Recently, several studies have demonstrated that heart failure (HF) may increase the risk of incident cancer. However, this association has not been statistically and systematically verified by any comprehensive pooled analyses. We performed a meta-analysis on cancer morbidity and co-mortality of adults with HF in a large sample size to explore the relationship between HF and the risk of developing cancer. From inception to April 2019, we searched PubMed and EMBASE for published relevant articles on patients with HF diagnosed with cancer afterwards, with reported outcomes of morbidity and mortality. Two investigators independently reviewed these included studies. Study data were independently extracted using predefined data extraction forms. Random and fixed-effects models were fit for the study duration. This analysis consisted of 4 cohort studies comprising 5,004,251 participants. The relative risk (RR) for incident cancer was 1.22 (95% confidence interval (CI), 1.13–1.33) indicating that patients with HF may have a higher risk of developing cancer. The pooled RR of co-mortality was 2.03 (95% CI, 1.13–3.65), indicating that HF associated with cancer increases the risk of mortality. In this meta-analysis and systematic review, our results demonstrated that heart failure may increase the risk of incident cancer and that HF associated with cancer increases the risk of mortality.
Heart failure (HF) and cancer are two major contributors to the burden of mortality accounting for 45.4% of deaths worldwide [22, 26] and will indisputably increase as the population ages, and advanced treatments enhance longevity [13, 27, 28].
Although generally considered as two separate disease entities, HF and cancer commonly coexist, share similar pathological mechanisms, and possess various potential interactions. Cancer survivors carry a greater burden of cardiomyopathy and HF compared with the non-cancer population . The use of cytotoxic chemotherapy, radiotherapy, molecular targeted therapies, and immune-modulating agents affects cardiomyocytes and vascular cells . Currently, cardio-oncology has become expansive in modern translational medical research area, gaining interest at a rapid pace . For all that cardiovascular complications such as potential for myocardial injury of cancer therapy have been well documented , knowledge regarding whether patients with HF have a higher risk of developing cancer is limited . Recent registries and clinical trials observed that a larger percentage of patients with HF die because of non-cardiac causes , and increasing attention has been paid to cancer. A basic study found that controlling risk factors of HF helps reduce the risk of incident cancer . Therefore, a diagnosis of HF therefore may be considered as a risk factor for incident cancer . Recently, three studies demonstrated an incremental morbidity of cancer in patients with HF [3, 11, 12]. However, one study found that HF did not increase the morbidity of cancer .
The association between HF and incident cancer is still controversial. To explore the morbidity and mortality of cancer in patients with HF, we conducted a meta-analysis using cohort studies with a sample size of 5 million population. Such association once validated, while it has profound significance in basic and clinical implications. Furthermore, it is hoped that the results can contribute to tumor-risk classification and treatment strategy.
Search strategy and literature selection
This meta-analysis follows Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement . No prespecified protocol was schemed for this analysis. A comprehensive search of relevant literature was conducted independently by two investigators.
To identify all potentially relevant literatures regarding the topic, a combination of the following terms were utilized to perform this search: “heart failure,” “cancer or tumor or malignancy,” “morbidity,” “mortality,” “relative risk or RR or odds ratio OR hazard ratio,” and “cohort or case control or case-control studies or cohort studies.” Furthermore, we manually retrieved references from reviews and original literatures. Subsequently, bibliographies of published references were searched for additional publications. Only full-text references published in peer-reviewed journals in English were included for this meta-analysis. The search was performed through April 2019 with no lower date limit in the PubMed (https://www.ncbi.nlm.nih.gov/pubmed) and EMBASE databases (https://www.embase.com). Relevant studies that reported morbidity and mortality of cancer in patients with HF were eligible for this meta-analysis. Prospective or retrospective observational studies were included if the following requirements were met: (1) study participations had HF as previously defined; (2) utilization of the current definition of cancer patterns; (3) inclusion of subjects were generally healthy (i.e., not acutely ill); (4) comprised information of defined events could be examined independently of other intervention components; (5) report of adjusted results of incident cancer in HF patients; and (6) risk ratio (RR), odds ratio (OR), hazard ratio (HR), or necessary raw data, and 95% confidence intervals (CIs) were reported.
Titles and abstracts of all retrieved records were independently reviewed by two authors to reject irrelevant articles, and the remaining articles were then assessed by screening the text throughout. Any controversy was settled by consensus or arbitration. After eliminating duplication, the eligibility of each report was assessed at the heading and abstract level, and the full text of the potential eligibility report was reviewed by two independent reviewers. Trials published simply as abstracts and with no additional data available from other sources were excluded. Conflicts were resolved by consensus or by a third reviewer.
