Recombinant human granulocyte colony-stimulating factor infusion and/or autologous peripheral blood stem cell transplantation as a rescue for sequential high-dose combination chemotherapy: a preliminary report.
BACKGROUND: Sequential cycles of combination chemotherapy with high-dose cyclophosphamide, etoposide and cisplatin (sHDCEP) can largely increase the total dose (TD) of drug delivered. If granulocyte colony-stimulating factor (G-CSF) and/or autologous peripheral blood stem cell (PBSC) rescue can shorten the duration of cytopenia between cycles of sHDCEP, the dose intensity (DI) can be increased as well. In order to explore the feasibility of delivering maximal TD and DI by administration of sHDCEP with G-CSF and/or PBSC rescue, this trial is undertaken to investigate the hematologic and nonhematologic toxicity observed with sHDCEP by G-CSF and/or PBSC rescue. METHODS: Patients with refractory malignancy and well preserved physiologic function for whom no available therapy is likely to cure or prolong the survival were eligible for the study. Each cycle of high-dose chemotherapy consisted of: cyclophosphamide 5,000 mg/m2, etoposide 1,500 mg/m2 and cisplatin 150 mg/m2. G-CSF and/or PBSC were administered alternatively after each cycle as rescue for myelosuppression. The next cycle was given to patient who showed response to the previous cycle after recovery from toxicity for a maximal of 4 cycles. RESULTS: Two cases of refractory malignancy with progressive disease were treated by sHDCEP for 7 cycles, including 4 cycles with G-CSF rescue, 2 cycle with PBSC rescue, and 1 cycle with G-CSF + PBSC rescue. In the 4 cycles rescued by G-CSF alone, we observed a slightly slower granulocyte and markedly prolonged platelet recovery in the subsequent cycle. By comparing the effect of G-CSF and/or PBSC rescue on hematologic recovery with the preceding cycle in the same patient, we found that G-CSF rescue provided faster granulocyte recovery than PBSC, but PBSC rescue provided faster platelet recovery than G-CSF. Rescue by larger number of PBSCs provided only faster platelet but not granulocyte recovery than rescue by adding G-CSF to a very small number of PBSCs. However, G-CSF plus the very small number of PBSCs provided shorter duration of both granulocytopenia and thrombocytopenia than rescue by G-CSF alone. The most common nonhematologic toxicity from sHDCEP included transient nausea, vomiting, diarrhea and mild impairment of liver function but we observed no significant or irreversible major organ damage. The side effect from PBSC collection was mild and toxicity from reinfusion of the thawed PBSCs was not obvious. Using G-CSF and/or PBSC rescue, sHDCEP was delivered repeatedly in no more than 4 weeks for the next-cycles except for patient 1 who had cycle 4 delayed because of prolonged platelet recovery by only G-CSF rescue in cycle 3. CONCLUSIONS: Our initial experience has shown that the nonhematologic toxicity from sHDCEP, G-CSF and PBSC rescue was well tolerated. Prolonged platelet recovery after sequential cycles of HDCEP by only G-CSF rescue delayed the next cycle of chemotherapy. Although the next cycle was delivered within 4 weeks by only PBSC rescue, concurrent infusion of G-CSF and larger number of PBSCs should provide the most rapid hematologic recovery. Sequential high-dose chemotherapy administered by this model is likely to provide the maximal delivery of TD and DI, and is worthy of further clinical trials.