Abstract
Background
Cardiovascular diseases are associated with higher cancer risk. However, their relationship with metastatic cancer, the primary determinant of cancer prognosis, has not been studied.
Objectives
This study aimed to determine the association between atherosclerotic cardiovascular disease and the presence of metastasis at the time of cancer diagnosis.
Methods
We analyzed data from 21,654 self-referred adults who were free of cancer and atherosclerotic cardiovascular disease at enrollment in a preventive health care program. To exclude silent cancers, a 1-year blanking period was implemented at the start of the follow-up. The relationship between atherosclerotic cardiovascular disease and metastatic cancer was assessed using cause-specific Cox regression, treating incident atherosclerotic cardiovascular disease as a time-dependent covariate. Interaction analysis further elucidated differences in metastasis risks between middle-aged adults (Q1-Q3 age ≤54 years) and older adults (Q4 age >54 years).
Results
Over a median follow-up of 6 years (Q1-Q3: 3-12 years), we recorded 1,333 cases of atherosclerotic cardiovascular disease (6.2%) and 1,793 cases of cancer (8.3%), of which 1,036 (4.8 %) were nonmetastatic and 757 (3.5%) were metastatic at diagnosis. After adjusting for shared risk factors, atherosclerotic cardiovascular disease was independently associated with an increased risk of cancer metastasis at the time of cancer diagnosis (HR: 1.75; 95% CI: 1.33-2.29). This association was more pronounced among middle-aged adults (HR: 1.64; 95% CI: 1.03-2.61; P = 0.036) than in older adults (HR: 1.11; 95% CI: 0.78-1.60; P = 0.56), with a significant interaction (Pinteraction = 0.039).
Conclusions
Atherosclerotic cardiovascular disease is associated with a significantly increased risk of cancer, specifically metastasis at the time of cancer diagnosis, particularly in middle-aged adults. Recognizing this association could enhance the prevention and treatment of metastatic cancer in patients with atherosclerotic cardiovascular disease.
Key Words: atherosclerosis, cancer, cardiovascular diseases, ischemic heart disease, metastasis, myocardial infarction, reverse cardio-oncology, stroke
Central Illustration
A growing body of research has explored the intersection between cardiovascular disease and cancer.1, 2, 3, 4, 5, 6, 7, 8 Both cardiovascular disease and cancer have numerous complex and overlapping risk factors,4,8, 9, 10, 11 including genetic, environmental, and lifestyle influences.8,9,11 Additionally, cardiovascular disease may lead to systemic inflammation and immunomodulation, potentially facilitating tumor growth8,12, 13, 14 by releasing protumorigenic factors from the diseased heart.13,15, 16, 17, 18, 19 Despite these insights, the exact mechanisms detailing how cardiovascular disease may accelerate cancer progression remain complex and poorly understood. The specific link between cardiovascular disease and cancer metastasis—a crucial determinant of malignancy and patient outcomes—remains unexplored. Metastasis not only marks the progression of disease but also influences disease outcomes.
In this study, we aimed to investigate whether atherosclerotic cardiovascular disease (ASCVD) is associated with an increased incidence of metastasis at cancer diagnosis. Given that both cardiovascular diseases and metastatic cancer are leading global causes of morbidity and mortality, recognizing and understanding their relationship could open unique opportunities for preventive interventions and targeted treatments. Such insights have the potential to significantly improve outcomes for patients affected by both conditions.
Methods
Study population
The study population comprises self-referred adults who were free of cancer and ASCVD at baseline. This cohort and its baseline characteristics have been described in prior studies.20,21 Participants were recruited from preventive health care settings where they presented as private individuals (Figure 1). They were assessed through questionnaires covering demographic characteristics and medical history. Subsequently, all selected participants underwent comprehensive health evaluations including blood tests, blood pressure measurements, standard physical examination by a physician, and an electrocardiogram. Additionally, each participant completed a standard exercise stress test according to the Bruce protocol. The study protocols and procedures were approved by the Institutional Review Board of the Chaim Sheba Medical Center (approval number SMC-4451-17).
Figure 1.
Workflow of Subject Selection and Analysis
This flowchart outlines the selection process and criteria for excluding participants from the study. ASCVD = atherosclerotic cardiovascular disease.
Inclusion and exclusion criteria
The database comprised records of 36,688 individuals screened between January 2000 and December 2018. The inclusion criteria for the current study were 1) having had a minimum of 2 doctors’ visits and 2) having a valid national identification number for cross-referencing with the INCR (Israel National Cancer Registry). Exclusion criteria included any history of cancer, hematologic malignancy, presence of atherosclerotic cardiovascular disease, or missing baseline data regarding cancer or ASCVD (Figure 1).
Study definitions and endpoints
Data on cancer diagnosis during the study follow-up were obtained from INCR, a national passive registry that collects reports on newly diagnosed cancer cases from all health care providers nationwide.22 New cancer cases were coded according to the International Classification of Diseases for Oncology, 3rd edition and the classification guidelines of the Surveillance, Epidemiology, and End Results program summary staging manual.23 The database was last updated on December 31, 2018.
The primary endpoint of this study was the occurrence of metastasis at the time of cancer diagnosis as defined by the Surveillance, Epidemiology, and End Results program as follows: “tumor cells that have broken away from the primary tumor, have traveled to other parts of the body, and have begun to grow at the new location.”23
A new diagnosis of atherosclerotic cardiovascular disease, defined as myocardial infarction or stroke, was identified from the medical records maintained by the primary health care providers and the Sheba Medical Center database. These records were coded using International Classification of Diseases-Ninth Revision codes. All new diagnoses were recorded in the medical summaries during annual visits. All diagnoses were adjudicated by 2 study investigators who were blinded to the subjects’ characteristics.
