Cyclophosphamide, Doxorubicin, Vincristine, Prednisone Dose Intensification With Granulocyte Colony-Stimulating Factor Markedly Depletes Stem Cell Reserve for Autologous Bone Marrow Transplantation

ONE APPROACH TO improve the efficacy of cancer chemotherapy is through dose escalation.1 The availability of hematopoietic growth factors has facilitated this approach by minimizing myelosuppression, as well as shortening the duration of chemotherapeutic cycles.2-5 Although definitive evidence for a benefit of this approach is still uncertain, a major concern is that repeated cycles of cytotoxic drugs followed by growth factors might result in damage to hematopoietic stem cells.

Autologous hematopoietic stem cell transplantation has become widely used in the treatment of patients with a variety of malignancies, including solid tumors and hematologic malignancies.6-8 Many of these patients have had extensive cytotoxic therapy before autologous transplantation, which may have damaged their hematopoietic stem cell compartment. Although a number of factors may affect posttransplant engraftment, including nonhematologic toxicities, recurrent disease, and the intensity of ablative therapy, it is the quality and quantity of the reinfused autologous stem cell product that is likely to be the major factor that affects engraftment.

In this study, patients with previously untreated advanced stage follicular lymphoma were selected on an intent to treat basis to receive both combination chemotherapy with cyclophosphamide, doxorubicin, vincristine, prednisone (CHOP) in an attempt to induce either a complete or very good partial remission (induction therapy) and autologous bone marrow transplantation (ABMT). We report here the results of two sequential trials that varied only in the dose intensity of CHOP and whether growth factor support accompanied this dose intensification. The two sequential studies will be referred to as: (1) standard dose CHOP (SD-CHOP) or (2) dose escalated CHOP (HD-CHOP) induction with granulocyte colony-stimulating factor (G-CSF ) support. The rationale for dose escalation of CHOP (HD-CHOP) was based on our observation that SD-CHOP induced complete remissions in only one third of patients before ABMT. In contrast to the virtually uniform hematologic engraftment observed in the patients whose induction therapy consisted of SD-CHOP, the most striking finding was markedly retarded engraftment in a highly significant number of patients whose induction therapy was HD-CHOP and G-CSF. These data suggest that induction therapy with intensive chemotherapy and growth factor support resulted in a decreased content and proliferative capacity of stem cells within transplanted autologous marrow. Moreover, this approach may limit the choice of future potentially curative treatment options that are dependent on adequate stem cell reserve.

Selection of patients and treatment protocol. Patient eligibility and selection were identical for each of the sequential studies. Patients were eligible for high dose therapy and ABMT after CHOP induction in first remission if they were: (1) physiologic age 55 years or less; (2) had previously untreated follicular low grade B-cell non-Hodgkin's lymphoma (NHL) as defined by the Working Formulation (WF )9 including follicular small cleaved cell (WF-B) and follicular mixed small cleaved and large cell (WF-C); and (3) had lymphoma cells that expressed the CD20 (B1) antigen as previously described.10 Patients had to have stage IIIB, IIIE, or III with masses >5 cm or stage IV disease. Patients with stage IV disease by virtue of minimal adenopathy (<1 cm) and <5% bone marrow [BM] involvement were excluded. Additional criteria for entry included the absence of comorbid disease of the heart, kidney, lung, and liver and a Karnofsky score above 80%. Eighty-three patients were registered and treated with six to eight cycles of SD-CHOP (cyclophosphamide 750 mg/m² intravenously [IV] day 1; doxorubicin 50 mg/m² IV day 1; vincristine 1.4 mg/m² (maximum 2 mg) IV day 1; and prednisone 100 mg/day by mouth days 1 to 5).11 Twenty patients were registered and received four courses of HD-CHOP every 3 weeks.12 This regimen consisted of: cyclophosphamide 1.5 g/m²/day IV day 1 and day 2; doxorubicin 50 mg/m² IV day 1; vincristine 1.4 mg/m² (maximum 2 mg) IV day 1; and prednisone 100 mg/d by mouth days 1 to 5. Mesna was given by IV bolus as uroprotectant. G-CSF 5 μg/kg/day subcutaneously (SC) was given days 4 to 18 of each cycle. All patients received prophylactic acyclovir and trimethoprim-sulfamethoxazole during each cycle. At the completion of SD-CHOP or HD-CHOP, patients in complete remission (CR) or minimal disease state went on to BM harvest. Minimal disease status was defined as lymph nodal mass less than 2 cm in its greatest diameter and histologic evidence of BM involvement of 10% or less of the intratrabecular space as determined by iliac crest biopsy. BM cellularity was also assessed on the biopsy specimen before harvest. After SD-CHOP, 10 patients (13%) with 1 to 3 masses >2 cm received involved field radiotherapy. After completion of HD-CHOP, two patients (11%) with 1 mass >2 cm received involved field radiotherapy of 2,200 to 2,500 cGy. Informed consent was obtained from all patients.

