Cardiovascular diseases (CVD) and cancer are some of the most recognized causes of mortality and morbidity worldwide.1 Heart failure (HF) coupled with cancer contribute to almost 50% of deaths in the middle-age population.2 Despite recent advances in treatment, HF mortality rate remains high.3 The 10-year survival rate stands at only 35%.4 Cancer is the second leading cause of death in the HF population after CVD, accounting for over 15% of all mortality.5,6

Heart failure and cancer have been the subject of much debate in the scientific community. Recent epidemiological and experimental studies have supported the link between these two conditions, suggesting that the relationship between HF and cancer relation may be a two-way street.

In one direction, a vast amount of evidence suggests that cancer patients have a higher risk of developing HF. In fact, the cardio-oncology field has recognized the negative impact of oncological treatment on the heart, demonstrating the cytotoxic contribution of these therapies to the development of HF.7 In the other direction, recent studies have suggested that a diseased heart may promote cancer development and progression.8–16 These two apparently distinct clinical entities share similar risk factors, symptoms, and pathophysiological mechanisms – inflammation, metabolic disturbances, neurohormonal and immune system activation, and endothelial dysfunction and may explain the coexistence of these two diseases.1,7

Further research into this topic is essential and could have a significant impact on clinical practice. The diagnosis of new cancers worsens HF prognosis, impairing patients’ already weak homeostasis. In addition, the incidence of cardiotoxicity caused by oncological treatments is higher, limiting the therapeutic options that can be offered, which contributes to higher mortality.17 Improved acknowledgment of this topic may enable risk stratification in predisposed patients promoting early prevention and optimal management in the era of increasingly personalized medicine. A one size-fits-all treatment approach is ineffective for these patients.18

In recent years, epidemiological studies have been conducted to identify whether patients with HF have an increased risk of incident cancer and its impact on mortality.8,19,20 Additionally, they have investigated the difference in incidence between multiple subtypes of cancer, as well as factors that may influence this relationship, such as age, sex, or number of hospitalizations. Even though there is a previous systematic review21 on the topic, it reports on only four studies and since then, larger, and more detailed studies have been published, some of them with conflicting results. Furthermore, to the best of our knowledge, our review is the first one that gives a site-specific cancer quantitative analysis to understand better which types of cancer may be more involved in this possible association.

In this systematic review and meta-analysis, we aim to analyze and discuss the most recent epidemiological and clinical evidence on new-onset cancer in HF population, exploring cancer incidence and mortality in this population, comparing it with non-HF subjects. We also aim to ascertain whether some types of cancer are more sensitive than others to this possible association, such as lung, breast, prostate, hematologic, skin, gastrointestinal and female reproductive system diseases.

A total of 2401 studies were retrieved using the above methodology. After removing 562 duplicates, 1839 records were assessed for eligibility. Concerning the screening phase, 1777 articles were not eligible for not meeting our inclusion criteria. Therefore, 62 eligible studies were selected for full reading and assessment.

Thirteen observational studies, published between 2013 and 2022, were included in the qualitative analysis and their results were reviewed and discussed (Figure 1). Of these 13 studies, seven assessed cancer incidence and mortality, five assessed only cancer incidence, while one assessed exclusively mortality. The main characteristics and outcomes of these studies have been summarized in Tables 1 and 2.

In general, studies included in the qualitative analysis demonstrated an overall bias of low risk according to National Institutes of Health quality assessment tool. Nine studies were ≥11 in a total of 14 questions and were considered to be good quality, whereas the remaining studies pointed between 8 and 10 in 14 and were considered to be fair quality. The major concerns were related with short timeframe for follow-up, the assessment of outcome of interest more than once or by different methods and the investigator's blindness was not reported in most of the studies (Figure S1 – supplementary material).

Incidence of all-cancer types

Accounting the 12 studies that report a risk estimate for all-cancer type incidence, 8 found a significantly association favoring a higher all-cancer type incidence in the group of patients with prior HF diagnosis, while four suggest no significant association (Table 2).

Of the studies reporting incidence of cancer, nine provided adjusted Cox or Poisson models (HRs or IRRs (incidence risk ratios)) of the risk of developing cancer in HF group comparing with the group of patients without HF. These studies were included in the quantitative analysis, that took in consideration data availability (Table S2 – supplementary material). The studies included in this analysis assessed a total of 7238643 patients (487931 in the HF group and 6750718 in control group).

