Portal de Periódicos da CAPES

Epirubicin and long‐term heart failure risk in breast cancer survivors

2018; Elsevier BV; Volume: 20; Issue: 10 Linguagem: Inglês

10.1002/ejhf.1215

1879-0844

Syed Mahmood, Ravi B. Patel, Javed Butler, Muthiah Vaduganathan,

Cancer-related cognitive impairment studies

This article refers to 'Long-term effect of epirubicin on incidence of heart failure in women with breast cancer: insight from a randomized clinical trial' by A. Banke et al., published in this issue on pages 1447–1453. The advent of anthracycline therapy in the 1960s for cancer treatment was followed by observations of dose-dependent cardiotoxicity and heralded the establishment of the sub-specialty of 'cardio-oncology'.1 In the treatment of breast cancer, anthracyclines are expected to continue to be part of cancer therapy regimens for the foreseeable future with epirubicin and doxorubicin being the preferred agents in Europe and the United States, respectively.2 Successes in breast cancer treatment have resulted in nearly 80% of patients surviving at 15-year follow-up3; as such, cardiovascular (CV) surveillance and disease prevention will be a cornerstone of cancer survivorship amongst breast cancer patients.2 The short-term incidence of heart failure (HF) after very high-dose epirubicin exposure (∼900 mg/m2) may exceed 5–10%, but substantial cardiotoxicity may occur at lower doses depending on background co-morbid disease burden and individual susceptibility.4, 5 The long-term CV sequelae of epirubicin treatment is less established and has largely been informed by observational studies. Clinical trials have been limited to smaller sample sizes with short follow-up and have traditionally focused on left ventricular ejection fraction as a surrogate marker of cardiotoxic risk. We discuss strategies of ascertainment of high-quality data regarding chemotherapy-related risks of HF in future clinical trials of cancer therapies. The Danish Breast Cancer Cooperative Group (DBCG) 89D trial randomized relatively young (average age 48 years) Nordic women with breast cancer to treatment with epirubicin (median dose 452 mg/m2) in the experimental arm or methotrexate in the control arm, with both study arms receiving concomitant cyclophosphamide and 5-fluorouracil.6 The trial was completed in 1998, but extended follow-up has been facilitated by data linkage with Danish administrative health records. In an earlier post-hoc analysis at median 12-year follow-up, 26 total cases of symptomatic HF were identified but no difference in HF risk was observed between the study arms, after accounting for variation in ischaemic heart disease and adjuvant treatment.7 Seventy-seven surviving study participants were also randomly selected to undergo echocardiography and there was no difference in left ventricular ejection fraction between the epirubicin and control arms (66% vs. 67%).7 In this issue of the Journal, Banke and colleagues undertake an exploratory analysis of the DBCG 89D cohort after an additional 5 years of follow-up (median follow-up now of 17 years).8 At this later follow-up, the authors identified six additional (32 total) cases of symptomatic HF corresponding to a low incidence of symptomatic HF at 3.7 per 1000 patient-years in the epirubicin arm, which was higher than the incidence of 1.4 per 1000 patient-years in the control arm. The authors conclude that the number needed to harm to cause one additional case of symptomatic HF was 29, which was nearly twice as high as the number needed to treat of 18 to save one life with epirubicin therapy.8 Consistent with the original DBCG 89D trial publication in 20076 of improved breast cancer survival with epirubicin-based chemotherapy regimens, the relatively low observed long-term incidence of symptomatic HF associated with epirubicin will therefore be reassuring to both patients and clinicians. The original DBCG 89D trial did not include protocolized cardiac imaging for HF surveillance and instead relied upon investigator-reported signs and symptoms, which were not different between study arms.6 Banke et al.8 leveraged International Classification of Diseases and Related Health Problems (ICD)-based Danish national registries to retrospectively identify Danish participants of DBCG 89D who experienced symptomatic HF. ICD codes for HF in administrative databases have high specificity (97–99%) in correctly identifying HF based on a meta-analysis of 11 studies.9, 10 At the same time, it is likely that rates of HF are underestimated as the sensitivity of HF detection in Danish administrative databases has been reported to be only 29%.10 As acknowledged by the authors, their event estimates miss the significant proportion of patients who may have developed left ventricular systolic dysfunction without overt symptoms or who died of sudden death outside the medical system. The low incidence of symptomatic HF observed in long-term follow-up of DBCG 89D may also be related to the low relative CV