Cancer Is a Comorbidity of Heart Failure

Graphical Abstract

(Enlarge Image)

Possible causes of the association of HF with cancer and ensuing clinical implications

In the last 5 years, growing attention has been paid to the occurrence of cancer in individuals who already have heart failure (HF).^[1]

The issue of cancer in HF was further addressed by Bruhn and colleagues in a study that is now published in this issue of the European Heart Journal. By using Danish nationwide registers, they evaluated the temporal trends in new-onset cancer between 1997 and 2016 in 103 711 subjects with HF, aged 30–80 years and without a history of cancer within 10 years before the index date.^[2] In contrast to what they hypothesized, they found that the 5-year incidence rates of cancer were relatively constant throughout the study period, ~20 per 1000 person-years, in the landscape of progressively decreasing age at the index date, cardiovascular (CV) mortality, and all-cause mortality. The results were consistent in sensitivity analyses and after considering several possible confounders, such as age, sex, and concomitant ischaemic heart disease. Notably, the 5-year risk of death from cancer among HF patients in Denmark also slightly declined between 1997 and 2016.

The authors are to be commended for performing a rigorous analysis of such a large population. In particular, they noted that the incidence of cancer was highest during the first year after the diagnosis of HF, suggesting a detection bias, i.e. detection of cancer is more likely than in later phases because medical evaluation is more frequent than subsequently. Therefore, they chose to set the index date to start the observation for cancer incidence 1 year after the diagnosis of HF.

The findings by Bruhn et al. are in good agreement with the prior observation that cancer mortality, as a proxy of cancer incidence since this latter was not recorded, was quite similar across 15 HF randomized controlled trials (RCTs) spanning 29 years, while CV and total mortality declined.^[3] In contrast, another group reported progressively higher cancer mortality in the face of lower CV mortality in HF patients, but they examined a relatively small, single-centre cohort.^[4] Remarkably, several HF RCTs demonstrated that cancer is the largest contributor to non-CV death.^[5]

The work by Bruhn et al. also has limitations, which are intrinsic to retrospective, community-based analyses and are largely acknowledged by the authors themselves. Methodological differences may indeed explain the discrepancy between the present analysis and previous ones, which described a higher risk of cancer in patients with HF than in controls in community-based cohorts from Europe, including Denmark.^[6,7] For instance, the International Classification of Diseases codes and other criteria to define HF and cancer, and the timing and duration of the follow-up greatly varied from one study to another. Furthermore, cancer may be found at a different frequency in HF with reduced vs. preserved ejection fraction, but this distinction was never considered because of the unavailability of left ventricular ejection fraction values.

Nonetheless, most epidemiological investigations indicate that cancer intersects the trajectory of HF more than in the past. Bruhn et al. argue that this trend is not the consequence of a more frequent diagnosis; rather, they reason that the substantial reduction in CV and global mortality makes it more likely that, over years, cancer complicates the history of patients with HF (Graphical Abstract).

HF may also be associated with cancer mechanistically, in a direct or indirect manner. A first, seminal study demonstrated that mice with an adenomatous polyposis coli mutation developed more and larger tumours when HF due to myocardial infarction (MI) was inflicted.^[8] Secreted factors were identified as the cause of this phenomenon. HF secondary to transverse aortic constriction induced breast and lung cancer growth.^[9] Moreover, in a breast cancer model, HF enhanced circulating Ly6Chi monocytes and their recruitment to the tumour site, which in turn restricted T-cell infiltration and antitumoural immune responses.^[10] More recently, however, no differences were observed in an orthotopic renal cancer model after MI or sham surgery.^[11] Hence, HF promotes some types of cancer in the experimental setting, and might also do so in the human situation.

Cancer initiation may be fuelled by the same risk factors as HF, such as smoking, diabetes, and obesity.^[12] Furthermore, events that are implicated in HF pathogenesis can also drive cancer formation and progression. Inflammation is a prime example in this regard.^[13] Interestingly, activation of inflammation-related intracellular signalling factors has been pinpointed as the reason why clonal haematopoiesis of indeterminate potential, a condition that predisposes to haematological malignancies, can lead to HF^[14] (Graphical Abstract).

It is not currently possible to disentangle the relative contribution of risk factors and biological pathways vs. changes in HF epidemiology to the association of HF with cancer. In perspective, the elucidation of the mechanistic links between HF and cancer may allow the identification of cancer biomarkers specific to HF patients, as well as novel therapeutic approaches for cancer in HF.

The work by Bruhn et al. also stimulates questions.

Is cancer screening performed in individuals with HF as systematically as it is done in the general population? This should be the case, given the number of contacts with the healthcare system that patients with HF typically have. Yet, it is possible that screening for some cancer types is only partially pursued in subjects who suffer from HF. Measurement of prostate-specific antigen concentrations might be neglected because prostate cancer is rarely the cause of death in men with comorbidities, hence the risk–benefit ratio of the diagnostic work-up and treatment in the absence of overt symptoms may be viewed as unfavourable by HF specialists.^[15] Similarly, colonoscopy for screening of colorectal cancer might be omitted, since it entails a preparation with fluid loss that can be troublesome in HF. All these suppositions need to be verified. Nonetheless, a better understanding of the use of cancer screening tools in the HF population is urgently needed.

Bruhn et al. also discuss the role of cancer in RCTs in HF. Incident cancer is normally not adjudicated in HF trials. However, RCTs could generate information on the outcomes of patients with HF who have cancer during follow-up, which would be complementary to that obtained from epidemiological studies. Moreover, data from RCTs would clarify the safety and efficacy of HF therapies in the subgroups with newly diagnosed cancer.

Finally and most importantly, the article by Bruhn et al. leaves the clinician with the confirmation that cancer is a comorbidity of HF to be aware of (Graphical Abstract). According to the Danish registers, the 5-year risk of ischaemic stroke, transient ischaemic attack, or systemic thrombo-embolism in men aged 70 years with atrial fibrillation (AF) was 6.8%, and in those with one risk factor in addition to AF it was 8.7%.^[16] In women of the same age, the corresponding risks were 8.2% and 9.1%, respectively. The 5-year risk of incident cancer in the HF cohort examined by Bruhn et al. was 9.0%. In pragmatic terms, therefore, the practitioner should be as concerned about cancer in HF as much as about ischaemic complications of AF.

Nevertheless, it remains to be established whether the management of HF should be modified when cancer is diagnosed, and this is the most clinically relevant task to be accomplished by the field of cardio-oncology focusing on cancer in HF.

Eur Heart J. 2023;44(13):1133-1135. © 2023 Oxford University Press
Copyright 2007 European Society of Cardiology. Published by Oxford University Press. All rights reserved.