In both studies, patients underwent BM harvest within 4 weeks of completing chemotherapy. In patients not receiving involved field radiotherapy, admission for ABMT was within 4 weeks of BM harvest. In both studies, preparative therapy for ABMT consisted of cyclophosphamide, 60 mg/kg of body weight, infused on each of two consecutive days before radiotherapy. Total body irradiation was administered in fractionated doses (200 cGy) twice daily on three consecutive days (total of 1,200 cGy) in all patients. Patients who were treated with SD-CHOP induction did not receive any hematopoietic growth factors after ABMT. All patients who received HD-CHOP induction received hematopoietic growth factors after ABMT. Supportive care was provided as previously described.13 

Collection, processing, and infusion of marrow. BM was obtained, treated in vitro with anti-B1 (CD20), -B5, and -J5 (anti-CD10) monoclonal antibodies and rabbit complement and cryopreserved as previously described.13 No patients were excluded from the protocol after BM harvest. Within 18 hours of the completion of total body irradiation, the cryopreserved marrow cells were reinfused as previously described.13 

Analysis of engraftment. Trilineage engraftment after ABMT was compared in patients who received SD-CHOP or HD-CHOP as induction therapy. The data is expressed as the percentage of patients not achieving an expected value of white blood cell (WBC), hematocrit (Hct), and platelet counts at days 100, 180, and 360 post-ABMT. The expected values were for day 100: Absolute neutrophil count (ANC) 1,200/μL; Hct 28; platelet count 50,000/μL. For day 180: ANC 1,600/μL; Hct 30; platelet count 75,000/μL. For day 360: ANC 2000/μL; Hct >30; platelet count 100,000/μL. These values correspond to the National Institutes of Health Common Toxicity Criteria (CTC) hematologic toxicities grade 2 at 100 days, grade 1 at 180 days, and grade 0 at 360 days post-ABMT.

Statistical methods. Changes in nadir WBC and platelet counts between cycle 1 and cycle 4 of HD-CHOP were compared using a paired t-test for the 17 patients with complete data on nadir counts. Incidence of fever and neutropenia in each cycle was compared using the Fisher exact test. Differences in WBC, hematocrit, and platelet counts at conditioning for transplant, as well as number of cells/kg reinfused and length of hospital stay, were compared between SD-CHOP and HD-CHOP using the Wilcoxon rank sum test. BM cellularity in SD-CHOP and HD-CHOP patients was compared using a two-sample t-test. The time to neutrophil engraftment, defined as the first of two consecutive days with ANC of 500/μL or higher and time to platelet engraftment, defined as the first of two consecutive days with platelets ≥20,000/mm3 with no platelet transfusion were compared using the Wilcoxon test. The numbers of patients not achieving expected hematologic recovery goals at 100, 180, and 360 days post-ABMT, as described above, were compared using the Fisher exact test. All P values reflected two-sided levels of significance.

This report summarizes the outcome of two sequential trials of CHOP induction followed by identical myeloablative therapy and ABMT in previously untreated, advanced stage follicular lymphoma patients. In the first study, SD-CHOP induction therapy was administered every 21 days for six to eight cycles without growth factor support. In the second study, HD-CHOP induction was administered every 21 days for four cycles with G-CSF support. The clinical results of the SD-CHOP study have been recently reported10 and, herein, we report the comparison of hematologic engraftment between these two studies.