risk of the enrolled cohort and variable application of potentially cardiotoxic combination chemotherapies. Age at time of breast cancer diagnosis is one of the greatest independent determinants of subsequent CV death. Indeed, in a population-based cohort study of nearly 100 000 patients with early-stage breast cancer,11 incidence of CV death at 10 years was much lower for patients who were diagnosed with breast cancer at age less than 66 years compared with those who were older at diagnosis (0.5% vs. 8%).11 Each additional decade of age at start of epirubicin therapy is associated with a nearly 30% increased rate of cardiotoxicity.5 An additional factor contributing to the low observed symptomatic HF risk relates to the low rates of CV and non-CV co-morbidities in the DBCG 89D trial cohort. In fact, the cumulative baseline prevalence of hypertension, coronary artery disease, and atrial fibrillation was only 1–2%.8 Presence of CV risk factors at the time of breast cancer diagnosis is associated with higher 10-year CV death incidence (17% vs. 2%).11 In fact, CV death incidence exceeds that of breast cancer death incidence among breast cancer patients who have a history of CV risk factors before cancer diagnosis and have survived past 5 years since cancer diagnosis (17% vs. 15%).11 The incremental HF risks of combination cancer therapies have increasingly been recognized in recent years. It is uncertain how many DBCG 89D participants who experienced recurrence were treated with trastuzumab, a therapy known to have additive and synergistic risk of HF to that of the anthracycline class. In the original analysis at 10-year follow-up, ∼40% of participants had breast cancer recurrence.6 It is probable that at least a portion of these participants with recurrence received trastuzumab, since the DBCG 89D study recruitment ended in 1998, and in 2000 trastuzumab was approved by the European Medicines Agency as a breast cancer therapy. However, the original DBCG 89D study did not include human epidermal growth factor receptor 2 status to better understand candidacy for trastuzumab. In treating individual patients, several international expert opinion-based cardio-oncology surveillance guidelines exist for post-anthracycline monitoring.2 The European Society of Medical Oncology (ESMO) suggests transthoracic echocardiography at 1 and 5 years following anthracycline therapy if patients receive higher-dose anthracyclines (∼400 mg/m2 of epirubicin) or experienced cardiotoxicity requiring treatment.12 In the conduct of clinical trials of cancer therapies, there is a lack of a uniform, comprehensive, evidence-based, and clinically meaningful metric for detecting chemotherapy-associated cardiotoxicity.7, 13 For instance, classifying HF grade using common terminology criteria for adverse events can involve five different domains and is subject to inter-investigator variability; and the sole reliance on changes in left ventricular ejection fraction may not necessarily have clinical relevance.13 There is an enduring need for dedicated adjunctive CV safety programmes with centralized adjudication committees to accurately capture CV clinical endpoints, including HF, in emerging cancer trials for several reasons.13 First, many chemotherapy regimens in contemporary oncology have life-prolonging benefits, and CV adverse effects may be dose- or therapy-limiting. Second, new clinically-important CV toxicities are becoming identified with a broader range of cancer therapies.14 Third, relying on administrative health records may miss a significant proportion of patients who may ultimately develop chemotherapy-related cardiotoxicity. Fourth, similar CV safety programmes have been successfully employed in the evaluation of commonly used therapies (for instance, in diabetes mellitus and gout), and have yielded new knowledge regarding preventing CV events in these areas. Finally, as the population of patients being evaluated and treated for breast cancer becomes older with greater medical co-morbidities and as they are treated with combination chemotherapy regimens (with or without concomitant radiation therapy), competing cardiotoxic risks will become increasingly important to identify, manage, and potentially prevent. Cross-disciplinary collaboration between oncologists and cardiologists will continue to be necessary to develop effective therapies in cancer medicine with rigorous surveillance infrastructure to detect and appropriately manage potentially cardiotoxic adverse events. Conflict of interest: S.S.M. has been supported by the Sarnoff Cardiovascular Research Foundation. J.B. has received research support from the NIH and European Union and has been a consultant for Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, CVRx, Janssen, Luitpold Pharmaceuticals, Medtronic, Merck, Novartis, Relypsa, Vifor Pharma, and ZS Pharma. M.V. is supported by the NHLBI T32 postdoctoral training grant (T32HL007604). R.B.P. has no conflicts of interest to declare.