Data extraction and quality assessment
The following information was extracted from each cohort independently by two investigators: name of the first author, publication year, geographic location, study design, sample size, demographic and health characteristics, study period, numbers of outcomes, adjustments or matching, morbidity of cancer, mortality, and duration of follow-up. We used the Newcastle-Ottawa Quality Assessment Scale for the quality assessment of the included studies.
Primary outcome: the morbidity of cancer in patients with HF; secondary outcome: the mortality of cancer in patients with HF.
Effect estimates of the morbidity and mortality of cancer in patients with HF were extracted, such as RR, OR, HR, or necessary raw data, and eventually presented as RR with 95% CIs for each trail. P value and I2 statistic were used to evaluate the heterogeneity across studies. A P value < 0.1 and I2 > 50% were indicative of at least moderate statistical heterogeneity. If the heterogeneity was significant, a pooled effect was calculated using a random or fixed-effects model. Statistical significance was defined at P < 0.05. I2 was used as the main evaluation method if the results of the two test methods were inconsistent. Due to the small sample size of the included articles, publication bias could not be evaluated by funnel plot, and the trim and fill method was performed to validate the robustness of results. All analyses were performed using the Review Manager, version 5.0.12 (Revman; The Cochrane Collaboration, Oxford, UK).
Selection and inclusion of studies
Figure 1 shows a flow diagram of this study selection. Two hundred twenty-seven records in the aggregate were identified using our predefined search terms; 63 duplicates were rejected. After screening titles and abstracts, 156 literatures were excluded for the following reasons: guidelines, animal studies, laboratory studies, editorials, review articles, letters, conference abstracts, or irrelevant to our study. Therefore, eight potentially eligible studies were retrieved for a further detailed evaluation. Another four literatures were rejected because they did not report effect estimates of interest (n = 2) and were basic research (n = 2). Finally, four articles were included in our meta-analysis.
Primary characteristics of the four included literatures are shown in Table 1, with 5,004,251 individuals (mean age, 65 ± 10.7 years) from four studies (all cohort studies) incorporated. The mean follow-up duration ranged from 5 to 20 years. The definition of HF and cancer reached a consensus in all studies. All studies were matched or adjusted for at least sex and age. Table 2 shows characteristics of patients included in each study. We further combined mortality findings from the four articles. Quality assessments of the studies are also shown in Table 1. All studies obtained eight stars for quality assessment (Table 1).
Four studies with 5,004,251 participants were included for cancer morbidity. A fixed-effects model was used in the calculation. Patients with HF have an increased risk of incident cancer (RR = 1.22, 95% CI = 1.13–1.33, P < 0.00001; heterogeneity: P = 0.31, I2 = 16%) (Fig. 2).
In the analysis of mortality, HF associated with cancer significantly increased the mortality compared without HF (RR = 2.03, 95% CI = 1.13~3.65, P = 0.02; heterogeneity: P < 0.00001, I2 = 93%). To trace the source of heterogeneity, we used sensitivity analysis by sequentially omitting each study. Unfortunately, we could not find the study that substantially influenced the pooled RR (Fig. 3).
The following are the main findings from this pooled analysis of observational studies: (i) Heart failure increased the risk of incident cancer and (ii) cancer associated with heart failure increases mortality.
To the best of our knowledge, our study is the first meta-analysis to explore the morbidity of cancer in patients with HF and co-mortality. Although we observed an association between HF and incident cancer, the following aspects should be considered when interpreting the outcomes. First, several biases, such as surveillance and self-selection biases, cannot be prevented when data were obtained from observational studies. Second, patients with HF may be more likely to be hospitalized and undergo more clinical tests, which may lead to overestimation of cancer. Third, although all the included literatures claimed that their research subjects were matched or adjusted, the baseline values of the four studies are different, and we could not exclude the bias. Therefore, the findings should be adopted cautiously.
