BACKGROUND: The purpose of this investigation was to: (1) determine incidence and predictors of mitoxantrone-induced early cardiotoxicity and (2) study left ventricular mechanics before and after receiving mitoxantrone. METHOD AND RESULTS: We retrospectively analyzed 80 subjects diagnosed with acute myeloid leukemia (AML) who underwent chemotherapy with bolus high-dose mitoxantrone. Echocardiographic measurements were taken at baseline and at a median interval of 55 days after receiving mitoxantrone. Thirty-five (44%) of the patients developed clinically defined early cardiotoxicity, 29 (36%) of which developed heart failure. There was a significant decrease in the ejection fraction (EF) not only in the cardiotoxicity group (17.6 +/- 14.8%, P < 0.001) but also in the noncardiotoxicity group (5.3 +/- 8.4%, P < 0.001). Decrease in global longitudinal strain (GLS) (-3.7 +/- 4.5, P < 0.001 vs. -2.4 +/- 4.3, P = 0.01) and global circumferential strain (GCS) (-5.6 +/- 9, P = 0.003 vs. -5.3 +/- 8.7, P < 0.001) was significant in both the cardiotoxicity and noncardiotoxicity group, respectively. A multivariate model including baseline left ventricular end-systolic diameter, baseline pre-E/A ratio, and baseline pre-E/e' ratio was found to be the best-fitted model for prediction of mitoxantrone-induced early clinical cardiotoxicity. CONCLUSION: High-dose mitoxantrone therapy is associated with an excellent remission rate but with a significantly increased risk of clinical and subclinical early cardiotoxicity and heart failure. Mitoxantrone-induced systolic dysfunction is evident from reduction in EF, increase in Tei index, and significant reduction in GLS and GCS. Baseline impaired ventricular relaxation evident from higher E/e' ratio and lower E/A ratio independently predicts increased risk of mitoxantrone-induced early cardiotoxicity.