Of these nine estimates, six reported that HF was associated with significantly increased risk of all-cancer patients, two indicates that HF was associated with non-significantly increased cancer risk and 1 that HF was associated with non-significantly decreased cancer incidence.

The pooled HR for risk of overall cancers in patients with prior HF diagnosis compared with those without HF was 1.30 (95% CI: 1.04–1.62, p=0.02; I2=99%, p<0.001) (Figure 2).

Figure 2.

Forest plot of the random effect models for estimating the pooled overall cancer incidence in patients with HF compared with patients without HF. The diamond represents the combined effect and the respective confidence interval.

(0,33MB).

Incidence by cancer siteLung cancer

Seven cohorts reported results for incidence of lung cancer. All studies showed higher lung cancer incidence in the group of patients with prior HF diagnosis. The pooled HR was 1.87 (95% CI: 1.28–2.73, p<0.001; I2=99%, p<0.001) indicating a significantly increased risk of lung cancer diagnosis in HF patients compared with the group without HF (Figure 3A).

Figure 3.

Forest plot of the random effect models for estimating the pooled cancer incidence by cancer site (A. Lung cancer; B. Women breast cancer; C. Gastrointestinal system cancer (including gastric and colorectal cancers); D. Haematologic cancers (leukaemia, lymphoma, and multiple myeloma); E. Prostate cancer; F. Female reproductive system cancers (ovarian and uterine cancers); G. Skin cancer (melanoma and non-melanoma skin cancers)) in patients with HF compared with patients without HF. The diamond represents the combined effect and the respective confidence interval.

(0,76MB).

Female breast cancer

For incidence of female breast cancer, seven cohorts reported results. Three of the studies indicate a significantly increased risk and four of them reported no significant association. Lam23 did not report an HR estimate, and therefore was only included in the qualitative analyses (Table 2). The pooled HR of the six remaining studies was 1.23 (95% CI: 0.97–1.55, p=0.08; I2=97%, p<0.001) showing that there is no significantly increased risk of breast cancer diagnosis in HF patients compared with the group without HF (Figure 3B).

Gastrointestinal cancers

Seven cohorts reported results for incidence of gastrointestinal system cancer. Five of the studies indicate a significant increased risk and four of them reported no significant association. The pooled HR was 1.22 (95% CI: 1.03–1.45, p=0.02; I2=97%, p<0.001) indicating significantly increased risk of gastrointestinal cancer diagnosis in HF patients compared with the group without HF (Figure 3C).

Hematologic cancers

Six cohorts reported results for incidence of hematologic cancer (leukemia, lymphoma, or multiple myeloma). Five of the studies indicate a significantly increased risk, and two of them reported no significant association. The pooled HR was 1.60 (95% CI: 1.23–2.08, p<0.001; I2=98%, p<0.001) indicating a significantly increased risk of hematologic cancer diagnosis in HF patients compared with the group without HF (Figure 3D).

Female reproductive system cancers

Four cohorts reported results for incidence of female reproductive system cancer (including ovarian and uterus). Three of the studies indicate a significantly increased risk and 1 of them reported no significant association. The pooled HR was 1.67 (95% CI: 1.27–221, p<0.001; I2=92%, p<0.001) indicating a significantly increased risk of female reproductive system cancer diagnosis in HF patients compared with the group without HF (Figure 3E).

Prostate cancer

Five cohorts reported results for incidence of prostate cancer. Only one of the studies indicate a significantly increased risk, three reported a significant decrease in cancer incidence and one indicated no significant association. The pooled HR was 0.97 (95% CI: 0.67–1.20, p<0.001; I2=97%, p=0.46) indicating no significantly decreased risk of prostate cancer diagnosis in HF patients compared with the group without HF (Figure 3F).

Skin cancers

Six cohorts reported results for incidence of skin cancer. Four of the studies indicate a significantly increased risk and three of them reported no significant association. The pooled HR was 1.19 (95% CI: 0.83–1.72, p=0.34; I2=99%, p<0.001) suggesting no significantly increased risk of skin cancer diagnosis in HF patients compared with the group without HF (Figure 3G).

