WHAT IS KNOWN AND OBJECTIVE: The CHOP regimen with rituximab (R-CHOP) remains the standard for chemotherapy in patients with aggressive non-Hodgkin's lymphoma (NHL). The cardiotoxicity of doxorubicin appears to be a key problem in clinical practice. We studied the cardiotoxicity of CHOP/R-CHOP regimen in a retrospective series. The prognostic factors of congestive heart failure (CHF) were investigated, including the impact of empirical cardioprotection by dexrazoxane. METHODS: Patients with an aggressive NHL between 1994 and 2005 were included. Cardiac events were defined as either a decline in resting left ventricular ejection fraction (LVEF) <50%, a decline in LVEF of ≥20% from baseline or as clinical evidence of CHF. The risk of cardiotoxicity was explored by the Kaplan-Meier method. RESULTS: The study included 180 consecutive patients. During the second period of the survey, cardioprotective therapy by dexrazoxane was administered to 45% of patients. The 5-year cumulative risks of cardiac events (29% vs. 8%) and clinical CHF (17% vs. 1*5%) varied significantly between the two periods of study (1994-2000 vs. 2001-2005). In multivariate analysis, use of dexrazoxane (HR = 0*1 [0*01-0*75], P = 0*02) and age < 60 years (HR = 0*4 [0*17-0*9], P = 0*03) appeared as protective factors of cardiac events. WHAT IS NEW AND CONCLUSION: Our study confirmed the weight of cardiac toxic effect of CHOP ± R regimen. Even if the use of dexrazoxane is highly debatable in curative situations, it may be an effective prevention of cardiotoxicity in aggressive NHL patients.