For each patient, a 10-year ASCVD risk score was calculated based on the American College of Cardiology/American Heart Association guidelines for the estimation of future cardiovascular disease risk in primary care settings.24 The ASCVD risk score calculation takes into account age, sex, race, systolic blood pressure, total cholesterol (mg/dL), high-density lipoprotein concentration (mg/dL), a history of diabetes mellitus, smoking status, and hypertension treatment. An elevated ASCVD score was defined as 7.5% or greater.24
The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation.25 Anemia was defined as hemoglobin <13 for men and <11.6 for women. For cardiorespiratory fitness,20,21 we first computed age-specific and sex-specific distributions of the metabolic equivalents from treadmill exercise tests. Participants were then categorized into 2 groups based on fitness quintiles: low fitness (quintiles 1-3) and high fitness (quintiles 4-5). Additionally, age groups were defined by age quartiles at baseline, distinguishing between middle-aged adults (Q1-Q3) and older adults (Q4).
Statistical analysis
Continuous data are presented as mean ± SD or median with 25th and 75th percentiles (Q1-Q3). Categoric data are presented as count (percentage). We assessed the cumulative incidence of ASCVD using Gray’s method, considering death as a competing risk.
The HRs and 95% CIs for incident metastatic and nonmetastatic cancer were evaluated using a cause-specific Cox proportional hazards regression model. To decrease the likelihood of cases of silent cancer, a 1-year blanking period was implemented at the start of the follow-up (Figure 1). Subsequent analysis included both univariable and multivariable cause-specific Cox regressions, targeting incident metastatic and nonmetastatic cancer as outcomes while considering death as a competing risk.26 The proportional hazards assumption was tested using Schoenfeld residuals (Supplemental Table 1).
Follow-up commenced at the initial visit for all participants, with ASCVD treated as a time-varying covariate. The model also included several continuous variables, including an elevated ASCVD baseline score, chronic kidney disease (eGFR <60 mL/min/1.73 m2), body mass index (BMI >30 kg/m2), anemia, and low-density lipoprotein (LDL) concentration (mg/dL). The selection of parameters for adjustments in the multivariate model was based on clinical relevance, which was determined before analysis. Mortality across all patients was analyzed using a Cox proportional hazards regression, with ASCVD, metastatic cancer at diagnosis, and nonmetastatic cancer included as time-varying covariates.
For matched landmark analysis, propensity scores were first calculated to adjust for the following variables: age, sex, BMI, smoking, diabetes mellitus, hypertension, and length of follow-up. Next, each patient with ASCVD was matched to a 1:1 ratio to a patient without the disease using a 0.1 caliper of the SD of the propensity scores (Supplemental Table 2).27
Subsequent analysis used Gray’s method to assess the cumulative incidence and determine the subdistribution HR (sHR) along with the 95% CI. In this analysis, ASCVD was treated as a landmark rather than a time-dependent variable, with nonmetastatic cancer and death as competing risks. The follow-up for each participant began at their first visit or, for those without ASCVD, at the time of diagnosis.
For sensitivity analysis regarding the blanking period, the primary analysis was repeated without this 1-year blanking period (Supplemental Methods, Supplemental Table 3). All P values were computed as 2-tailed, and a P value <0.05 was considered statistically significant. All statistical analyses were performed with SPSS software v28.0.1.1 (IBM Corp). Propensity scores and analyses using Gray’s method, including its graphics, were calculated using STATA BE v17 (StataCorp). Additional graphics were generated using Graph Pad Prism v9.5.0 (GraphPad Software).
Data availability
All requests for raw and analyzed data and related materials will be reviewed by Sheba Medical Center’s legal department to verify whether the request is subject to any confidentiality obligations.
Results
The final cohort comprised 21,654 subjects, all free from cancer and ASCVD at baseline (Figure 1, Table 1). The median age was 46 years (Q1-Q3: 40-53 years). Of these, 16,241 (75%) were classified as middle-aged (age quartiles 1-3, Q1-Q3: 39-48 years), and 15,138 (70%) were men (Figure 1, Table 1).
Table 1.
