The use of CFU-GM and CD34+ cell enumeration for assessing harvest quality and factors affecting peripheral blood progenitor cell (PBPC) harvest and engraftment were investigated in 45 women with high-risk and metastatic breast cancer scheduled for dose-intensive cyclophosphamide, thiotepa, and carboplatin (CTCb). PBPC were mobilized with standard breast cancer regimens or cyclophosphamide (1.5 g/m2) and 5 micrograms/kg/day G-CSF and used together with G-CSF for hematopoietic support post-CTCb. There was a significant correlation between peripheral blood CD34+ cells/microliter and harvest CD34+/kg (r = 0.73, p < 0.0001) and between harvest CFU-GM and CD34+ cells/kg (r = 0.5, p < 0.0001). CFU-GM clonogenic assays were of no clinical use beyond that of CD34+ cell enumeration, with the latter allowing for real-time decisions regarding harvesting. Multiple stepwise regression identified the number of prior chemotherapy cycles as the only significant clinical predictor of CD34+ cell yield. For 34 patients proceeding to CTCb with PBPC support, multiple stepwise regression identified as the best predictors for engraftment CFU-GM and CD34+ cells/kg for neutrophils and CFU-GM, CD34+ cells/kg, and the number of prior cycles of chemotherapy for platelets, respectively. A threshold dose of 1 x 10(6) CD34+ cells/kg, obtained in 87% of these heavily pretreated breast cancer patients, was adequate to ensure engraftment within 15 days. There was no significant difference in length of hospital stay or blood product use between patients receiving 1-2.5 x 10(6) CD34+ cells/kg and greater than 2.5 x 10(6) CD34+ cells/kg, although median time to engraftment of neutrophils (9 days versus 8 days, p = 0.007) and platelets (12 days versus 9 days, p = 0.006) was significantly longer. The established threshold of > or = 1 x 10(6) CD34+ cells/kg will allow for more confident consideration of using aliquots of this threshold dose for hematopoietic support in sequential high-dose regimens inclusive of CTCb.