These four studies (all cohort studies) showed the relationship between HF and subsequent cancer. All studies excluded individuals with a prior cancer diagnosis. After adjusting for age, sex, and Charlson comorbidity index, three studies [3, 11, 12] demonstrated that patients with HF had ≥ 60% higher risk of developing cancer. Moreover, the risk of incident cancer was ongoing even after the first stage of diagnosis and medication. However, one study , with a large sample size, found no association. Basic studies found that cardiac-excreted factors may be the causation that cancer is one common comorbidity of HF , and the diagnosis of HF may therefore be taken as a risk factor of cancer . In fact, HF and cancer also possess various similarities and possible interactions , including a number of shared risk factors and common trigger mechanisms. The number of evidence suggesting that inflammation, obesity, oxidative stress, diabetes mellitus, hypertension, smoking, diet, and physical inactivity are all contributors to the development of both heart failure and cancer has been increasing . These risk factors shared a common pathophysiological end point of chronic inflammation, which is associated with both cardiomyopathic and neoplastic processes [7, 8]. Inflammation is an established component of carcinogenesis , and HF is characterized by chronic inflammation and neurohormonal activation [10, 23]. Presumably, chronic inflammation increased the risk of developing cancer in patients with HF. Evidence suggesting that chronic disease patients with an inflammatory component had an increased risk of incident cancer has been accumulating. Type 2 diabetes, breast cancer, colorectal carcinoma, and gastrointestinal tumor have been found to be more continual, suggesting an association with poor prognosis [5, 15, 19]. At the tissue level, it has been demonstrated that inflammatory bowel disease was associated with colorectal cancer that indicates an association between chronic inflammation and cancer . Given the shared risk factors and pathological mechanisms for cancer and cardiac disease, it is not surprising that individuals exposed in risk factors share the same probability of developing cancer.
Because most heart failure was caused by prior myocardial infarction, we considered that MI may be a critical factor. Approximately 20% of patients with HF had a higher frequency of prior myocardial infarction [3, 11], and analysis of comorbidity that showed an increased risk of cancer within the HF cohort revealed previous myocardial infarction , and 12% more of anterior MI in comparison with ST-segment MI. Furthermore, patients with reduced ejection(EF) seemed to have a more prominent risk of having subsequent cancer . However, no statistically significant difference was found in the association of preserved and reduced EF. One study  excluded part of individuals due to a prior MI; hence, data of EF or functional capacity were not uniformly collected.
Also, difference was found in the rate in sex, wherein men with HF were observed to have a 55% increased risk of developing incident cancer compared with men without HF, whereas women with HF had a 71% greater risk . By contrast, HF was more likely to occur in women, with adverse risk factors and larger infarctions , suggesting that sex is a relevant factor. Thus, it is inappropriate that Physicians’ Health Study (PHS) only enrolled men; therefore, the results may not be generalizable to the whole population.
After classifying the types of cancer based on system involvement, we observed a distribution of different types of cancers in the HF and non-HF cohorts. Among patients with HF, the most common malignant diagnoses were respiratory, digestive, and hematologic. However, no statistically significant differences in the associations between HF and cancer subtypes were found.
We also observed that both HF and cancer were strongly associated with all-cause mortality, but patients with both, cancer and HF, demonstrate even worse prognosis than both diseases alone [3, 11, 12]. The 5-year survival estimate for HF cases and non-HF controls was 53% and 77%, respectively . Interestingly, a frequency of lung cancer in HF cohort was also observed. We considered that smoking, a shared common risk factor of these two conditions, may be the cause of this phenomenon. Because the 5-year survival rate after lung cancer only vary from 4 to 17% depending on stage and regional differences , the quantity of lung cancer could possibly lead to the overall increased mortality in HF cohort. Furthermore, tumor load impairs cardiac severely , and cancer treatments, such as chemotherapy and radiation, cause cardiomyocyte injury. This vicious cycle may therefore be the reason of mortality elevation.
Some methodical limitations of this analysis should be acknowledged. Bias cannot be ruled out. First, one study  ( post hoc analysis of randomized controlled trials) included only men, which may bring a gender bias and the interventions may have influenced the results. Second, only a small number of literatures was available for inclusion. A high heterogeneity (I2 = 93%) was calculated among these included studies. Third, because HF, ischemic heart disease, and cancer share risk factors, adjustment for risk factors, i.e., smoking status, alcohol consumption, and body mass index (BMI), would be necessary; unfortunately, this information was unavailable for further analysis. Finally, eligible studies published in other languages might have been missed in this meta-analysis because we included only studies published in English. And due to the small number of included trials (less than ten), it was impossible to conduct a sufficient additional analysis of publication bias. Therefore, a potential publication bias cannot be fully excluded in this analysis.
In conclusion, our systematic review and meta-analysis indicated that HF may lead to an incident cancer afterward and may also increase the co-mortality rate. Considering the limitations of the current analysis, more prospective cohort studies are warranted to evaluate the association between HF and incident cancer.