Characteristics of Study Population
| All (N = 21,654) | Middle-Aged Adults (n = 16,241) | Older Adults (n = 5,413) | |
|---|---|---|---|
| Age, y | 46 (40-53) | 43 (39-48) | 59 (56-63) |
| Follow-up, y | 6 (3-12) | 6 (2-12) | 7 (3-13) |
| Male | 15,138 (70) | 11,236 (69) | 3,902 (73) |
| Smoking | 3,625 (17) | 2,965 (18) | 660 (13) |
| High fitness | 8,661 (40) | 6,659 (41) | 2,002 (37) |
| Weight | |||
| Underweight | 132 (0.6) | 121 (0.6) | 10 (0.2) |
| Normal weight | 10,392 (48) | 5,470 (44) | 1,882 (36) |
| Overweight | 7,458 (34) | 6219 (38) | 2536 (46) |
| Obesity | 2,754 (13) | 1,857 (12) | 879 (16) |
| BMI | 25.8 ± 4 | 26 ± 4 | 27 ± 4 |
| Missing BMI data | 918 (4) | 812 (5) | 106 (2) |
| Atherosclerotic cardiovascular disease score >7.5 | 3,829 (18) | 843 (5) | 2,986 (56) |
| Comorbidities | |||
| Atrial fibrillation | 422 (2) | 136 (1) | 286 (5) |
| Diabetes mellites | 1,566 (7) | 733 (5) | 833 (16) |
| Impaired fasting glucose | 2,409 (11) | 1,351 (8) | 1,058 (20) |
| Hypertension | 4,899 (23) | 2,405 (15) | 2,494 (47) |
| SBP, mm Hg | 122.4 ± 16.7 | 119 ± 15 | 132 ± 18 |
| DBP, mm Hg | 76.9 ± 10.7 | 76 ± 11 | 80 ± 10 |
| eGFR <60 ml/min/1.73 m2 | 884 (4.1) | 212 (1) | 672 (13) |
| eGFR ≥60-<90 mL/min/1.73 m2 | 13,534 (62) | 9,526 (58) | 4,008 (75) |
| eGFR ≥90 mL/min/1.73 m2 | 7,224 (33) | 6,563 (40) | 661 (12) |
| Missing eGFR data | 38 (0.2) | 30 (0.2) | 8 (0.1) |
| Creatinine | 1.00 ± 0.18 | 0.98 ± 0.18 | 1.04 ± 0.18 |
| Anemia | 5,156 (24) | 4,169 (26) | 987 (19) |
| Hemoglobin | 14.5 ± 1.3 | 14.4 ± 1.3 | 14.7 ± 1.2 |
| HDL | 48.5 ± 12.5 | 48.2 ± 12.3 | 49.3 ± 13.1 |
| LDL | 122 ± 28 | 121 ± 28 | 126 ± 28 |
| Total cholesterol | 192.0 ± 34.2 | 190 ± 34 | 198 ± 34 |
| Triglycerides | 124.7 ± 70.4 | 123 ± 72 | 130 ± 66 |
Values are median (Q1-Q3), n (%), or mean ± SD. Clinical characteristics of the final cohort (N = 21,654) are divided into middle-aged adults (age quartiles 1-3) and older adults (age quartile 4).
BMI = body mass index; DBP = diastolic blood pressure; eGFR = estimated glomerular filtration rate; HDL = high-density lipoprotein; LDL = low-density lipoprotein; SBP = systolic blood pressure.
Overall, the incidence of ASCVD within the study population was low (Table 1). Compared with older adults (age quartile 4, Q1-Q3: 56-63 years), middle-aged adults had a lower prevalence of key risk factors such as hypertension, diabetes mellitus, obesity, and elevated ASCVD risk scores. Additionally, fitness levels were higher among middle-aged adults, but they also showed a higher prevalence of smoking (Table 1).
Incidence of ASCVD, cancer, and metastatic cancer
During a median follow-up of 6 years (Q1-Q3: 3-12 years), ASCVD was diagnosed in 1,333 (6%) of the subjects. Among the middle-aged adults (16,241 participants), 562 (3.4%) developed the disease, whereas in the older adult group (n = 5,413), 771 (14.2%) were diagnosed.
Gray’s competing risk analysis, with death as a competing risk, showed that the cumulative probability of being diagnosed with ASCVD at 6 years was significantly lower in middle-aged adults (2.1% ± 0.2%) compared with older adults (10.4% ± 0.4%) (P < 0.001). Additionally, unadjusted Cox regression analysis showed that compared with middle-aged adults, older adults were 4.2 times more likely to develop ASCVD during the follow-up period (95% CI: 3.28-4.72; P < 0.001).
The same model showed that, beyond age, an ASCVD risk score >7.5% was the strongest predictor of developing ASCVD (Supplemental Figure 1A). Additional risk factors included chronic kidney disease, obesity, and impaired fasting glucose (Supplemental Figures 1B to 1F). Overall, these results confirmed the association between traditional risk factors and the onset of the disease.
Next, we examined the variables associated with the development of nonmetastatic cancer within our cohort. During the follow-up period, cancer was diagnosed in 1,793 (8.3%) patients, 1,036 (4.8%) of whom had nonmetastatic cancer. Variables associated with nonmetastatic cancer at diagnosis included chronic kidney disease, impaired fasting glucose, obesity, and anemia (Figure 2). We also found an association between ASCVD (HR: 1.96; 95% CI: 1.56-2.47) and a high ASCVD risk score (HR: 3.16; 95% CI: 2.73-3.65) with nonmetastatic cancer (Figure 2). Overall, our analysis indicates that traditional risk factors, along with incident ASCVD, were associated with the development of nonmetastatic cancer.
Figure 2.
Multivariable Cause-Specific Analysis for Variables Associated With Metastatic and Nonmetastatic Cancer
A multivariable, cause-specific Cox regression model shows that atherosclerotic cardiovascular disease (ASCVD) significantly increases the risk of metastatic cancer at diagnosis. ASCVD was analyzed as a time-dependent variable, incorporating a 1-year blanking period for adjustments. eGFR = estimated glomerular filtration rate; LDL = low-density lipoprotein; MET = metastatic cancer.
ASCVD was associated with metastasis at cancer diagnosis
Of the 1,793 patients diagnosed with cancer, 757 (3.5%) had metastatic cancer at diagnosis. This included 268 cases of metastatic skin cancer (29%), with nearly all (260) being metastatic melanoma (Figure 3). We also found 71 cases of metastatic breast cancer (8%), 60 cases of metastatic lung cancer (7%), and 211 cases (23%) of metastatic cancer of unknown origin (Figure 3).
