Cardiotoxicity of breast cancer treatment: no easy solution for an important long-term problem

This editorial refers to ‘Prevention of cardiac dysfunction during adjuvant breast cancer therapy (PRADA): a 2 × 2 factorial, randomized, placebo-controlled, double-blind clinical trial of candesartan and metoprolol’†, by G. Gulati et al., on page 1671. Breast cancer is the most common cancer in women in the developed world. Therapeutic options for patients with breast cancer have improved, and many patients with early breast cancer receive adjuvant systemic therapy including anthracyclines, taxanes, trastuzumab, hormonal therapy, and radiotherapy. Unfortunately, these effective treatments are associated with adverse cardiovascular effects, both acute and late-onset.1,2 Prolonged follow-up is needed to estimate the effect of late cardiac damage on ultimate outcome. Heart failure (HF) is one of the most severe cardiac side effects of systemic adjuvant therapy.3 Anthracyclines are widely used to treat breast cancer. Its anticancer effects take place through the inhibition of topoisomerase leading to blockade of DNA. Although very effective, its use is mostly limited by its cardiac side effects. The incidence of HF in contemporary trials involving adjuvant doxorubicin is ∼1%, but this incidence rate increases when anthracyclines are followed by trastuzumab treatment (see also a recent review4). However, this might be an underestimation since screening has only been performed with relatively insensitive diagnostics (i.e. left ventricular ejection fraction at rest),5 and morbidity and mortality for cancer limited follow-up duration and may have concealed the cardiac effects. Although the initial insult to the myocardium most probably takes place during treatment, as revealed by the acute release of troponins during treatment,6 HF may only become evident years after the chemotherapy. A strong predictor of cardiotoxicity is the cumulative dose of anthracycline; however, there are large interindividual differences.1 In addition, interaction with radiotherapy and the presence of cardiovascular risk factors such as hypertension do play an important role. From clinical practice it is known that some patients can withstand higher dosages of anthracyclines with no effect on cardiac function, whereas in others a low dose of anthracyclines already leads to a severely depressed left ventricular function. Whether traditional HF treatment targeted at the renin–angiotensin–aldosterone system (RAAS) and/or sympathetic nervous system (i.e. beta-blockers) may prevent cardiotoxicity is not really clear. In the present issue of the journal, the investigators of the ‘Prevention of cardiac dysfunction during adjuvant breast cancer therapy’ (PRADA) study prospectively studied the effect of the beta-blocker metoprolol and the angiotensin receptor blocker (ARB) candesartan in patients with breast cancer and no prior cardiac disease, scheduled for adjuvant systemic therapy including anthracyclines (i.e. epirubicin).7 The primary endpoint of this 2 × 2 factorial double-blind placebo-controlled study was change in left ventricular ejection fraction (LVEF) measured by magnetic resonance imagimng (MRI), shortly after completion of systemic treatment. There was no interaction between the treatment groups and, consequently, the data for candesartan and metoprolol were analysed separately. Patients in the control group (placebo) had a 2.6% decrease in LVEF at the end of the study, and this was somewhat prevented by candesartan (–0.8%) (P = 0.026 between groups). No effect of candesartan was observed on any of the other endpoints including global longitudinal strain (GLS), diastolic LV function, or the biomarkers brain natriuretic peptide (BNP) and troponin. Metoprolol did not influence the decline in LVEF during treatment; however, a small but significant improvement in diastolic function (E/E') was observed with this drug. Both treatments were well tolerated without serious adverse events. In the last decade, only a few small sized randomized studies have been reported which investigated the role of RAAS inhibition and beta-blockers in patients receiving cardiotoxic cancer treatment. The largest of them (until now) was a placebo-controlled randomized trial studying the effect of enalapril in 114 patients who received high dose chemotherapy.8 In this single-centre trial, only high-risk patients were recruited. High risk was defined as troponin release during chemotherapy. The authors reported a large effect of RAAS inhibition: in 43% of the control group LVEF decreased >10% at 1 year after completion of treatment, whereas in none of the enalapril-treated patients was this observed. Comparable results were observed in a Turkish