Purpose: Olanzapine has been shown to be a safe and effective agent for the prevention of chemotherapy-induced nausea and vomiting (CINV). Olanzapine may also be an effective rescue medication for patients who develop breakthrough CINV despite having received guideline-directed CINV prophylaxis. Methods: A double-blind, randomized phase III trial was performed for the treatment of breakthrough CINV in chemotherapy-naive patients receiving highly emetogenic chemotherapy (cisplatin, ≥ 70 mg/m2 or doxorubicin, ≥ 50 mg/m2 and cyclophosphamide, ≥ 600 mg/m2), comparing olanzapine to metoclopramide. Patients who developed breakthrough emesis or nausea despite prophylactic dexamethasone (12 mg IV), palonosetron (0.25 mg IV), and fosaprepitant (150 mg IV) pre-chemotherapy and dexamethasone (8 mg p.o. daily, days 2-4) post-chemotherapy were randomized to receive olanzapine, 10 mg orally daily for 3 days or metoclopramide, 10 mg orally TID for 3 days. Patients were monitored for emesis and nausea for 72 h after taking olanzapine or metoclopramide. Two hundred seventy-six patients (median age 62 years, range 38-79; 43 % women; Eastern Cooperative Oncology Group (ECOG) PS 0,1) consented to the protocol. One hundred twelve patients developed breakthrough CINV and 108 were evaluable. Results: During the 72-h observation period, 39 out of 56 (70 %) patients receiving olanzapine had no emesis compared to 16 out of 52 (31 %) patients with no emesis for patients receiving metoclopramide (p < 0.01). Patients without nausea (0, scale 0-10, M.D. Anderson Symptom Inventory) during the 72-h observation period were those who took olanzapine, 68 % (38 of 56), and metoclopramide, 23 % (12 of 52) (p < 0.01). There were no grade 3 or 4 toxicities. Conclusions: Olanzapine was significantly better than metoclopramide in the control of breakthrough emesis and nausea in patients receiving highly emetogenic chemotherapy. © 2013 Springer-Verlag Berlin Heidelberg.