Figure 3.
Types of Cancer
This figure displays crude counts of metastatic cancers and total cancer cases across different cancer types.
To determine the association between ASCVD and metastasis at cancer diagnosis while considering shared risk factors, we used an adjusted, cause-specific Cox regression, with incident ASCVD as a time-varying covariate. First, we adjusted for age and sex. Both older age and male sex were associated with higher rates of metastasis at cancer diagnosis (Table 2). Furthermore, the relationship between ASCVD and metastasis remained statistically significant (HR: 1.20; 95% CI: 1.01-1.43; P = 0.037) (Table 2).
Table 2.
Multivariable Analysis for Variables Associated With Metastatic Cancer With Adjustment for Age and Sex
| HR | 95% CI | P Value | |
|---|---|---|---|
| Age, y | 1.07 | 1.06-1.07 | <0.001 |
| Male | 1.16 | 1.05-1.28 | 0.004 |
| Atherosclerotic cardiovascular disease | 1.20 | 1.01-1.43 | 0.037 |
The data were analyzed using a Cox proportional hazards regression, treating atherosclerotic cardiovascular disease as a time-dependent variable, following a 1-year blanking period.
Subsequently, the model was further refined to control for shared risk factors, including the ASCVD risk score (incorporating age, sex, smoking, and other shared risk factors), chronic kidney disease (eGFR <60 mL/min/1.73 m2), obesity (BMI >30 kg/m2), anemia, LDL cholesterol, and impaired fasting glucose. We verified the proportional hazards assumption for all covariates using Schoenfeld residuals (Supplemental Table 1). The analysis revealed that incident ASCVD was independently associated with a 75% increase in the risk of metastasis at cancer diagnosis (adjusted HR: 1.75; 95% CI: 1.33-2.29; P < 0.001) (Figure 2). Overall, after adjusting for multiple risk factors, ASCVD remained a significant independent predictor of metastasis at diagnosis.
Risk of metastatic and nonmetastatic cancer and ASCVD was strongest in middle-aged adults
We then used ASCVD as a time-dependent variable to identify specific subgroups at increased risk for cancer (metastatic or nonmetastatic). We observed that the increased risk of metastasis and nonmetastatic cancer was age dependent. Specifically, middle-aged adults (age ≤54 years, median age 43 years, Q1-Q3: 39-48 years) exhibited a 64% increased risk for metastasis at cancer diagnosis (adjusted HR: 1.64; 95% CI: 1.03-2.61; P = 0.036) and a 104% increased risk for nonmetastatic disease at cancer diagnosis (adjusted HR: 2.04; 95% CI: 1.39-2.98; P < 0.001) (Figure 4). In contrast, in older adults (age >54 years, median age = 59 years, Q1-Q3: 56-63 years), the disease was not significantly associated with metastasis risk (adjusted HR: 1.11; 95% CI: 0.78-1.60; P = 0.56; Pinteraction = 0.039) or nonmetastatic cancer (adjusted HR: 0.97; 95% CI: 0.72-1.32; P = 0.85) (Figure 4). Overall, this analysis highlights that the association of ASCVD with metastatic and nonmetastatic cancer at diagnosis is most pronounced in middle-aged adults.
Figure 4.
Variables Associated With Metastatic and Nonmetastatic Cancer by Age Groups
A multivariable, cause-specific Cox regression model reveals that ASCVD significantly increases the risk of metastatic cancer in middle-aged adults but not in older adults. The disease was analyzed as a time-dependent variable with adjustments made for a 1-year blanking period. Abbreviations as in Figure 2.
ASCVD, metastatic cancer, and mortality
We used a Cox regression model to explore whether the co-occurrence of ASCVD and metastatic cancer correlates with a poorer prognosis. This model adjusted for factors including an ASCVD score >7.5, eGFR <60 mL/min/1.73 m2, obesity, LDL, impaired fasting glucose, and anemia. It also treated incident ASCVD, metastatic cancer at diagnosis, and nonmetastatic cancer as time-dependent variables, with incident metastatic and nonmetastatic cancer events being mutually exclusive. The analysis revealed significant associations with increased mortality for metastatic cancer at diagnosis, incident nonmetastatic cancer, and ASCVD (for ASCVD, adjusted HR: 4.27; 95% CI: 3.47-5.25; P < 0.001) (Figure 5).
Figure 5.
Atherosclerotic Cardiovascular Diseases Negatively Affect the Prognosis of Patients With Metastatic Cancer
Using a multivariable Cox regression model with incident ASCVD and cancer status as time-varying covariates, this analysis confirms that incident ASCVD significantly worsens the prognosis for patients with metastatic cancer. Abbreviations as in Figure 2.
Importantly, we found a significant interaction between ASCVD and metastasis at diagnosis, indicating that the increased risk of death associated with metastasis was more pronounced in subjects with ASCVD than in those without (adjusted HR: 1.31; Pinteraction < 0.001). Overall, our findings suggest that ASCVD not only increases the incidence of metastasis at diagnosis but is also significantly associated with mortality when it co-occurs with cancer.
Risk for incident metastasis by propensity-matched landmark analysis
To validate our findings and further investigate the association between ASCVD and metastasis at diagnosis, we performed a propensity score–based landmark analysis. Patients were matched on age, sex, BMI, smoking, diabetes mellitus, hypertension, and length of follow-up. Among 1,333 patients with ASCVD, 1,259 were successfully matched (Table 3, Supplemental Table 2). Follow-up began at the time of diagnosis for patients with ASCVD and at the initial visit for those without. The median follow-up was 7 years for both groups, with quartiles ranging from 4 to 12 years for the ASCVD group and 3 to 12 years for the group without the disease.
