCT (CD34⫹ cell doses between 5 and 9.9 ⫻ 106/kg). The differences were both statistically significant and clinically relevant, with a 2-week reduction in the time to platelets ⱖ20,000/␮L and a 4-week reduction in time to platelet recovery ⱖ50,000/␮L. Faster platelet recovery resulted in 6 fewer platelet transfusions. HD-ASCT patients also showed a trend toward faster neutrophil recovery—3 days faster to ANC ⬎500/␮L and to ANC ⬎1000/␮L—and required 6 fewer units of transfused RBCs. Other studies have generally shown an inverse correlation between the CD34⫹ dose and hematopoietic engraftment rate [1,2,4-6,10-13]. However, poor interinstitution CD34 assay reproducibility makes the validity of analysis of pooled data uncertain [19,25]. In addition, previous studies suggest that factors such as differing forms of chemotherapy and conditioning regimens [1,13,26,27] and having received a prior bone marrow transplantation [4] may affect engraftment rates. Finally, different types of cancer may affect posttransplantation platelet recovery times [27,28]. Therefore, extrapolation of these previously described dose-response relationships to the use of HD-ASCT in AML may not be valid.

High CD34ⴙ Dose

Figure 1. Kaplan-Meier cumulative probability curves illustrating the time to platelet recovery to ⬎20,000/␮L, ⬎50,000/␮L, and ⬎100,000/␮L for high-dose versus standard-dose autologous stem cell transplantation (HD-ASCT and SD-ASCT, respectively). In the adjusted model, HD-ASCT was associated with faster platelet engraftment to 20,000/␮L (P ⫽ .001) and to 50,000/␮L (P ⫽ .01), but not to 100,000/␮L (P ⫽ .59).

In this study, we examined a specific, relatively homogeneous patient cohort: patients with AML in first remission who were treated with uniform therapy (after remission induction) at a single institution. We accounted for possible confounders with multivariate regression analysis. Variables included age, sex, cytogenetic risk group, remission induction chemotherapy regimen, and time between consolidation and harvest. We also controlled for platelet engraftment time during consolidation, because a slow pretransplantation platelet recovery rate may have been a marker for a poor marrow microenvironment. Finally, because the CD34⫹ cell dose requirement was decreased after November 1996, we controlled for whether the harvest occurred after this time point. This controlled for

BB&MT

those who had a CD34⫹ cell dose ⬍10 ⫻ 106/kg by design (ie, cells were harvested after November 1996 with a goal CD34⫹ of ⱖ5 ⫻ 106/kg) versus those who could not reach a threshold of 10 ⫻ 106/kg because of poor stem cell mobilization (n ⫽ 3) before November 1996. Our data, adjusted for these variables, demonstrate that infused CD34⫹ cell doses ⬎10 ⫻ 106/kg are associated with faster engraftment. Our study had a relatively small number of patients. Although this does not affect the positive results of this study, the lack of significance in negative results may be a result of low power. For example, there may be some survival advantage to high-dose transplantation because of decreased amounts of thrombocytopenic-associated complications; con647

N. Gunn et al.

versely, high-dose harvesting may result in the transplantation of higher numbers of tumor cells, leading to higher rates of relapse. Such outcomes may become manifest only in a study with increased power. Nevertheless, no differences were found between groups in the EFS rates or in the number of morbid complications of prolonged thrombocytopenia. Our findings offer evidence that ASCT with CD34⫹ cell doses ⬎10 ⫻ 106/kg is associated with faster platelet engraftment that is clinically meaningful and results in decreased need for transfusion of platelets and RBCs. In addition, the ability to achieve rapid and sustained platelet engraftment early after transplantation may facilitate the use of posttransplantation therapies such as tumor vaccines and interleukin-2, which are generally started when adequate levels of blood count recovery have occurred.

11.

12.

13.

14.

15.