Cancer mortality

Five cohorts reported results for cancer-specific mortality in HF patients compared with patients without HF. All studies found a significant increase in cancer death when these two diseases coexist, except for the Selvaraj24 where the result was not statistically significant (Table S5 – supplementary material). The pooled HR was 2.17 (95% CI: 1.23–3.84, p=0.008; I2=98%, p<0.001) (Figure 4). Additionally, some of the studies found an increased all-cause mortality in HF group.10,12

Figure 4.

Forest plot of the random effect models for estimating the pooled cancer mortality in patients with HF compared with patients without HF. The diamond represents the combined effect and the respective confidence interval.

(0,2MB).

This systematic review and meta-analysis summarize all cohort studies presently available and suggests that there is a significantly increased risk of incident cancer among HF patients. The pooled HR across the nine studies was 1.30 (95% CI: 1.04–1.62, p=0.02), showing that HF patients are at 30% greater risk of being diagnosed with a new malignancy. This estimate was robust across sensitivity analyses performed. We observed that this association increases with age, as expected, and seems to be higher in women.

The association of HF and site-specific cancer risk reached statistical significance for lung, gastrointestinal, hematologic and female reproductive system cancers. However, no significant association was found in breast, prostate, and skin cancers. The cancer types associated with higher risk were lung and female reproductive system, followed by hematologic cancer. This consistent association suggests the shared biological link between HF and certain site-specific cancers.

In sum, there are several potential explanations for the observed association between increased incidence of cancer in patients with pre-existing chronic HF.

First, it is acknowledged that HF and cancer share common risk factors (smoking, aging, genetic predisposition, obesity, and diabetes mellitus), which can explain, at least in part, the co-occurrence of these two conditions, and combating the modifiable factors will inevitably reduce both HF and cancer.7,25,26 Although, even after the adjustment made for some of these factors by the several studies included (Table 2) the correlation seems to persist and might be further explained by other mechanisms.

Second, some common pathophysiological mechanisms are involved in both diseases. It is known that inflammation plays an important role in these two conditions.27,28 Maladaptive chronic inflammation leads to progressive myocardial injury, development of vascular dysfunction and reduces cardiac tissue survival. In fact, independent of its etiology, HF is associated with an increase in circulating proinflammatory cytokines, such as tumor necrosis factor-α interleukin-1, and interleukin-6.29 On the other hand, chronic inflammation predisposes to development of cancer and affects the tumorigenesis and tumor-permissive state by promoting the same proinflammatory cytokines and chemokines.2 In line with this, the CANTOS trial (Canakinumab Anti-Inflammatory Thrombosis Outcome Study) demonstrated that canakinumab, an interleukin-1-targeting antibody, reduces cardiovascular events in patients with a history of myocardial infarction (MI) with a moderate increase in C-reactive protein levels, as well as a reduction in lung cancer incidence in these patients.7,30

Moreover, the metabolic changes caused by toxic intermediate accumulation and the use of unbalanced substrates can disrupt cardiac cell homeostasis and promote cancer growth. In response to pathophysiological insult and stress, metabolic reprogramming occurs as an adaptive event in both cancer and cardiac cells indicating that both cells share the same metabolic pathways.2,31

Angiogenesis adaptations are another common mechanism favoring this linkage. Cardiomyocyte hypertrophy causes a mismatch between capillary density and increased oxygen demand during the early stages of chronic pressure overload, stimulating microvascular expansion by inducing angiogenic factor secretion, such as vascular endothelial growth factor.32 Angiogenesis is critical for tumor growth and dissemination in cancer using dysfunctional tumor vessels to spread throughout the body.18

Also, there is the hypothesis that neurohormonal activation may also account for the increased risk of cancer observed in HF. This is supported by a large body of experimental data demonstrating that sympathetic nervous system (SNS) and renin–angiotensin–aldosterone system activation promote cancer progression and dissemination via multiple mechanisms. The SNS's pro-oncogenic effects are primarily mediated by β-AR (beta arrestin), which is expressed by both cancer cells and, more importantly, non-malignant cells in the tumor microenvironment.33