Table 3.
Patient Characteristics in the Propensity Score Matching Analysis
| Before Matching |
After Propensity Score Matching |
|||
|---|---|---|---|---|
| No Cardiovascular Disease (n = 20,321) | Cardiovascular Disease (n = 1,333) | No Cardiovascular Disease (n = 1,259) | Cardiovascular Disease (n = 1,259) | |
| Baseline age, y | 46 (40-53) | 55 (49-63) | 55 (49-63) | 55 (49-63) |
| Follow-up, y | 6 (3-12) | 7 (4-12) | 7 (3-12) | 7 (4-12) |
| Male | 13,944 (69) (69%) | 1,194 (90) | 1,107 (88) | 1,124 (89) |
| Smoking | 3,394 (17) | 231 (18) | 197 (16) | 226 (18) |
| High fitness | 6,850 (40) | 282 (27) | 317 (31) | 275 (28) |
| Weight | ||||
| Underweight | 131 (0.6) | 1 (0.1) | 0 (0) | 1 (0.1) |
| Normal weight | 9,998 (49) | 394 (30) | 382 (31) | 378 (30) |
| Overweight | 6,782 (33) | 676 (51) | 654 (52) | 646 (51) |
| Obesity | 2,515 (12) | 239 (18) | 223 (18) | 234 (19) |
| BMI, kg/m2 | 26 ± 4 | 27 ± 4 | 27 ± 4 | 27 ± 4 |
| Missing BMI data | 895 (4) | 23 (1) | 0 (0) | 0 (0) |
| ASCVD score >7.5 | 3,139 (15) | 690 (52) | 638 (51) | 680 (54) |
| Comorbidities | ||||
| Atrial fibrillation | 294 (1) | 128 (10) | 51 (4) | 120 (10) |
| Diabetes mellites | 1,259 (6) | 307 (23) | 288 (23) | 293 (23) |
| Impaired fasting glucose | 2,112 (10) | 297 (22) | 257 (21) | 271 (22) |
| Hypertension | 4,121 (20) | 788 (58) | 721 (58) | 729 (58) |
| SBP, mm Hg | 122 ± 16 | 132 ± 18 | 132 ± 18 | 132 ± 18 |
| DBP mm Hg | 77 ± 11 | 81 ± 10 | 82 ± 10 | 81 ± 10 |
| eGFR <60 mL/min/1.73 m2 | 743 (4) | 141 (11) | 152 (12) | 131 (11) |
| eGFR ≥60 to <90 mL/min/1.73 m2 | 12,511 (62) | 997 (75) | 934 (75) | 950 (75) |
| eGFR ≥90 mL/min/1.73 m2 | 7,034 (35) | 190 (14) | 173 (13) | 178 (14) |
| Missing eGFR data | 33 (0.2) | 5 (0.4) | 0 (0) | 0 (0) |
| Creatinine | 0.99 ± 0.18 | 1.09 ± 0.16 | 1.09 ± 0.16 | 1.09 ± 0.16 |
| Anemia | 2,824 (14) | 63 (5) | 63 (5) | 62 (5) |
| Hemoglobin | 14.4 ± 1.3 | 15.0 ± 1.2 | 15.0 ± 1.1 | 15.0 ± 1.2 |
| HDL | 48.7 ± 12.5 | 44.2 ± 11.1 | 45.9 ± 11.8 | 44.2 ± 11.1 |
| LDL | 122 ± 28 | 127 ± 30 | 125 ± 28 | 127 ± 30 |
| Total cholesterol | 191 ± 34 | 199 ± 34 | 196 ± 33 | 199 ± 34 |
| Triglycerides | 123 ± 70 | 147 ± 78 | 136 ± 70 | 147 ± 78 |
Values are median (Q1-Q3), n (%), or mean ± SD. Clinical characteristics of patients with and without ASCVD, before and after matching each ASCVD patient to a control patient.
ASCVD = atherosclerotic cardiovascular disease risk score; other abbreviations as in Table 1.
Gray’s competing risk method showed that the risk of metastasis at diagnosis was 3.0% ± 0.5% vs 1.1% ± 0.3% at 5 years and 7.2% ± 0.9% vs 3.6% ± 0.6% at 10 years for patients with vs without ASCVD, respectively (Figure 6A). Additionally, using Gray’s method, patients with ASCVD were found to be 39% more likely to present with metastasis at diagnosis during follow-up according to the landmark analysis model (sHR: 1.39; 95% CI: 1.03-1.87; P = 0.032).
Figure 6.
Propensity-Matched Landmark Analysis for the Risk of Metastatic Cancer at Diagnosis
Using Gray’s method, with nonmetastatic cancer and death as competing risks, this analysis uses atherosclerotic cardiovascular disease (ASCVD) as a landmark after a 1-year blanking period. Each patient with ASCVD was matched by propensity scores to a control patient. Error bands represent 95% CIs. (A) ASCVD was associated with an increased risk of metastatic cancer at diagnosis. This association was (B) strong in middle-aged adults but (C) null in older adults (C). SHR = subdistribution HR.
