Mitoxantrone is a conventional agent for relapsed or refractory acute lymphoblastic leukemia (ALL). However, an effective combination with other drugs and a feasible dosage has not been identified. A retrospective study of 46 patients with relapsed or refractory ALL was conducted to determine the efficacy of mitoxantrone and Ara-C treatment with (MEC) and or without etoposide (MC). Twenty-seven and 19 patients received MC and MEC chemotherapy, respectively. Twenty-two (48%) patients showed overall response [complete response (CR), 33%; CR with incomplete platelet recovery (CRp), 15%], and 10 of 22 responders received allogeneic stem cell transplantation (SCT). Median overall survival (OS) was 6.2 months (95% confidence interval, 3.41-9.0). Thirteen (48%) patients in the MC group and 9 (47%) in the MEC group achieved CR/CRp (p = 0.96). Treatment-related mortalities in the MC and MEC groups were 3 (11%) and 4 (21%), respectively (p = 0.36). However, the MEC group frequently presented with grade 3 or higher bacteremia/candidemia (p = 0.013). No difference in OS was observed between the two groups (p = 0.769). In conclusion, salvage therapy consisting of mitoxantrone and Ara-C without etoposide appeared to be an effective bridge therapy to allogeneic SCT for patients with refractory or relapsed ALL.