study evaluating the role of spironolactone in patients simultaneously treated with anthracyclines.9 Patients randomized to spironolactone did not exhibit a reduction in their LVEF, whereas LVEF decreased by 14% in the control group. For beta-blockers, similar, favourable results have also been reported; however, this is based on only two small trials with ∼50 patients. Both trials showed a preservation of LVEF in patients treated with nebivolol and carvedilol, respectively, compared with placebo.10,11 Possible effects of ARBs on cardiac unloading, myocardial remodeling or other pleiotropic effects. ARB, angiotensin receptor blocker; ROS, reactive oxygen species. Why did the PRADA study show only small effects with an ARB on LVEF, and remarkably, no effect on diastolic function, and why was there a discrepancy between the LVEF and peak systolic GLS? Compared with LVEF, GLS is regarded a better parameter to assess myocardial contractility, especially since it is considered load independent. Candesartan will most probably reduce peripheral resistance and therefore may lead to an increase in LVEF even prior to cancer treatment (in particular since in the candesartan group ∼10% of patients are hypertensive compared with 2% in the placebo group). To understand whether this may be the case, it would be of interest to know the LVEF data of the different groups after the first dose of chemotherapy and after the final dose, shortly after completion of therapy. Could it be the wrong timing of drug administration? In a recent observational study, the chances for complete recovery after a decline in LVEF after anthracycline use were largest if treatment with RAAS inhibition and beta-blockers was started within 2 months after completion of chemotherapy. In that sense, the current trial was optimal, since treatment with candesartan and metoprolol was initiated at the start of the systemic cancer treatment, although follow-up after completion of treatment was rather short. Another question is whether ARBs are the optimal drugs to prevent cardiotoxicity. The mechanisms for anthracycline-induced cardiac injury are largely unknown, but the generation of intracellular reactive oxygen species (ROS), inhibition of topoisomerase 2b, and mitochondrial iron accumulation are thought to play a role.12 In this respect, the use of the iron chelator dexrazoxane has indeed, in several studies, been shown to reduce the incidence of new-onset HF, without an assumed compromising cancer treatment effect.13 Interestingly, it appears that derazoxane may also have a direct effect on topoisomerase 2b—a potential missing link in the explanation of the cardioprotective effects of dexrazoxane. How does an ARB fit into this concept? In experimental settings they reduced oxidative damage in vascular endothelium, possibly by acting as a free radical scavenger;14 whether this is also the case in cardiomyocytes is currently unknown. Although no difference in troponin release was observed in the PRADA study, it has to be noted that this was based on only two time points. Whether the small effects of candesartan on LVEF are related to cardiac unloading, myocardial remodelling, or that pleiotropic effects play a role is unknown (Figure 1). Did the PRADA study examine the wrong population? The population studied in PRADA consisted of patients with no prior cardiac disease and few cardiovascular risk factors. Greater treatment benefit may have been achieved if only high-risk patients would have been targeted or patients with pre-existing cardiac conditions.8 In a recent meta-analysis on the effects of beta-blockers, angiotensin-converting enzymes, and ARBs, a significantly better LVEF was observed in the treatment groups compared with placebo (64% vs. 57%, respectively).15 This effect was most prominent in patients treated with a high cumulative dose of anthracyclines, again indicating that high-risk patients in particular benefit from intervention. Clearly longer follow-up and more work is needed to identify high-risk patients and to better understand the large interindividual effects of anthracyclines on the heart. Post-hoc studies from PRADA should focus on the potential long-term effects and on the group of patients with troponin release, and investigate whether beta-blockers and/or ARBs may have an effect in patients susceptible to cardiac damage, potentially leading to patient-tailored medicine. In conclusion, cancer drug-associated HF is a substantial health problem, and a better understanding of the pathophysiology may lead to the design of specific, targeted drug therapies. In addition, larger trials are needed with longer follow-up to identify beneficial treatment strategies for these patients. Conflict of interest: none declared.