Consistent with the primary analysis, interaction analysis within the matched landmark framework showed that the risk associated with ASCVD was moderated by age (Figures 6B and 6C). Specifically, among middle-aged adults, ASCVD was associated with a significantly increased independent risk of metastatic cancer at diagnosis, with an sHR of 1.98 (95% CI: 1.33-2.95; P = 0.001). In contrast, among older adults, this association was not statistically significant (sHR: 0.78; 95% CI: 0.49-1.23; P = 0.28). Overall, the results from the matched landmark analysis confirmed our initial findings showing that ASCVD is associated with a higher incidence of metastasis at diagnosis, particularly among middle-aged adults.
Discussion
In this study, we demonstrate that ASCVD is associated with an increased risk of both metastatic and nonmetastatic disease at cancer diagnosis. To our knowledge, this is the first study to present detailed evidence that incident ASCVD is specifically an independent predictor of metastasis at cancer diagnosis among middle-aged adults (Central Illustration). Moreover, the co-occurrence of ASCVD and cancer metastasis at diagnosis correlated with poorer survival outcomes. These findings underscore the critical importance of recognizing the deadly interplay between ASCVD and metastatic cancer. Enhancing awareness of this association could significantly improve the prevention, early detection, and treatment of metastatic cancer.
Central Illustration.
Atherosclerotic Cardiovascular Diseases Were Associated With an Increased Risk of Metastatic Cancer at Diagnosis
In an analysis of 21,654 individuals free from cancer and atherosclerotic cardiovascular disease (ASCVD) at baseline, who were participants in a preventive health care program, it was demonstrated that incident ASCVD is associated with an elevated risk of metastatic and nonmetastatic cancer. eGFR = estimated glomerular filtration rate; LDL = low-density lipoprotein.
Comparison with previous reports
The link between ASCVD and metastasis at diagnosis has not been previously reported. This study uniquely unveils this link, thereby confirming and expanding on previous reports showing the link between ASCVD and cancer.1, 2, 3, 4, 5, 6, 7, 8 Moreover, we determined that ASCVD and cancer co-occurrence worsen patient prognosis.2,3 Importantly, our analysis suggests that the observed increase in risk within the entire cohort is primarily driven by ASCVD in middle-aged adults. This heightened risk is not exclusively associated with conditions like myocardial infarction or stroke. For instance, Saliba et al2 demonstrated that middle-aged adults with moderate to severe aortic stenosis face an increased risk of nonhematologic cancer.2 Conversely, Lam et al7 found no significant association between prevalent heart failure in postmenopausal women and an increased incidence of breast cancer, with the mean age of patients in their cohort being 63 years, 3 years older than the older adult subgroup in our cohort. In line with these reports, our study indicates that the risk of metastatic and nonmetastatic cancer at diagnosis among patients with ASCVD is most pronounced and limited to middle-aged adults.
Notably, we also found that an elevated ASCVD score and low eGFR were strongly associated with metastatic and nonmetastatic cancer at diagnosis. An elevated ASCVD score, which encompasses multiple risk factors common to both ASCVD and cancer, has been previously linked to future cancer risks by Lau et al.9 Similarly, chronic kidney disease has been recognized as a predictor of cancer.28 These findings confirm and extend previous research to include cancer metastasis at diagnosis.
Our results are consistent with prior studies showing that the predictive power of risk factors for ASCVD or cancer tends to diminish with age.29, 30, 31 For example, major risk factors such as hypertension, smoking, and diabetes appear to be less predictive in older adults.32,33 Similar findings have been found in studies measuring the influence of BMI, smoking, and diabetes on the risk of pancreatic cancer.30,31 The reasons why the associations between ASCVD and metastasis or nonmetastatic cancer at diagnosis are not evident in older adults remain uncertain, possibly because of biological mechanisms or the interplay of competing risk factors. Finally, nontraditional risk factors, such as frailty, may complicate prognostication in older adults.34
Our analysis demonstrates that ASCVD is associated with an increased risk of both metastatic and nonmetastatic cancer at diagnosis. The similarity in risk levels is indicated by the overlapping CIs of their HRs (Figure 2). This increased cancer risk may partly stem from overlapping risk factors and heightened medical surveillance in patients with heart diseases, who often undergo more diagnostic tests that could inadvertently detect cancer.35 In addition, the use of anticoagulants and antithrombotic agents prescribed to these patients not only increases the risk of bleeding but also potentially raises the frequency of colonoscopies, particularly for colorectal cancer detection.35
Thus, integrating imaging modalities for the early detection of atherosclerotic cardiovascular disease and cancer could offer advantages. For example, cardiac computed tomography used to detect coronary atherosclerosis might concurrently serve as a screening tool for lung cancer.36 Therefore, integrating various imaging techniques could facilitate the early diagnosis of cancer, ultimately helping to prevent its progression to metastatic cancer.
Strengths and limitations
Our database provides several key advantages. Each patient in our cohort underwent a comprehensive evaluation at the start of the follow-up, allowing for a detailed characterization of baseline health with an array of health metrics. The routine use of exercise stress tests confirmed or ruled out the diagnosis of ischemic heart disease. Additionally, the examination of each case of ASCVD by a blinded investigator further enhances the validity of our data. Finally, the INCR, a national registry, ensures complete follow-up data with no loss of information.
Despite the strengths of our study, we acknowledge several limitations. Our cohort predominantly comprises White individuals of high socioeconomic status and has inadequate representation of women, which may limit the generalizability of our findings. Second, we lacked data on diet, alcohol consumption, drugs, environmental hazards, and lifestyle, all of which could influence the risk of metastatic cancer. Third, INCR provides information solely on the primary diagnosis of cancer (metastatic or nonmetastatic), thereby limiting our analysis to the association of ASCVD with cancer metastasis at the initial diagnosis rather than the progression from nonmetastatic to metastatic cancer. Finally, our multivariable model may not adjust for all risk factors and confounders, a common limitation inherent to observational studies.
