論文ID: CJ-19-0279
Background: Heart failure (HF) and cancer (CA) are becoming increasingly prevalent as the population ages. We aimed to evaluate prior history and occurrence of CA and its prognostic impact on HF.
Methods and Results: Consecutive hospitalized HF patients (n=2,103) were divided into 2 groups according to prior history of CA: non-prior-CA group (n=1,828) and prior-CA group (n=275). Compared with the non-prior-CA group, the prior-CA group were older, and had higher prevalence of chronic kidney disease, anemia, and atrial fibrillation (P<0.05). In contrast, sex, other comorbidities, levels of natriuretic peptide and ejection fraction were comparable between groups. We focused on newly diagnosed CA after discharge for HF. In the follow-up period (median 623 days), 114 (6.2%) patients in the non-prior-CA and 17 (6.2%) patients in the prior-CA groups were newly diagnosed as having CA. Additionally, 83 (3.9%) CA-related patient deaths occurred (median 776 days). In the Kaplan-Meier analysis (median 1,037 days), not only all-cause death but also cardiac event rate was significantly higher in the prior-CA group than in the non-prior-CA group (log-rank P<0.01). In the Cox proportional hazard analysis, CA history was a predictor of cardiac event rate (HR 1.450, 95% CI 1.134–1.822), as well as all-cause death (HR 2.483, 95% CI 2.034–3.030).
Conclusions: Prior-CA history was associated with high cardiac event and mortality rates. CA is notable comorbidity in HF patients.
Cardiovascular disease (CVD) and cancer (CA) are the 2 principal causes of death in Western countries.1 CVD and CA share many common risk factors and comorbid conditions.2 A recently focus has been on “onco-cardiology”; that is, patients presenting with cardiotoxicity caused by CA treatment (e.g., chemotherapy, radiation) or CA-associated thrombosis.3–5 On the other hand, clinicians often encounter active CA or a history of CA when dealing with patients with CVD in various clinical settings.6 CA may render patients more prone to developing CVD and heart failure (HF), in particular through many, mostly indirect, pathogenic pathways, including cardiovascular complications of CA therapy, while CVD may affect the clinical course and prognosis of CA patients. As a result, there is a significant interplay between CA and CVD.
Both HF and CA are becoming increasingly prevalent as the population ages.7 On the other hand, although the importance of noncardiac comorbidities (e.g., chronic kidney disease [CKD], anemia, chronic obstructive pulmonary disease [COPD]) and general status (e.g., cachexia,8 nutritional status9) in patients with HF are being increasingly recognized as having a significant effect on prognosis in HF patients, especially in HF with preserved ejection fraction (HFpEF),10,11 the prevalence of CA and its effect on the prognosis of HF have received less attention.12–14 It has recently been reported that HF patients have an increased risk of CA, which persists after the first year of diagnosis of HF, with a worsened prognosis compared with that of CA patients without HF.12,15
Here, we aimed to clarify (1) whether a prior-CA history affects cardiac prognosis and all-cause death in HF patients, and the underlying clinical and pathophysiological parameters (e.g., other comorbidities, anemia, nutrition, inflammation, cardiac function), and (2) the incidence of CA, and (3) CA-related death in HF patients.
This was a prospective observational study in which 2,103 decompensated HF patients, who were discharged from Fukushima Medical University Hospital between 2010 and 2016, were enrolled. The decompensated HF was diagnosed by several cardiologists based on the recent HF guidelines.16,17 Patients who had been admitted because of acute coronary syndrome (ACS) and had undergone hemodialysis were excluded. We divided the 2,103 patients into 2 groups according to prior history of CA: without prior history of CA (non-prior-CA group, n=1,828) and with prior history of CA (prior-CA group, n=275). We reviewed the medical records and defined for each patient the presence of metastasis, CA status, including the medical history of previous and current CA types and treatments, and determined whether they currently had CA, had completed treatment within 5 years, or completed treatment more than 5 years prior. CA types were classified by anatomic site and the primary involved organ, and the date of CA diagnosis was used as the diagnosis date. We defined the CA status according to the timing of treatment: currently has CA, recent CA (within 5 years), non-recent CA (>5 years ago).