Angsutararux P et al (2015) Chemotherapy-Induced cardiotoxicity: overview of the roles of oxidative stress. Oxidative Med Cell Longev 2015:795602
Balkwill F, Mantovani A (2010) Cancer and inflammation: implications for pharmacology and therapeutics. Clin Pharmacol Ther 87(4):401–406
Banke A et al (2016) Incidence of cancer in patients with chronic heart failure: a long-term follow-up study. Eur J Heart Fail 18(3):260–266
Blaes A et al (2017) Cardio-oncology related to heart failure: common risk factors between cancer and cardiovascular disease. Heart Fail Clin 13(2):367–380
De Bruijn KM et al (2013) Systematic review and meta-analysis of the association between diabetes mellitus and incidence and mortality in breast and colorectal cancer. Br J Surg 100(11):1421–1429
Deswal A, Basra SS (2013) Incident cancer in patients with heart failure: causation or mere association? J Am Coll Cardiol 62(10):887–888
Diakos CI et al (2014) Cancer-related inflammation and treatment effectiveness. Lancet Oncol 15(11):e493–e503
Dick SA, Epelman S (2016) Chronic heart failure and inflammation: what do we really know? Circ Res 119(1):159–176
Fung BM et al (2019) Cancer risk in primary sclerosing cholangitis: epidemiology, prevention, and surveillance strategies. World J Gastroenterol 25(6):659–671
Gullestad L, Aukrust P (2005) Review of trials in chronic heart failure showing broad-spectrum anti-inflammatory approaches. Am J Cardiol 95(11A):17C–23C discussion 38C-40C
Hasin T et al (2013) Patients with heart failure have an increased risk of incident cancer. J Am Coll Cardiol 62(10):881–886
Hasin T et al (2016) Heart failure after myocardial infarction is associated with Increased risk of cancer. J Am Coll Cardiol 68(3):265–271
Heidenreich PA et al (2013) Forecasting the impact of heart failure in the United States: a policy statement from the American Heart Association. Circ Heart Fail 6(3):606–619
Hirsch FR et al (2017) Lung cancer: current therapies and new targeted treatments. Lancet 389(10066):299–311
Holden SE et al (2015) Glucose-lowering with exogenous insulin monotherapy in type 2 diabetes: dose association with all-cause mortality, cardiovascular events and cancer. Diabetes Obes Metab 17(4):350–362
Hutton B et al (2015) The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. Ann Intern Med 162(11):777–784
Koene RJ et al (2016) Shared risk factors in cardiovascular disease and cancer. Circulation 133(11):1104–1114
Libby P (2006) Inflammation and cardiovascular disease mechanisms. Am J Clin Nutr 83(2):456S–460S
Lin HC et al (2014) Effects of metformin dose on cancer risk reduction in patients with type 2 diabetes mellitus: a 6-year follow-up study. Pharmacotherapy 34(1):36–45
Lund LH et al (2014) Association between cardiovascular vs. non-cardiovascular co-morbidities and outcomes in heart failure with preserved ejection fraction. Eur J Heart Fail 16(9):992–1001
Meijers WC et al (2018) Heart failure stimulates tumor growth by circulating factors. Circulation 138(7):678–691
Murphy SL et al (2017) Deaths: final data for 2015. Natl Vital Stat Rep 66(6):1–75
Reina-Couto M et al (2014) Impaired resolution of inflammation in human chronic heart failure. Eur J Clin Investig 44(6):527–538
Selvaraj S et al (2018) Lack of association between heart failure and incident cancer. J Am Coll Cardiol 71(14):1501–1510
Springer J et al (2014) Prevention of liver cancer cachexia-induced cardiac wasting and heart failure. Eur Heart J 35(14):932–941
Thavendiranathan P, Nolan MT (2017) An emerging epidemic: cancer and heart failure. Clin Sci (Lond) 131(2):113–121
Torre LA et al (2016) Global Cancer incidence and mortality rates and trends-an update. Cancer Epidemiol Biomark Prev 25(1):16–27
Voors AA et al (2017) Development and validation of multivariable models to predict mortality and hospitalization in patients with heart failure. Eur J Heart Fail 19(5):627–634
The work was supported by the National Key Research and Development Program of China (Grant Nos. 2017YFC1700400 and 2018YFC1704901) and the National Natural Science Foundation of China (Grant Nos. 81725024 and 8143009)
Conflict of interest
The authors declare that they have no conflict of interest.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Zhang, H., Gao, Y., Wang, L. et al. Does heart failure increase the risk of incident cancer? A meta-analysis and systematic review. Heart Fail Rev 25, 949–955 (2020). https://doi.org/10.1007/s10741-019-09876-0
- Heart failure
- Cancer morbidity
- Systematic review