Summary, Future Research, and Implications
We propose that ASCVDs are linked to an increased risk of cancer, specifically metastasis, at the initial diagnosis of cancer. Our findings highlight the need for an integrated approach to health care that considers the interactions between concomitant diseases. Understanding the mechanisms that connect ASCVD with metastatic and nonmetastatic cancer could facilitate earlier diagnosis and prevention and might also lead to new therapies targeting the pathways shared by ASCVD and cancer. Until then, we recommend that patients with ASCVD, especially those with concomitant cancer, should manage shared risk factors, maintain a healthy lifestyle, undergo periodic screenings, and seek prompt medical intervention to reduce the risk of cancer.
Perspectives.
COMPETENCY IN MEDICAL KNOWLEDGE: ASCVD is associated with an increased risk of metastatic and nonmetastatic disease at the initial diagnosis of cancer. The association is strongest in middle-aged adults and weaker among older adults.
TRANSLATIONAL OUTLOOK: Understanding the mechanisms linking ASCVD to cancer, specifically metastatic disease, could facilitate earlier diagnosis and prevention and lead to new therapies that target shared pathways of ASCVD and cancer metastasis.
Funding Support and Author Disclosures
This project was supported by research grants from the Israel Cancer Association, the Israel Science Foundation, and the Seymour Feffer Foundation. A PhD scholarship from Mrs Tuna Gursoy supported Tal Caller in partial fulfillment of requirements for the PhD degree (Faculty of Medical and Health Sciences, Tel Aviv University). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Footnotes
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the Author Center.
Appendix
For an expanded Methods section and supplemental tables and figures, please see the online version of this paper.
Contributor Information
Jonathan Leor, Email: leorj@tauex.tau.ac.il.
Elad Maor, Email: elad.maor@sheba.health.gov.il.
Appendix
References
- 1.Bertero E., Canepa M., Maack C., Ameri P. Linking heart failure to cancer: background evidence and research perspectives. Circulation. 2018;138:735–742. doi: 10.1161/CIRCULATIONAHA.118.033603. [DOI] [PubMed] [Google Scholar]
- 2.Saliba W., Bental T., Shapira Y., et al. Increased risk of non-hematological cancer in young patients with aortic stenosis: a retrospective cohort study. Cardiooncology. 2021;7(1):37. doi: 10.1186/s40959-021-00123-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Banke A., Schou M., Videbaek L., et al. Incidence of cancer in patients with chronic heart failure: a long-term follow-up study. Eur J Heart Fail. 2016;18:260–266. doi: 10.1002/ejhf.472. [DOI] [PubMed] [Google Scholar]
- 4.Bell C.F., Lei X., Haas A., et al. Risk of cancer after diagnosis of cardiovascular disease. JACC CardioOncol. 2023;5:431–440. doi: 10.1016/j.jaccao.2023.01.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Roderburg C., Loosen S.H., Jahn J.K., et al. Heart failure is associated with an increased incidence of cancer diagnoses. ESC Heart Fail. 2021;8:3628–3633. doi: 10.1002/ehf2.13421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bertero E., Robusto F., Rulli E., et al. Cancer incidence and mortality according to pre-existing heart failure in a community-based cohort. JACC CardioOncol. 2022;4:98–109. doi: 10.1016/j.jaccao.2021.11.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lam P.H., Barac A., Nohria A., et al. Temporal associations and outcomes of breast cancer and heart failure in postmenopausal women. JACC CardioOncol. 2020;2:567–577. doi: 10.1016/j.jaccao.2020.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wilcox N.S., Amit U., Reibel J.B., Berlin E., Howell K., Ky B. Cardiovascular disease and cancer: shared risk factors and mechanisms. Nat Rev Cardiol. 2024;45(14):1209–1223. doi: 10.1038/s41569-024-01017-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lau E.S., Paniagua S.M., Liu E., et al. Cardiovascular risk factors are associated with future cancer. JACC CardioOncol. 2021;3:48–58. doi: 10.1016/j.jaccao.2020.12.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ameri P., Bertero E., Lombardi M., et al. Ischaemic heart disease in patients with cancer. Eur Heart J. 2024;45(14):1209–1223. doi: 10.1093/eurheartj/ehae047. [DOI] [PubMed] [Google Scholar]
- 11.Liu E.E., Suthahar N., Paniagua S.M., et al. Association of cardiometabolic disease with cancer in the community. JACC CardioOncol. 2022;4:69–81. doi: 10.1016/j.jaccao.2022.01.095. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Libby P., Kobold S. Inflammation: a common contributor to cancer, aging, and cardiovascular diseases—expanding the concept of cardio-oncology. Cardiovasc Res. 2019;115:824–829. doi: 10.1093/cvr/cvz058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lancellotti P., Marechal P., Donis N., Oury C. Inflammation, cardiovascular disease, and cancer: a common link with far-reaching implications. Eur Heart J. 2019;40:3910–3912. doi: 10.1093/eurheartj/ehz645. [DOI] [PubMed] [Google Scholar]