As described previously,18,19 we diagnosed several comorbidities, which often coexist and are associated with adverse prognosis in patients with HF. CKD was defined as an estimated glomerular filtration rate <60 mL/min/1.73 cm2 according to the Modification of Diet based on Renal Disease formula.20 Anemia was defined as hemoglobin level <12.0 g/dL in females and <13.0 g/dL in males.17 Atrial fibrillation (AF) was identified by ECG performed during hospitalization and/or from medical records. Presence of COPD was defined as forced expiratory volume in 1 s/forced vital capacity <70% by spirometry according to the Global Initiative for Chronic Obstructive Lung Disease and the American Thoracic Society/European Respiratory Society guidelines.21 Peripheral artery disease (PAD) was diagnosed according to the guidelines using computed tomography, angiography, and/or ankle-brachial index.22 Smoking was determined by medical interview, and defined as patients who had a current or past smoking habit.
We compared the clinical features, laboratory data and echocardiographic findings, and post-discharge prognosis. The patients were followed up until 2018 for cardiac events, occurrence of CA and all-cause death. Fortunately, we were able to follow-up on all patients. Cardiac event was defined as worsened HF and/or cardiac death. Worsened HF was defined as unplanned hospitalization for HF in accordance with the Framingham criteria diagnosed by experienced cardiologists. Cardiac death was classified by experienced cardiologists as death caused by worsened HF in accordance with the Framingham criteria, ventricular fibrillation documented by ECG or implantable device, or ACS. Status and date of death were determined from the patients’ medical records. In cases where medical records were unavailable, status was ascertained by a telephone call to the patient’s referring hospital physician. Those administering the survey were blind to the analyses. The investigation conformed to the principles outlined in the Declaration of Helsinki, the study protocol was approved by the Ethical Committee of Fukushima Medical University, and reporting of the study conformed to STROBE, along with references to STROBE and the broader EQUATOR guidelines.23 Written informed consent was given by all study subjects.
Measurement of Laboratory and Echocardiographic ParametersBlood samples were obtained from all patients at hospital discharge. The B-type natriuretic peptide (BNP) levels were measured using a specific immunoradiometric assay (Shionoria BNP kit, Shionogi, Osaka, Japan), and the high-sensitivity troponin T levels were measured using an electrochemiluminescence immunoassay (Elecsys Troponin T hs, Roche Diagnostics Ltd., Rotkreuz, Switzerland).24 Echocardiography was performed by experienced echocardiographers based on standard techniques.25 The left ventricular ejection fraction (LVEF) was calculated using the Simpson’s method in a 4-chamber view and HF patients were categorized into 3 groups: HF with reduced EF (HFrEF; LVEF <40%), HF with mid-range EF (HFmrEF; LVEF 40–49%) and HF with preserved EF (HFpEF; LVEF ≥50%).26 The RV-FAC, was defined as (end-diastolic area−endsystolic area)/end-diastolic area×100. All measurements were performed using an ultrasound system (ACUSON Sequoia, Siemens Medical Solutions USA, Inc., Mountain View, CA, USA).
Statistical AnalysisCategorical variables are expressed as numbers and percentages. The chi-square test was used for comparisons of categorical variables and was followed by Fisher’s exact test when appropriate. Normality was confirmed using the Shapiro-Wilk test. Parametric variables were presented as mean±SD, and non-parametric variables (e.g., BNP, troponin I and C-reactive protein) are presented as median and interquartile range. Parametric variables were compared using Student’s t-test or analysis of variance. Non-parametric variables were compared using the Mann-Whitney U test or Kruskal-Wallis test. Kaplan-Meier analysis was used for presenting the cardiac event rate and all-cause death, and the log-rank test was used for initial comparisons. The prognostic impact was evaluated by a Cox proportional hazard analysis. To assess potential heterogeneity of associations between prior-CA history and prognosis, because several confounding factors had multicollinearity (e.g., COPD and smoking, diuretics and sodium), we conducted subgroup analyses, but not a multivariate Cox proportional hazard analysis. A multiple Cox proportional hazard model including prior-CA, subgroup factor, and their interaction was fit to obtain the interaction P values. The subgroups were based on the presence/absence of possible confounding factors or the median of other variables. Interactions between prior-CA and clinically relevant variables that are generally known to affect CA or prognosis in HF patients, including age, sex, body mass index, New York Heart Association functional class, systolic blood pressure, heart rate, etiology of HF, hypertension, diabetes, dyslipidemia, CKD, anemia, AF, coronary artery disease (CAD), COPD, stroke, PAD, alcohol, smoking, use of angiotensin-converting enzyme (ACE) inhibitor, angiotensin-receptor blocker, β-blocker, calcium-channel blocker, diuretic, inotropic agent, digitalis, antiplatelet agents, aspirin, anticoagulants, statin, BNP, troponin T, total protein, albumin, sodium, C-reactive protein, LVEF and RV-FAC, were estimated by the Cox proportional hazards regression model. P<0.05 was considered statistically significant for all analyses, except for the subgroup analysis with an interaction P-value of <0.002 in accordance with the Bonferroni correction to avoid type 1 statistical errors. All analyses were performed using a statistical software package (SPSS ver. 24.0, IBM, Armonk, NY, USA).