Patient characteristics. Twenty patients (median age, 44) with previously untreated, advanced stage follicular lymphoma were considered for induction therapy with HD-CHOP and G-CSF followed by high dose chemoradiotherapy and ABMT in first remission. The characteristics of these patients are detailed in Table 1. All of these patients had progressive disease at the time of consideration and patients with minimal stage III/IV disease were excluded. At diagnosis, 18 patients had follicular small cleaved cell histology and two patients had follicular mixed small cleaved and large cell lymphoma. The majority of patients had bulky disease with 60% having masses greater than 5 cm, and 19 patients had stage IV disease by virtue of BM involvement. The characteristics of the patients treated with HD-CHOP who underwent ABMT were similar to those of 77 patients with previously untreated follicular lymphoma who received SD-CHOP before ABMT (Table 1).10 There was no statistically significant difference between the two groups of patients.

Treatment with HD-CHOP. After four cycles of HD-CHOP and G-CSF, only six of the 20 patients were in a clinical CR (31%). Thirteen patients achieved minimal disease state, while one patient progressed with a high grade lymphoma within 1 month after completing cycle 4. At the time of harvest, seven patients had histologic evidence of BM involvement. The CR rate after HD-CHOP was similar to that seen in previously untreated patients with follicular NHL treated with six to eight courses of SD-CHOP before ABMT, where the CR rate was 36%.10 

Treatment with HD-CHOP and G-CSF was associated with profound myelosuppression in these patients with follicular lymphoma (Table 2). With each successive cycle, a greater degreee of leukopenia and thrombocytopenia was observed. Whereas the median nadirs for WBC and platelets were 500/μL and 70,000/μL for cycle 1, by cycle 4, these had significantly decreased to 200/μL and 18,000/μL, respectively (P = .02 for WBC, P = .0001 for platelets). Despite this progressive myelosuppression, the interval between administration of each course of therapy remained quite constant, with a median of 21 days for cycle 1 to 2 and cycle 3 to 4, and 21.5 days for the interval between cycle 2 to 3. Fever associated with neutropenia requiring hospitalization was seen in 24 of 80 (30%) courses of HD-CHOP and G-CSF administered despite administration of prophylatic trimethoprim-sulfamethoxazole during each cyle. No treatment-related deaths were seen. The incidence of fever and neutropenia was not statistically different between cycles (P = .56).

ABMT course and outcome. Nineteen of the 20 patients underwent BM harvest and ABMT. The patient who developed a histologic transformation to high grade lymphoma died of progressive disease 3 months after completing HD-CHOP. At BM harvest, all patients had WBC above 3000/μL, while two patients had Hct less than 28, and one patient had a platelet count less than 100,000/μL. BM cellularity was assessed before harvest in both groups of patients. Patients who received SD-CHOP had a mean cellularity of 46% (range, 15% to 90%), while patients treated with HD-CHOP had mean cellularity of 69% (range, 30% to 90%). These were significantly different (P = .0007) using a two sample t-test. At the time of beginning myeloablative therapy, the median WBC, platelet count, and Hct were significantly lower than that seen in patients in the antecedent study where SD-CHOP induction was administered before ABMT (Table 3). In the 19 patients who received HD-CHOP, the mean number of anti-B–cell monoclonal antibody-purged BM mononuclear cells reinfused was 4.92 × 10⁷/kg (range, 1.94 × 10⁷ to 11.42 × 10⁷), which is not significantly different (P = .31) from patients treated with SD-CHOP induction (4.06 × 10⁷/kg). CD34⁺ cells were not enumerated in the marrow either before or after ex vivo purging in either study.

