The Breast Cancer, Early Disease: Toxicity from Therapy with Epirubicin Regimens – Cardiac Assessment and Risk Evaluation (BETTER-CARE) Study : design, set-up and analysis of a prospective multicentre collaboration to investigate the pharmacogenetics of anthracycline cardiotoxicity, using cardiovascular magnetic resonance imaging

• Main Author: Kotwinski, P. J.
• Published: University College London (University of London) 2013
• Subjects: 610
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.626031

Anthracyclines are highly effective chemotherapy drugs. However, their use is associated with anthracycline cardiotoxicity (AC), which may result in premature heart failure. Risk of AC is related to cumulative dose, but is also highly idiosyncratic. Effective tests to predict susceptibility are lacking: serial measurement of left ventricular ejection fraction (LVEF) detects cardiotoxicity only after significant damage has occurred. I hypothesised that: (i) Genetic variation underlies susceptibility to cAC. (ii) Risk of cAC could be predicted after the 1st dose of anthracycline by measuring oxidative stress (urinary F2-Isoprostanes, uF2IP), and (iii) also by cardiovascular magnetic resonance imaging (CMR) tissue characterisation (EGRE and STIR). I tested these hypotheses in a prospective gene-environment interaction study, whereby 164 women underwent measurement of LVEF, before and >12 months after anthracyclinechemotherapy. Sixty subjects participated in the 1st dose sub-studies. The mean change in LVEF at follow-up (ĢLVEF) was -2.2%. Thirty-four subjects (20.7%) experienced a fall in LVEF .5% (the cAC group). LVEF fell by 0.8% in the remaining 130 (minimally affected controls). An a priori tiered genetic analysis was performed. High probability variants (Tier 1) were identified (i) by cross referencing human GWAS and gene-expression / proteomic data from animal models of AC (N=12) (ii) from associations with AC reported by others (N=4). Significant associations were discovered with CLCNKA rs10927887 (OR 4.7 [2.1-10.5]) and PDE4D rs1588265 (OR 3.3 [1.51-7.37]), p=0.003 and p=0.045 respectively (Bonferroni corrected). There were correlations between ĢLVEF and percentage change in uF2IP (r=0.30, p=0.04), EGRE (r=-0.35, p=0.01) and STIR (r=-0.29, p=0.03). My results support all 3 hypotheses, and suggest that it may be possible to develop tests to identify individuals at increased risk of cAC and premature heart failure before the majority of the damage is caused. Recognition of this susceptibility could inform treatment decisions and /or identify those requiring greater cardiac surveillance.