Of the 2,103 HF patients, 275 (13.1%) had a prior history of CA (e.g., stomach, colorectum, blood lymph, prostate, breast and lung), including 80 with metastasis and 195 without metastasis, 11 current, 120 recent and 144 non-recent CA cases, and 188 underwent surgical operation, 69 underwent chemotherapy and 44 underwent radiation therapy (Figure 1). The types of prior and newly diagnosed CA and cause of CA-related death are presented in Figure 2. The comparison of clinical features between the prior-CA and non-prior-CA groups in the present study is shown in Table. Although age was significantly higher, body mass index was lower, and the prevalence of CKD, anemia, AF, COPD, and stroke were significantly higher in the prior-CA group than in the non-prior-CA group, we found no significant difference in sex, prevalence of ischemic etiology, hypertension, diabetes mellitus, dyslipidemia, CAD, or alcohol and smoking habit between the 2 groups. Regarding medication, the use of calcium-channel blockers was more frequent in the prior-CA group than in the non-prior-CA group. Use of other treatments, including ACE inhibitors, angiotensin II receptor blockers, mineral receptor antagonists, diuretics, inotropic agents, digitalis, aspirin, statins, implantable cardiac defibrillator and cardiac resynchronization did not significantly differ between the groups. In the laboratory data (Table), hemoglobin, total protein and albumin were significantly lower in the prior-CA group than in the non-prior-CA group. In contrast, other laboratory data, including BNP, troponin I, iron, ferritin, sodium, creatinine and C-reactive protein, were comparable between the groups. Echocardiographic parameters (Table) showed no statistical differences between the 2 groups.
Study flowchart. Of 2,103 heart failure (HF) patients, 275 (13.1%) had a prior history of cancer (CA), including 11 current, 120 recent and 144 non-recent cases of CA, and 188 underwent surgical operation, 69 underwent chemotherapy and 44 underwent radiation therapy.
Numbers of each type of cancer (CA): prior-CA at discharge, newly diagnosed CA post-discharge and causes of CA-related death. Prior-CA at discharge were stomach, colorectum, blood lymph, prostate, breast etc. Newly diagnosed CA post-discharge were stomach, colorectum, lung, blood lymph, liver etc. Causes of CA-related death were liver, stomach, blood lymph, lung, colorectum etc.
| Non-prior-CA group (n=1,828) |
Prior-CA group (n=275) |
P-value | |
|---|---|---|---|
| Cancer | |||
| Cancer status (non-recent/recent/currently) (n, %) | – | 223 (81.2)/41 (14.9)/ 11 (4.0) |
– |
| Operation (n, %) | – | 188 (68.4) | – |
| Chemotherapy (n, %) | – | 69 (25.1) | – |
| Radiation therapy (n, %) | – | 44 (16.0) | – |
| Demographic data | |||
| Age (years) | 66.1±14.8 | 73.3±11.3 | <0.001 |
| Male sex (n, %) | 1,124 (61.5) | 165 (60.0) | 0.637 |
| Body mass index (kg/m2) | 23.4±4.3 | 22.6±4.0 | 0.005 |
| New York Heart Association class III or IV (n, %) | 73 (4.0) | 12 (4.4) | 0.771 |
| Etiology of heart failure | 0.065 | ||
| Ischemic | 506 (27.7) | 71 (25.8) | |
| Myopathy | 472 (25.8) | 54 (19.6) | |
| Valvular | 412 (22.5) | 73 (26.5) | |
| Other | 439 (24.1) | 76 (28.4) | |