- 14.de Boer R.A., Hulot J.S., Tocchetti C.G., et al. Common mechanistic pathways in cancer and heart failure. A scientific roadmap on behalf of the Translational Research Committee of the Heart Failure Association (HFA) of the European Society of Cardiology (ESC) Eur J Heart Fail. 2020;22:2272–2289. doi: 10.1002/ejhf.2029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Meijers W.C., Maglione M., Bakker S.J.L., et al. Heart failure stimulates tumor growth by circulating factors. Circulation. 2018;138:678–691. doi: 10.1161/CIRCULATIONAHA.117.030816. [DOI] [PubMed] [Google Scholar]
- 16.Avraham S., Abu-Sharki S., Shofti R., et al. Early cardiac remodeling promotes tumor growth and metastasis. Circulation. 2020;142:670–683. doi: 10.1161/CIRCULATIONAHA.120.046471. [DOI] [PubMed] [Google Scholar]
- 17.Koelwyn G.J., Newman A.A.C., Afonso M.S. Myocardial infarction accelerates breast cancer via innate immune reprogramming. Nat Med. 2020;26:1452–1458. doi: 10.1038/s41591-020-0964-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Caller T., Rotem I., Shaihov-Teper O., et al. Small extracellular vesicles from infarcted and failing heart accelerate tumor growth. Circulation. 2024;149(22):1729–1748. doi: 10.1161/CIRCULATIONAHA.123.066911. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Tani T., Oikawa M., Misaka T., Ishida T., Takeishi Y. Heart failure post-myocardial infarction promotes mammary tumor growth through the NGF-TRKA pathway. JACC CardioOncol. 2024;6:55–66. doi: 10.1016/j.jaccao.2023.10.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Fardman A., Banschick G.D., Rabia R., et al. Cardiorespiratory fitness and survival following cancer diagnosis. Eur J Prev Cardiol. 2021;28:1242–1249. doi: 10.1177/2047487320930873. [DOI] [PubMed] [Google Scholar]
- 21.Fardman A., Banschick G.D., Rabia R., et al. Cardiorespiratory fitness is an independent predictor of cardiovascular morbidity and mortality and improves accuracy of prediction models. Can J Cardiol. 2021;37:241–250. doi: 10.1016/j.cjca.2020.05.017. [DOI] [PubMed] [Google Scholar]
- 22.Robinson E., Silverman B.G., Keinan-Boker L. Using Israel’s National Cancer Registry Database to track progress in the war on cancer: a challenge for the health services. Isr Med Assoc J. 2017;19:221–224. [PubMed] [Google Scholar]
- 23.Ruhl J.L., Callaghan C., Hurlbut A., Ries L., Schussler N. National Cancer Institute; 2018. Summary Stage 2018: Codes and Coding Instructions. [Google Scholar]
- 24.Arnett D.K., Blumenthal R.S., Albert M.A., et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. J Am Coll Cardiol. 2019;74:e177–e232. doi: 10.1016/j.jacc.2019.03.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Levey A.S., Schmid C.H., Zhang Y.L., et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–612. doi: 10.7326/0003-4819-150-9-200905050-00006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Austin P.C., Lee D.S., Fine J.P. Introduction to the analysis of survival data in the presence of competing risks. Circulation. 2016;133:601–609. doi: 10.1161/circulationaha.115.017719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Haukoos J.S., Lewis R.J. The propensity score. JAMA. 2015;314(15):1637–1638. doi: 10.1001/jama.2015.13480. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Lowrance W.T., Ordonez J., Udaltsova N., Russo P., Go A.S. CKD and the risk of incident cancer. J Am Soc Nephrol. 2014;25:2327–2334. doi: 10.1681/ASN.2013060604. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Shield K.D., Rehm J. Global risk factor rankings: the importance of age-based health loss inequities caused by alcohol and other risk factors. BMC Res Notes. 2015;8:231. doi: 10.1186/s13104-015-1207-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Yuan C., Kim J., Wang Q.L., et al. The age-dependent association of risk factors with pancreatic cancer. Ann Oncol. 2022;33:693–701. doi: 10.1016/j.annonc.2022.03.276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Li D., Morris J.S., Liu J., et al. Body mass index and risk, age of onset, and survival in patients with pancreatic cancer. JAMA. 2009;301:2553–2562. doi: 10.1001/jama.2009.886. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Tian F., Chen L., Qian Z.M., et al. Ranking age-specific modifiable risk factors for cardiovascular disease and mortality: evidence from a population-based longitudinal study. EClinicalMedicine. 2023;64 doi: 10.1016/j.eclinm.2023.102230. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Kaneko H., Yano Y., Okada A., et al. Age-dependent association between modifiable risk factors and incident cardiovascular disease. J Am Heart Assoc. 2023;12 doi: 10.1161/JAHA.122.027684. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Razavi A.C., Dzaye O., Michos E.D., et al. Nontraditional risk markers for incident coronary artery calcium among persons ≥65 years of age. JACC Adv. 2024;3 doi: 10.1016/j.jacadv.2023.100755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.de Boer R.A., Aboumsallem J.P., Bracun V., et al. A new classification of cardio-oncology syndromes. Cardiooncology. 2021;7:24. doi: 10.1186/s40959-021-00110-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Handy C.E., Quispe R., Pinto X., et al. Synergistic opportunities in the interplay between cancer screening and cardiovascular disease risk assessment: together we are stronger. Circulation. 2018;138:727–734. doi: 10.1161/CIRCULATIONAHA.118.035516. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All requests for raw and analyzed data and related materials will be reviewed by Sheba Medical Center’s legal department to verify whether the request is subject to any confidentiality obligations.