Risk prediction model for long-term heart failure incidence after epirubicin chemotherapy for breast cancer - A real-world data-based, nationwide classification analysis

Background: Dilated cardiomyopathy (DCM) incidence during and after anthracycline therapy is highly dependent on anthracycline cumulative dose (CD), but its detailed risk factors remained unexplored. Our aim was to assess heart failure (HF) incidence after epirubicin therapy and construct a HF risk-prediction score. Methods and results: A retrospective study was conducted by anonymized integration of nationwide healthcare databases. All the analysed patients were diagnosed with breast carcinoma confirmed by histology from 2007 to 2016. Participants did not undergo chemo-or radiotherapy or suffer HF/DCM during the preceding 3 years. The HF endpoint was established by assignment of I50 International Classification of Diseases (ICD) codes upon discharge from hospital or issuance of an autopsy report. 8068 patients treated with epirubicin were analysed. The 3-10-year HF cumulative incidence was 6.9%. Using binomial logistic regression the independent predictors were identified. A CD-dependent and significant effect on HF was revealed for epirubicin (threshold dose: 709 mg/m(2), odds ratio (OR): 1.76) and docetaxel (CD: >510 mg/m(2), OR: 1.59; CD <= 510 mg/m(2), OR: 1.28, respectively). HF risk increased with age, even over 40. A risk-prediction score derived from regression coefficients consisting of age, diabetes mellitus, hypertension, coronary artery disease, stroke, epirubicin CD, docetaxel CD, capecitabine, gemcitabine, bevacizumab and cancer stage was able to classify HF risk over a wide range (2-30%). Conclusion: Long-term HF risk for patients treated with epirubicin was stratified by our risk-prediction score with a nearly 15-fold difference between the lowest and highest groups. (C) 2019 Elsevier B.V. All rights reserved.