| Comorbidities | |||
| Hypertension (n, %) | 1,299 (71.1) | 206 (74.9) | 0.187 |
| Diabetes mellitus (n, %) | 740 (40.5) | 114 (41.5) | 0.759 |
| Dyslipidemia (n, %) | 1,323 (72.4) | 193 (70.2) | 0.450 |
| Chronic kidney disease (n, %) | 982 (53.7) | 178 (64.7) | 0.001 |
| Anemia (n, %) | 929 (50.8) | 182 (66.2) | <0.001 |
| Atrial fibrillation (n, %) | 690 (37.7) | 122 (44.4) | 0.036 |
| Coronary artery disease (n, %) | 607 (33.2) | 90 (32.7) | 0.875 |
| Chronic obstructive pulmonary disease (n, %) | 404 (22.1) | 81 (29.5) | 0.007 |
| Stroke (n, %) | 344 (18.8) | 66 (24.0) | 0.043 |
| Peripheral artery disease (n, %) | 211 (11.5) | 33 (12.0) | 0.825 |
| Alcohol (n, %) | 170 (9.3) | 21 (7.6) | 0.371 |
| Smoking (n, %) | 966 (52.8) | 132 (48.0) | 0.134 |
| Treatments | |||
| Angiotensin-converting enzyme inhibitor (n, %) | 860 (47.0) | 140 (50.9) | 0.232 |
| Angiotensin-receptor blocker (n, %) | 589 (32.2) | 100 (36.4) | 0.172 |
| Mineral receptor antagonist (n, %) | 733 (40.1) | 105 (38.2) | 0.545 |
| β-blockers (n, %) | 1,364 (74.6) | 191 (69.5) | 0.069 |
| Calcium-channel blocker (n, %) | 648 (35.4) | 126 (45.8) | 0.001 |
| Diuretics (n, %) | 1,205 (65.9) | 181 (65.8) | 0.974 |
| Inotropic agents (n, %) | 199 (10.9) | 28 (10.2) | 0.726 |
| Digitalis (n, %) | 193 (10.6) | 34 (12.4) | 0.368 |
| Antiplatelet agents (n, %) | 903 (49.4) | 127 (46.2) | 0.320 |
| Aspirin (n, %) | 777 (42.5) | 125 (45.5) | 0.357 |
| Anticoagulants (n, %) | 1,038 (56.8) | 140 (50.9) | 0.067 |
| Statin (n, %) | 711 (38.9) | 100 (36.4) | 0.421 |
| Implantable cardiac defibrillator (n, %) | 223 (12.4) | 42 (15.6) | 0.140 |
| Cardiac resynchronization therapy (n, %) | 175 (9.6) | 29 (10.5) | 0.612 |
| Laboratory data | |||
| B-type natriuretic peptide (pg/mL)§ | 223.5 (77.2–569.3) | 266.7 (110.9–226.7) | 0.766 |
| Troponin I (ng/mL)§ | 0.040 (0.017–0.061) | 0.032 (0.019–0.085) | 0.477 |
| Hemoglobin (g/dL) | 12.6±2.3 | 11.7±2.3 | <0.001 |
| Iron (μg/dL) | 79.6±42.0 | 74.9±38.8 | 0.200 |
| Ferritin (ng/mL) | 177.4±292.5 | 151.9±182.7 | 0.334 |
| Unsaturated iron binding capacity (μg/dL) | 230.9±73.2 | 236.6±73.5 | 0.372 |
| Total protein (g/dL) | 6.9±0.7 | 6.8±0.8 | 0.006 |
| Albumin (g/dL) | 3.7±0.6 | 3.6±0.6 | 0.007 |
| Sodium (mEq/L) | 139.0±3.7 | 139.1±3.8 | 0.899 |
| C-reactive protein (mg/dL)§ | 0.19 (0.06–0.84) | 0.17 (0.05–1.09) | 0.829 |
| Echocardiography | |||
| Left ventricular ejection fraction (%) | 50.7±15.7 | 52.2±13.3 | 0.193 |
| Right ventricular fractional area change (%) | 41.6±14.6 | 40.3±11.4 | 0.336 |
§Data are presented as median (interquartile range). CA, cancer.
We next focused on newly diagnosed CA after discharge for HF, and future CA-related death of hospitalized HF patients. As shown in Figure 1 and Figure 2, during the follow-up period (median 623 days, range 95–2,496 days), 114 (6.2%) patients in the non-prior-CA group and 17 (6.2%) patients in the prior-CA group were newly diagnosed as having CA (e.g. stomach, colorectum, lung, blood lymph, liver). Additionally, as shown in Figure 1 and Figure 2, 83 (3.9%) CA-related deaths occurred among the patients (e.g., liver, stomach, lung, blood lymph and colorectum) within around 2 years (median 776 days, range 145–2,243 days).