After HD-CHOP and G-CSF induction and ABMT, there were no acute in-hospital treatment-related deaths. After SD-CHOP induction and ABMT, there were two acute in-hospital treatment-related deaths both from diffuse alveolar hemorrhage syndrome. Fever associated with neutropenia was seen in all 19 HD-CHOP patients with four documented bacterial infections. Of the 77 patients who received SH-CHOP induction, 71 patients developed fever with neutropenia and nine had documented bacterial infections. The median hospital stay was 31.5 days for patients undergoing HD-CHOP induction, which is significantly longer than that seen in patients who received SD-CHOP induction therapy before ABMT, where the median duration of hospitalization was 26 days (P = .03). Two patients who received HD-CHOP and were platelet and RBC transfusion-dependent have died after relapse, one of hemorrhage and disseminated candidiasis and one of gastrointestinal bleeding at 8 and 23 months post-ABMT.

Hematologic engraftment. The time to attain an ANC of 500/μL was not significantly greater in patients who received HD-CHOP (median, 24 days) than seen in patients who received SD-CHOP (median, 19 days) (P = .13). The time to platelet engraftment was significantly greater in patients who received HD-CHOP induction (median, 56 days; range, 17 to 570+) than seen in patients who received SD-CHOP (median, 27 days; range, 14 to 75) (P = .0005). None of the patients treated with SD-CHOP received any hematopoietic growth factors after ABMT, whereas all patients who were treated with HD-CHOP induction received hematopoietic growth factors after ABMT. The median number of days that the HD-CHOP patients received hematopoietic growth factors post-ABMT was 27 days (range, 15 to 57). Despite hematopoietic growth factor support in the HD-CHOP group, it became clear with follow-up that trilineage engraftment in the patients who received HD-CHOP induction therapy was significantly prolonged as compared with patients treated with SD-CHOP induction therapy.

Hematologic engraftment was analyzed at 100, 180, and 360 days post-ABMT. The values correspond to NCI-CTC hematologic toxicities of grade 2 at 100 days, grade 1 at 180 days, and grade 0 at 360 days post-ABMT. As seen in Table 4, at 100, 180, and 360 days post-ABMT, patients who received HD-CHOP had marked delay in achieving the expected neutrophil count at those time intervals, of 1,200/μL, 1,600/μL, and 2,000/μL, respectively. This is highly significantly different from the patients who received SD-CHOP induction before ABMT, where the vast majority attained the expected neutrophil count for those intervals. In marked contrast, the patients who received HD-CHOP, failed to develop adequate numbers of neutrophils. At 1 year, 63% of the HD-CHOP patients had an ANC of under 2,000/μL, whereas 89% of the SD-CHOP patients had an ANC of 2,000/μL or above. Similarly, platelet engraftment was markedly delayed in the patients who received HD-CHOP, whereas by 360 days, 75% had a platelet count of less than 100,000/μL (Table 5). In contrast, 89% of patients who received SD-CHOP induction therapy had a platelet count of 100,000/μL or greater at 1 year. Red blood cell (RBC) engraftment was also markedly impaired in the 19 patients who received HD-CHOP before ABMT (Table 6). Although not statistically significant, RBC engraftment was delayed as compared with patients who received SD-CHOP. At 1 year post-ABMT, 31% of the HD-CHOP patients did not have an Hct of greater than 30 as compared with 19% of the SD-CHOP patients.

At 1 year post-ABMT, four of 18 HD-CHOP patients in remission were still platelet transfusion-dependent and three of these four patients were also RBC-dependent. The marked impairment in platelet engraftment is reflected by the observation that two of 15 patients who are alive and in remission are still platelet and RBC transfusion-dependent at 26 months post-ABMT. In contrast, none of the 77 patients treated with SD-CHOP induction were platelet or RBC transfusion-dependent at or beyond 1 year post-ABMT. We examined the relationship between BM cellularity before harvest and engraftment in both groups of patients. We found that low cellularity (less than 50%) was not significantly associated with delayed neutrophil engraftment in either patients treated with SD-CHOP or HD-CHOP induction. Among the patients receiving HD-CHOP, low BM cellularity before harvest was significantly associated with delayed platelet engraftment (P = .02), whereas no such association was seen among the patients treated with SD-CHOP. Moreover, there was no significant association between the marrow cellularity before harvest and number of BM mononuclear cells per kg reinfused among the HD-CHOP treated patients (not shown). This suggests that the engraftment problems of the HD-CHOP patients with low marrow cellularity may represent poor marrow stem cell reserve, rather than simply a problem of transfused cell dose numbers.

