BACKGROUND: There is great interpatient variability in the number of peripheral blood stem cells collected, as measured by CD34+ cell content, alter the administration of chemotherapy and a growth factor. The ability to predict patients who fail to yield adequate quantities of CD34+ cells would be of value. However, very few reports include large numbers of patients treated in an identical fashion. STUDY DESIGN AND METHODS: Between 1991 and 1995, 497 consecutive patients with a variety of malignant diseases received cyclophosphamide (4 g/m(2)), etoposide (600 mg/m(2)), and granulocyte-colony-stimulating factor (6 mu g/kg/day) for mobilization and collection of a target dose greater than or equal to 2.5 x 10(6) CD34+ cells per kg. Multivariate analyses were performed to determine the factors associated with failure to achieve this target harvest. RESULTS: A median of 14.71 x 10(6) CD34+ cells per kg (range, 0.08-137.55) was harvested with a median of 2 (range, 1-11) apheresis procedures. Ninety-one percent of patients yielded greater than or equal to 2.5 x 10(6) CD34+ cells per kg. Patients with Stage II-III breast cancer, who had pretreatment platelet counts greater than or equal to 150 x 10(9) per L and patients who underwent II prior chemotherapy regimen had improved CD34+ cell yields. However, most patients with adverse risk factors yielded greater than or equal to 2.5 x 10(6) CD34+ cells per kg. CONCLUSION: A regimen of cyclophosphamide, etoposide, and granulocyte-colony-stimulating factor led to the successful collection of adequate numbers of CD34+ cells in most patients without excessive toxicity. These observations confirm previous reports that intense prior therapy adversely affects the quantity of CD34+ cells harvested. Pretreatment and posttreatment variables did not predict with any certainty the small fraction of patients who fail to yield greater than or equal to 2.5 x 10(6) CD34+ cells per kg via multiple apheresis procedures.