In the follow-up period (median 1,037 days, range 15–2,975 days), of the 2,103 patients, there were 595 patients with cardiac events, including 280 cardiac deaths and 315 worsening HF, and 566 all-cause deaths. As shown in Figure 1, there were 280 (13.3%) cardiac deaths, 83 (3.9%) CA-related deaths and 203 (9.7%) noncardiac and non-CA-related deaths (e.g., respiratory failure, renal failure and infection). In the Kaplan-Meier analysis (Figure 3), not only all-cause death but also the cardiac event rate was significantly higher in the prior-CA group than in the non-prior-CA group (log-rank, P<0.01, respectively). We also compared the prognostic impact of CA among the HFrEF, HFmrEF and HFpEF groups (Figure 4). Except for the cardiac event rate in the HFrEF group, cardiac event rate and all-cause death were significantly higher in the prior-CA group than in the non-prior-CA group.
Kaplan-Meier analysis for cardiac event rate and all-cause mortality for the non-prior-CA and prior-CA groups. Cardiac event rate and all-cause mortality were significantly higher in the prior-CA group than in the non-prior-CA group (log-rank, P<0.01, each). CA, cancer.
Kaplan-Meier analysis for cardiac event rate and all-cause mortality for the non-prior-CA and prior-CA groups in HFrEF, HFmrEF and HFpEF groups. In each subgroups, except for cardiac event rate in the HFrEF groups, cardiac event rate and all-cause mortality were significantly higher in the prior-CA group than in the non-prior-CA group. EF, ejection fraction; HF, heart failure (r, reduced; mr, mid-range; p, preserved).
In prior-CA group (n=275), we also compared differences in prognostic impacts of sex, presence of metastasis, CA status and CA treatment (operation, chemotherapy and radiation) in Supplementary Figures 1–4. Cardiac event and all-cause mortality rates in each sex did not significantly differ (Supplementary Figure 1). The metastasis group (n=80) showed a tendency of a higher cardiac event rate, as well as all-cause mortality, than the non-metastasis group (n=195) (Supplementary Figure 2) The cardiac event rate and all-cause mortality in each CA status did not significantly differ (Supplementary Figure 3). Cardiac event rate and all-cause mortality did not significantly differ between the presence or absence of each treatment (Supplementary Figure 4).
In the univariate Cox proportional hazard analysis (Supplementary Figures 5,6), prior-CA was a predictor of both cardiac events and all-cause death (cardiac event, hazard ratio [HR] 1.450, 95% confidence interval [CI] 1.154–1.822, P=0.001; all-cause death, HR 2.483, 95% CI 2.034–3.030, P<0.001). In addition, subgroup analysis revealed an association between prior-CA and (1) cardiac event rate and (2) all-cause mortality rate in the subgroups after adjustment for interactions between prior-CA and prespecified clinically important variables (Supplementary Figures 5,6). There was no significant interaction between the prognostic impact of prior-CA on either the cardiac event rate or all-cause mortality, or with other important variables (e.g., age, sex, other comorbidities, treatment and cardiac function).
In the present study, 13.1% of the 2,103 hospitalized HF patients had a prior history of CA (e.g., stomach, colorectum, blood lymph, prostate, breast and lung). These patients had a higher prevalence of CKD, anemia, AF, COPD and stroke, as well as higher cardiac event and all-cause mortality rates. Also, 6.2% of HF patients, regardless of the presence of prior-CA history, were newly diagnosed as having CA (e.g., stomach, colorectum, lung, blood lymph and liver) after discharge for HF within a median of 623 days. Lastly, 3.9% of HF patients died from CA (e.g., liver, stomach, lung, blood lymph and colorectum) detected prior to HF hospitalization or newly detected CA after discharge, within a median follow-up of 776 days. To the best of our knowledge, the present study is the first to report the detailed associations between CA and cardiac prognosis, as well as all-cause death, while considering the order of diagnosis of both CA and HF in hospitalized HF patients.
