To evaluate the interest in assessing left ventricular diastolic function at baseline for prediction of trastuzumab-mediated cardiotoxicity (TMC) in the setting of adjuvant treatment for breast cancer. The study included 118 women presenting with HER2-positive early-stage invasive breast cancer. Patients received trastuzumab therapy over 1 year, concurrent with six cycles of docetaxel (n = 53), or following anthracycline-based chemotherapy with a cumulative dose of 300 mg/m(2) (n = 45) or 600 mg/m(2) (n = 20) of epirubicine. RNA was performed before anthracycline-based chemotherapy, before trastuzumab treatment (baseline), and every 3 months during treatment. Left ventricular ejection fraction (LVEF) and peak ejection rate (PER) were calculated to evaluate LV systolic function; peak filling rate (PFR), and time to peak filling rate (TPFR) were also calculated to evaluate LV diastolic function. Eighteen patients (15%) developed grade 1 or 2 TMC during follow-up. No significant difference was observed for age, cardiovascular risk factors, fasting blood glucose level, heart rate, systolic blood pressure, baseline LVEF, PER, and PFR between patients with and without TMC. In contrast, patients with TMC showed a longer TPFR at baseline (median [Q1-Q3]: 165 ms [149-190] vs. 142 ms [130-162]; P < 0.001). Furthermore, by logistic regression analysis, baseline TPFR >180 ms and the cumulative dose of epirubicin remained independent predictors of TMC. Patients receiving 600 mg/m(2) of epirubicin before trastuzumab showed a higher incidence of TMC (35%) than did both patients who previously received 300 mg/m(2) of epirubicin (13%) and those who received only docetaxel associated with trastuzumab (9%). Impaired left ventricular diastolic function before treatment is an independent predictor of trastuzumab-mediated cardiotoxicity. The evaluation of diastolic function could allow optimal risk stratification before the introduction of trastuzumab.