Adding Insult to Injury: Heart Failure and Incident Cancer

This article refers to ‘Incidence of cancer in patients with chronic heart failure: a long-term follow-up study’, by A. Banke et al., published in this issue on pages 260–266 Heart failure (HF) is a clinical syndrome associated with a multitude of co-morbidities such as diabetes, chronic kidney disease, COPD, anaemia, and cachexia.1-4 More recent research has focused on device therapies,5 catheter-based valve repair,6 iron deficiency,7 and skeletal muscle wasting.8-10 Cancer, however, has not been very high on the list of important contributing factors in HF. This is noteworthy for several reasons. As early as 15 years ago, the overall prognosis of HF had been found comparable with or even worse than that of some malignant cancers.11 While cancer survival rates have improved only modestly in recent years,12 the introduction of novel therapies for HF has had significant effects on overall survival rates, and new drugs such as the neprilysin inhibitor sacubitril are being developed and introduced into routine management of HF.13 With improved HF survival, the likelihood of receiving a cancer diagnosis increases for all patients with HF. When it comes to the interaction between cardiology and oncology, however, research and clinical interest have stayed rather focused on HF development as a result of cardiotoxic chemotherapy. Risk stratification tools such as the measurement of troponins have been developed to identify patients at risk of developing HF during chemotherapy,14 and screning algorithms have been suggested in addition to treatment regimens.15, 16 Still, the evidence has not been judged sufficient to render guideline-based recommendations possible. Few studies have been undertaken to decipher the interplay between the tumour and the cardiovascular system in an approach to understand whether or not cancer itself, independent of chemotherapy, may have detrimental effects on the cardiovascular system. One study in patients with colorectal cancer recently suggested that this is indeed the case, as cancer patients had significantly reduced heart rate variability and exercise capacity, and mildly reduced LVEF.17 Such changes were already prevalent in chemotherapy-naive patients. In this issue of the journal, Banke et al. report incidence rates of cancer among 9307 patients with newly diagnosed HF in Denmark between 2002 and 2009.18 They describe that during a mean follow-up of 4.5 years, 975 of these patients also received a new diagnosis of cancer, and that the most common cancer types were those of the lung and the skin. The fact that the median time from the diagnosis of HF to the diagnosis of cancer was 862 days gives rise to the suspicion that a longer follow-up may have significantly altered the incidence rate. Risk factors associated with incident cancer in the HF cohort were previous myocardial infarction and the use of ACE inhibitors. Neither factor comes as a surprise. Antihypertensive drugs have repeatedly been proposed to be involved in cancer development, but several large studies have recently refuted this hypothesis. For example, a large network meta-analysis including >320 000 patients could not detect any difference in the risk of cancer associated with the use of ARBs, ACE inhibitors, beta-blockers, calcium channel blockers, or diuretics.19 The only suspicion for increased risk of cancer that remained in this analysis was for a combination therapy of ACE inhibitor and ARB. However, this therapeutic approach has, a least for HF, largely been abandoned in recent years. Other workers have, already 40 years ago, argued that high blood pressure, a co-morbidity frequently observed among patients with HF, may itself be a risk factor for cancer at some sites.20 This point also leads to exploring other co-morbidities of HF-associated risk factors. Previous myocardial infarction as a risk factor for cancer of the lung has to be viewed in the context of smoking as a risk factor for both. Since ischaemic heart disease may predispose for the development of HF, this avenue of thought can easily be followed. Banke et al. compared their findings on incident cancer rates in patients with newly diagnosed HF with those of the Danish population aged 18 years or older. The rates of newly diagnosed cancer were always higher in patients with HF than in the background population. This remained true even after adjusting for age. The authors themselves discuss the problem of surveillance bias, simply because patients with a significant cardiovascular disease such as HF are under constant medical surveillance, making it more likely that they would receive any new diagnosis than subjects not under such care, as in the background population. In addition, we do not know a lot about the medical therapy of the background population, for example with regards to ACE inhibitor use, and whether or not there is an overlap in the risk factors idenfied for HF patients. The fact that mortality rates are higher in cancer patients with HF than in those who do not have cancer also requires further scrutiny. The authors argue that their background population may have had other types of cancer, simply due to having a different age range, but there may also be more or less aggressive chemotherapy approaches in patients who do not vs. those who do have HF. In addition, there might be a competing death risk, particularly in elderly subjects, when aggressive cardiovascular treatment leads to a longer survival time and thus a longer time actually to diagnose cancer.21 Be that as it may, Ann Banke and co-workers have helped to establish cancer as a co-morbidity of HF worth looking for. Thus, a broader picture of symptoms should be considered when a HF patient complains, again, about increasing shortness of breath or fatigue that cannot be reconciled with improving echo images or natriuretic peptide levels. With increasing survival rates in our patients, this is an undertaking that will ultimately pay off. Conflict of interest: none declared.