Prevalence and Prognostic Impact of CA in CVD PatientsTabata et al recently reported that of 701 patients with CAD who underwent coronary intervention, 16.3% had a CA history, which was associated with increased risk of target lesion revascularization and total cardiovascular events, as well as all-cause death within 1 year.27 In addition, patients with high levels of C-reactive protein currently being treated for CA had a worse prognosis.27 Iannaccone et al recently reported that of 14,475 patients with ACS, 6.4% had a CA history and more frequently presented with ST-segment elevation myocardial infarction, and higher incidence of bleeding events, recurrence of ACS and death within 1 year, compared with ACS patients without CA (bleeding, HR 1.5, 95% CI 1.1–2.1; re-ACS or death, HR 2.1, 95% CI 1.8–2.5).28 Kruijsdijk et al reported that patients with manifest vascular disease (CAD, cerebrovascular disease, abdominal aorta aneurysm, and PAD) were at increased risk of both cardiovascular and CA-related death, particularly patients with PAD or polyvascular disease.29
There are only a few reports regarding the prevalence and prognostic impact of CA on all-cause death, rather than on cardiac prognosis, in HF patients.12–14 Ather et al reported that CA coexisted in 26.6% of HFpEF patients, and was associated with higher all-cause mortality rates (HR 1.24, 95% CI 1.03–1.49), whereas in 18.6% of patients with HFrEF, CA was not associated with all-cause death (HR 1.10, 95% CI 0.98–1.24) in a retrospective study of treated HF patients with a 2-year follow-up period.13 Tribouilloy et al reported that CA coexisted in 12% of HFpEF patients, and was associated with higher all-cause mortality rate (HR 2.46, 95% CI 1.66–3.63); however, in 8.2% of HFrEF patients, CA was not associated with death (HR 1.54, 95% CI 0.92–2.58) in 799 consecutive hospitalized HF patients with a 5-year follow-up period.14 Hasin et al reported that patients with HF who were diagnosed as having CA after HF had a higher risk of death compared with HF patients without CA (HR 1.56, 95% CI 1.22–1.99) within a mean follow-up period of 7.7 years.12 However, the details of CA type, etiology of HF, and prognostic difference related to the order of diagnosis of CA and HF were not considered, and the effect of CA on cardiac prognosis still remains unclear.12–14 In the present study, 13.1% of the patients had prior-CA, which was associated with a higher cardiac event rate in both patients with HFpEF and HFmrEF, but not in HFrEF patients, and was associated with all-cause death in HFrEF, HFmrEF and HFpEF patients. In addition, the prognostic impact of CA on the cardiac event and all-cause mortality rates was consistent across etiologies (i.e., ischemic and non-ischemic) of HF subgroups. Furthermore, the only-prior-CA group had worse prognosis than the non-CA group. We could not explain the reason for this prognostic difference related to the order of diagnosis of CA and HF.
Occurrence of CA in HF PatientsSakamoto et al reported that the occurrence of CA (e.g., stomach, lung, prostate and colorectum) was significantly higher in HF patients than in control subjects among Japanese patients (2.27% vs. 0.59%, P<0.01, 95% CI 1.89–2.71).30 Hasin et al reported that the prevalence of prior-CA was similar in newly diagnosed HF patients compared with controls; however, the occurrence of CA was high in HF patients compared with the controls.12 Over 9,203 person-years of follow-up and a mean follow-up of 7.7 years, HF patients had a higher risk of being diagnosed as having CA compared with the controls, after adjusting for body mass index, smoking and comorbidities (HR 1.68, 95% CI 1.13–2.5).12 In addition, Hasin et al reported that patients who developed HF after ACS had an increased risk of CA than those without HF after myocardial infarction (HR 2.16, 95% CI 1.39–3.35).31 Banke et al reported that HF patients were at increased risk of CA, which persisted after the first year of HF diagnosis among Danish HF patients.15 In a mean follow-up period of 4.5 years, 975 (10.5%) of 9,307 HF patients without a prior diagnosis of CA had experienced several newly identified types of CA (e.g., colorectum, liver/biliary, lung, skin, breast, kidney/urinary tract and lymph/blood), and these incidence rates were significantly higher than those in the background population without HF (incidence rate ratio 1.24, 95% CI 1.15–1.33).15 These HF patients with newly diagnosed CA were associated with older age, prevalence of ischemic heart disease, and use of ACE inhibitors, β-blockers, aspirin and statins.15 Cumulative incidence was significantly higher among HF patients than in the background population in each age group (e.g., <60 years, 60–69 years and ≥70 years).15 Their prognosis was worse than that of CA patients without HF.12,15 Unfortunately, in the current study, no data were available to compare the occurrence of CA in healthy controls.
