Background G-CSF is a novel therapy for ischemic heart disease (IHD), but its safety and efficacy in acute or ‘no option’ refractory IHD (R-IHD) remains unclear. Also, the few completed studies generally did not adapt G-CSF to IHD patients but merely administered G-CSF according to usual hematological practice. To address these issues we evaluated the safety and efficacy of G-CSF therapy in 20 patients (18 males, 2 females, mean age 62.4) with R-IHD. Three approaches were incorporated to enhance the efficacy of G-CSF in R-IHD: Repeated, controlled myocardial ischemia was induced during G-CSF therapy to activate myocardial chemokine expression as a homing signal for mobilized angioblasts; Two cycles of G-CSF were administered to enhance the likelihood of clinically meaningful angiogenesis; During the second cycle of G-CSF an intracoronary infusion of CD133+ cells was performed. Methods After baseline cardiac assessment (CA) (angina questionnaire, exercise stress test (EST), nuclear Sestamibi (MIBI) and dobutamine stress echocardiographic (DSE) imaging), patients received open-label G-CSF commencing at 10μg/kg s/c for 5 days, with an EST on days 4 and 6 (to facilitate angioblast trafficking). After 3 months CA and the same regimen of G-CSF +ESTs were repeated but in addition, leukapheresis and a randomized double blinded intracoronary infusion of CD133+ or unselected cells were performed. Final CA was 3 months thereafter. Results Eight events fulfilled pre-specified adverse event (AE) criteria, but despite the poor outlook and co-morbidities of R-IHD patients there were no deaths or AEs resulting in long-term sequelae. Administration of 2 cycles of G-CSF resulted in step-wise improvements in angina frequency, quality of life, EST performance and Duke treadmill score (all p < 0.005). Pre-specified MIBI and DSE results trended towards improvement (p > 0.1), while a composite post hoc computation indicated improved ischemia/perfusion compared to baseline (p = 0.025). Intracoronary infusion of CD133+ or unselected cells was safe but neither enhanced the response to G-CSF. Conclusions Administering G-CSF to R-IHD patients may be performed safely and shows promise as a therapy for debilitating angina. A phase II/III study of this therapy is warranted.