REFERENCES 1. Bentley SA, Brecher ME, Powell E, Serody JS, Wiley JM, Shea TC. Long-term engraftment failure after marrow ablation and autologous hematopoietic reconstitution: differences between peripheral blood stem cell and bone marrow recipients. Bone Marrow Transplant. 1997;19:557-563. 2. Beyer J, Schwella N, Zingsem J, et al. Hematopoietic rescue after high-dose chemotherapy using autologous peripheralblood progenitor cells or bone marrow: a randomized comparison. J Clin Oncol. 1995;13:1328-1335. 3. Schmitz N, et al. Randomised trial of filgrastim-mobilised peripheral blood progenitor cell transplantation versus autologous bone-marrow transplantation in lymphoma patients. Lancet. 1996;347:353-357. 4. Bernstein SH, Nademanee AP, Vose JM, et al. A multicenter study of platelet recovery and utilization in patients after myeloablative therapy and hematopoietic stem cell transplantation. Blood. 1998;91:3509-3517. 5. Ketterer N, Salles G, Raba M, et al. High CD34(⫹) cell counts decrease hematologic toxicity of autologous peripheral blood progenitor cell transplantation. Blood. 1998;91:3148-3155. 6. Salazar R, Maroto P, et al. Factors affecting CD34⫹ cell mobilization with cyclophosphamide ⫹ G-CSF and hematopoietic recovery after high-dose chemotherapy and stem cell support in solid tumor patients. Proc Am Soc Clin Oncol. 1996;17(95a): (abstr. 368). 7. Sola C, Salazar R, et al. High-dose chemotherapy (HDC) and peripheral blood stem cell (PBSC) autologous transplantation: influence of the number of infused CD34⫹ cells in hematopoietic recovery and support measures required. Proc Am Soc Clin Oncol. 1996;15: (abstr. 1746). 8. Kiss JE, Rybka WB, Winkelstein A, et al. Relationship of CD34⫹ cell dose to early and late hematopoiesis following autologous peripheral blood stem cell transplantation. Bone Marrow Transplant. 1997;19:303-310. 9. Glaspy J, Lu C, Wheeler C, et al. Economic rationale for infusing optimal numbers of CD34⫹ cells in peripheral blood progenitor cell transplants (PBPCT). Blood. 1997;90(suppl 1): 370a (abstr 1646). 10. Benedetti G, Patoia L, Giglietta A, Alessio M, Pelicci PG, Grignani F. Very large amounts of peripheral blood progenitor cells eliminate severe thrombocytopenia after high-dose mel648

16.

17.

18.

19.

20. 21. 22. 23. 24.

25.

26.

27.

28.

phalan in advanced breast cancer patients. Bone Marrow Transplant. 1999;24:971-979. Dercksen MW, Rodenhuis S, Dirkson MKA, et al. Subsets of CD34⫹ cells and rapid hematopoietic recovery after peripheralblood stem-cell transplantation. J Clin Oncol. 1995;13:1922-1932. Lowenthal RM, Fabe`res C, Marit G, et al. Factors influencing haemopoietic recovery following chemotherapy-mobilised autologous peripheral blood progenitor cell transplantation for haematological malignancies: a retrospective analysis of a 10-year single institution experience. Bone Marrow Transplant. 1998;22:763-770. Weaver CH, Hazelton B, Birch R, et al. An analysis of engraftment kinetics as a function of the CD34 content of peripheral blood progenitor cell collections in 692 patients after the administration of myeloablative chemotherapy. Blood. 1995;86: 3961-3969. Haas R, Mohle R, Fruhauf S, et al. Patient characteristics associated with successful mobilizing and autografting of peripheral blood progenitor cells in malignant lymphoma. Blood. 1994;83:3787-3794. Bensinger W, Appelbaum F, Rowley S, et al. Factors that influence collection and engraftment of autologous peripheralblood stem cells. J Clin Oncol. 1995;13:2547-2555. van der Wall E, Richel DJ, Holtkamp MJ, et al. Bone marrow reconstitution after high-dose chemotherapy and autologous peripheral blood progenitor cell transplantation: effect of graft size. Ann Oncol. 1994;5:795-802. Beguin Y, Baudoux E, Sautois B, et al. Hematopoietic recovery in cancer patients after transplantation of autologous peripheral blood CD34⫹ cells or unmanipulated peripheral blood stem and progenitor cells. Transfusion. 1998;38:199-208. Linker CA, Ries CA, Damon LE, et al. Autologous stem cell transplantation for acute myeloid leukemia in first remission. Biol Blood Marrow Transplant. 2000;6:50-57. Sutherland DR, et al. The ISHAGE guidelines for CD34⫹ cell determination by flow cytometry. International Society of Hematotherapy and Graft Engineering. J Hematother. 1996;5:213226. Cox DR, Oakes D. Analysis of Survival Data. London: Chapman & Hall; 1984. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-481. Crowley J, Hu M. Covariance analysis of heart transplant survival data. J Am Stat Assoc. 1977;72:27-36. Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69:239-241. Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;81: 515-526. Siena S, Bregni M, Brando B, et al. Flow cytometry for clinical estimation of circulating hematopoietic progenitors for autologous transplantation in cancer patients. Blood. 1991;77:400-409. Millar BC, Millar JL, Shepard V, et al. The importance of CD34⫹/CD33⫺ cells in platelet engraftment after intensive therapy for cancer patients given peripheral blood stem cell rescue. Bone Marrow Transplant. 1998;22:469-475. Domenech J, Linassier C, Gihana E, et al. Prolonged impairment of hematopoiesis after high-dose therapy followed by autologous bone marrow transplantation. Blood. 1995;85:3320-3327. Pecora AL, Preti RA, Gleim GW, et al. CD34⫹CD33⫺ cells influence days to engraftment and transfusion requirements in autologous blood stem-cell recipients. J Clin Oncol. 1998;16: 2093-2104.