Underlying Mechanism of the Relationship Between CA and HFWe did not investigate the biological basis for the increased cardiac event rate that we found, but we speculate on the following. First, some types of CA and HF share certain lifestyle-related risk factors, such as hypertension, diabetes, dyslipidemia, COPD, smoking, alcohol, obesity and salt intake, which are associated with the occurrence or worsening of both CA and HF.2 Second, CA itself may have carcinogenic effects (e.g., related to the proliferation of a mutated clone), carcinogenicity and carcinogenesis, hyperviscosity, chronic inflammation through the action of inflammatory cytokines, endothelial dysfunction, coronary atherosclerosis,32 undernutrition state, cachexia,33,34 and anemia, which are also adverse prognostic factors for HF. Third, several CA treatments, including chemotherapy and radiation therapy, may cause cardiotoxicity and could lead to arrhythmia, CAD and worsening HF. Fourth, several medications for HF such as ACE inhibitors15,35–37 and digitalis,38 are assumed to be associated with the occurrence of CA. Although the anti-CA effect of digoxin is suspected in basic science, in a retrospective cohort study with a 10-year follow-up, taking digoxin increased the risk of CA, including breast, liver, and lung CA.38 Although the risk of CA from antihypertensive drugs (e.g., ACE inhibitors, angiotensin-receptor blockers, calcium-channel blockers, β-blockers and diuretics) has been much debated,15,35–37 recent network meta-analyses using 70 randomized controlled trials with 324,168 patients, suggested no evidence of CA increase or CA-related death, except for a combination of ACE inhibitor and angiotensin-receptor blocker, which may increase CA risk.39
Management of CA in HF PatientsThe present study also revealed a new problem regarding the treatment of HF patients with CA. In the aging society, CA patients may frequently experience HF, and HF patients may develop CA. Screening and treatment considering both CA in HF patients and HF in CA patients are important. With regard to the treatment of HF and CA, angiotensin-neprilysin inhibitor is expected to improve CA, as well as HF.40 In addition, it has been recently reported that sodium glucose cotransporter-2 inhibitor decreased not only cardiovascular events,41 but also the occurrence of CA.42,43 Thus, the importance of screening for CA in HF patients and of HF in CA patients, and management of both diseases will become more important.
Study Strengths and LimitationsThere are several strengths to our study. It is the first to show the association of CA with cardiac prognosis, as well as all-cause death in HF patients, taking into consideration the multifaceted background using subgroup analysis and occurrence of CA within a given population (n=2,103). Also, our study included more subjects than previous studies,14,27 and we could follow up all patients.
The present study also has several limitations. First, as a prospective cohort study of a single center, the present results may not be generalizable. Although we performed a subgroup analysis considering several confounding factors as much as possible, we cannot rule out residual confounding variables, and the effects of differences in backgrounds between the 2 groups may not be fully adjusted. However, we would like to emphasize that HF patients with prior-CA experience an adverse prognosis, which is accompanied by several comorbidities. Second, our study had no data regarding CA in the general population without HF. Thus, we could not compare the occurrence of CA between subjects with and without HF. Third, because asymptomatic or less active CA may have been underestimated, there may have been a lead-time bias in diagnosis. Fourth, the present study included only variables during hospitalization for decompensated HF, and we did not take into consideration changes in medical parameters and post-discharge treatment. Fifth, this was a prospective observational study, so the causal relationships and mechanisms of CA with worsened prognosis could not be fully explained. Sixth, we were unable to obtain detailed chemotherapy information. Therefore, the present results should be viewed as preliminary, and further studies with a larger population are needed.
Prior-CA history is associated with higher cardiac event and all-cause mortality rates in hospitalized HF patients.
We thank Ms. Kumiko Watanabe and Ms. Hitomi Kobayashi for their outstanding technical assistance.
This study was supported in part by a grant-in-aid for Scientific Research (No. 16K09447) from the Japan Society for the Promotion of Science.
Dr. Yoshihisa and Dr. Misaka belong to department of advanced cardiac therapeutics supported by Fukuda-denshi Co, Ltd. This company is not associated with contents of this study.
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-19-0279