Third, HF may promote oncogenesis through release of cardiac circulating factors. Relevant preclinical models of CVD, including surgical models of MI and subsequent HF, as well as aortic stenosis, identifying several candidate systemic factors including cardiac-specific circulating factors (e.g., serpinA3, periostin), that drive CVD-induced acceleration of colon, breast, and lung cancer.31,34

Finally, mutual pathogenic gene variants may also predispose individuals to both cancer and HF. Clonal hematopoiesis of indeterminate potential (CHIP), as mentioned above, is an emerging entity in which somatic blood mutations are present in individuals without established hematologic abnormalities. Studies have demonstrated that the presence of CHIP is associated with increased risk of hematologic cancer, CVD, disease progression and mortality in HF.35 Moreover, several genes associated with familial cardiomyopathies appear to be relevant genetic variants in somatic cancer cells. Interestingly, somatic mutations in titin, dystrophin, and desmoglein2 have been associated with different stages of the carcinogenesis.2,36,37

Furthermore, the diagnosis of cancer in HF patients seems to have a strong impact in these patient's prognosis. Most studies reported an increase in cancer-specific death in HF patients (HR 2.17 (95% CI: 1.23–3.84)), as well as an increase in overall mortality, suggesting that cancer is more lethal in HF patients. This rise in mortality can be explained by various factors. On one hand, the impact of the tumor itself on the already weak homeostasis of HF patients can make them prone to the development of more cardiotoxicities during or after antineoplastic therapies and reduce chances to survive oncological surgery, therefore less intensive oncological therapy is applied limiting therapeutic choices. On the other hand, optimization and up-titration of cardiological therapy may be compromised by cancer diagnosis and sometimes represent a significant problem in maintaining an optimal medical therapy for these patients.17,38,39 Also, end-stage cancer patients are often excluded from therapies that would improve their life expectancy, such as devices implantation.38

Even though we think that this phenomenon might be in part attributable to heightened surveillance, it also suggests that patients with HF tend to have hidden diseases of various forms, which could be identified under a close re-evaluation. These results found in the most recent studies suggest that doctors should be more encouraged to consider the possibility of coexisting disease throughout the follow-up. In individuals with HF, prompt diagnosis of de novo cancer can therefore be crucial. In HF patients, clinicians frequently focus on the cardiovascular consequences. Patients with HF have worse outcomes from superimposed cancer than cancer patients or patients with HF alone. The clinical decision can be modified in the presence of the concurrent disease and the decision to invest in some invasive procedures can be rejected in patients diagnosed with cancer at a more advanced stage.

Therefore, HF populations may need greater attention in cancer screening since early cancer detection could lead to better clinical outcomes. Given the various risks associated with each form of site-specific disease, the monitoring may focus particularly on certain types of disease (such as lung cancer). Further research is needed to determine whether an active, focused cancer surveillance program can improve the prognosis for HF, although we must emphasize that the improved prognosis that might be attributable to cancer screening in patients with HF would not help in reducing the social burden, because cancer screenings are not generally recommended at present, and they may not be cost-effective.16 Instead, we suggest that clinicians should always consider the possibility of cancer when examining HF patients; and oncologists should assess baseline cardiovascular risk in patients before starting a cancer therapy that might be related with cardiovascular toxicity.40

The overlap between HF and cancer is considerable and complex. Future research in these fields will likely clarify this relationship, which we believe will help to understand and treat these both lethal diseases better.

In conclusion, our systematic review and meta-analysis revealed that HF may result in a subsequent increase in cancer incidence. These findings support previous studies published in the literature. The cancer types that had a higher incidence in HF patients were lung, female reproductive system, and hematologic cancers. Furthermore, we conclude that having a diagnosis of HF at baseline carries an increased risk of death from cancer, underlining the importance of noncardiac morbidity and of cancer surveillance in the management of HF patients.

However, because of the significant degree of heterogeneity in the studies, the results should be evaluated carefully. More research in this area is required to clarify the contradictory clinical data and identify the underlying mechanisms.

Nevertheless, all things considered, each HF patient who develops cancer has their own unique characteristics, and the clinical and therapeutic management should be personalized to provide the best cardiological and oncological care and to improve their prognosis.

The authors have no conflicts of interest to declare.