We report a complication of the use of high dose induction therapy with hematopoietic growth factor support not previously observed in clinical stem cell transplantation. Despite the significant myelosuppression observed after each cycle of HD-CHOP and G-CSF support, trilineage hematologic recovery occurred with acceptable median WBC, RBC, and platelet counts at admission for ABMT. Therefore, before ABMT, there was no indication that hematologic engraftment would be different from that seen in patients with follicular NHL undergoing ABMT at our institution after SD-CHOP induction therapy10 or, in fact, all other induction regimens used during the past decade.14 The limited WBC and platelet engraftment clearly present by 100 days post-ABMT is evidence for injury to both committed and primitive stem cells. Furthermore, the lack of adequate hematologic reconstitution at 1 year and beyond after ABMT suggests significant damage to primitive stem cells in a sizeable subset of these patients, as it is presumably the primitive stem cells that are responsible for maintaining long-term hematopoiesis after transplantation.15-18 

Experimentally, hematopoietic stem cells vary greatly in their self-renewal and proliferative capacity. Most primitive stem cells under steady state conditions are proliferatively inactive. Under normal physiologic conditions, there is a large functional reserve in the primitive stem cell population to prevent marrow exhaustion. Based on animal models, there are a number of situations where apparent stem cell exhaustion or depletion can be elicited.19-23 These include the serial transplantation of marrow into lethally irradiated mice; exposure of marrow to radiation or cytotoxic drugs that damage stem cells; exposure of the marrow to great proliferative stress; and most recently, exposure of marrow to combinations of cytotoxic drugs and cytokines.22,24,25 Therefore, our working hypothesis is that high dose chemotherapy and G-CSF recruited both committed and primitive stem cells into a proliferative phase, and that G-CSF allowed us to retreat patients at a time interval when these stem cells were still susceptible to damage by cytotoxic therapy.

Using the treatment approach described above, it appears that dose escalation and G-CSF support for induction therapy before ABMT resulted in a severe and potentially irreparable toxicity. This marked delay in engraftment to date has not been associated with excess bleeding or infection. To date, five patients who have received SD-CHOP have developed myelodysplasia/acute myeloblastic leukemia. The follow-up in patients who received HD-CHOP is too short to assess this toxicty. Although no clinically relevant toxicity has been observed in patients who received HD-CHOP and G-CSF, our results suggest that one must be prudent in the use of high dose induction therapy and hematopoietic growth factor support before myeloablative therapy. Inadequate hematologic engraftment might be avoided by either shortening the time that growth factor support is administered, administering cytokines that might protect stem cells, lengthening the interval between cycles, or attempting to repetitively harvest additional stem cells either from the marrow or peripheral blood.26 Sufficient numbers of stem cells could potentially be harvested after the first cycle of dose intensified treatment, an approach that has been successfully done for patients with intermediate and high grade lymphomas.27 However, it would be important that a tumor-free stem cell product be obtained through either ex vivo purging and/or stem cell enrichment procedures. Although it is presently not known whether the peripheral blood stem cell compartment would be similarly depleted by multiple cycles of high dose CHOP and G-CSF, there is no reason to believe that it would not. A recent study suggested that stem cell toxins had a greater effect on peripheral blood stem cell progenitors than on BM progenitors.28 Therefore, attention to animal models and markers of hematopoietic stem cell function should be evaluated in all attempts to dose intensify induction therapy with growth factor support.

We are indebted to the nurses, medical oncology fellows, and social workers of the Dana-Farber Cancer Institute; the housestaff of the Brigham and Women's Hospital and Beth Israel Hospital for their excellent care of these patients. We also thank the technicians of the Clinical Immunology Laboratory and the Blood Component Lab for processing of the bone marrows; the oncologic surgeons of the Brigham and Women's Hospital and New England Deaconess Hospital for their surgical assistance; and the staffs of the Departments of Radiology and Infectious Disease for their assistance in patient care.