Coronary microvascular dysfunction is an independent predictor of developing cancer in patients with non-obstructive coronary artery disease

Abstract

Aims

Cardiovascular disease and cancer share common pathogenesis and risk factors. Coronary microvascular dysfunction (CMD), reflecting impaired coronary microvascular dilation in response to stress, is related to a higher risk of major cardiovascular events; however, its association with cancer has not been explored.

Methods and results

A retrospective study on 1042 patients with non-obstructive coronary artery diseases (NOCADs) was performed. Data regarding demographic, clinical history, diagnostic coronary reactivity test, and cancer occurrence were collected. Coronary microvascular dysfunction was defined as coronary flow reserve (the ratio of hyperaemic blood flow to resting blood flow) ≤2.5. Thirty-four per cent had CMD (67.4% female and the average age was 52.4 ± 12.2 years). Of 917 patients with no history of cancer, 15.5% developed cancer during follow-up [median of 9 (4, 16) years]. Kaplan–Meier analysis showed that CMD patients had lower cancer-free survival compared with those without CMD (log-rank P = 0.005). Cox proportional hazard analyses showed that after adjusting for age, sex, hypertension, diabetes, smoking, and glomerular filtration rate, CMD is independently associated with cancer [hazard ratio, 1.4; 95% confidence interval (CI), 1.09–2.04; P = 0.04]. The rate of major adverse cardiovascular events (MACE) was significantly higher in CMD patients compared with that in non-CMD patients who had a previous history of cancer [odds ratio (OR), 2.5; 95% CI, 1–6.2; P = 0.04] and those with no history of cancer (OR, 1.4; 95% CI, 1.01–1.9; P = 0.044).

Conclusion

Coronary microvascular dysfunction is associated with cancer incidence in patients presenting with NOCADs. This study emphasizes follow-up in patients with CMD to evaluate the risk of MACE as well as potential malignant diseases.

Introduction

Cardiovascular disease (CVD) and cancer are the two leading causes of morbidity and mortality worldwide. They share common pathophysiology and risk factors including age, smoking, obesity, and dietary intake.¹ Previous studies suggested a bidirectional association between cancer and coronary artery disease (CAD). Patients diagnosed with cancer are more susceptible to developing CVD compared with a healthy population.² Likewise, recent studies revealed that the risk of cancer increased in patients with a history of CVD, possibly due to the common underlying mechanisms in both conditions including the role of oxidative stress and chronic inflammation.^3,4

Coronary microvascular dysfunction (CMD) is defined as impaired vasodilation of coronary arterioles and inadequate increase of coronary flow reserve (CFR) in response to stress or hyperaemic stimuli.^5,6 Coronary microvascular dysfunction is highly prevalent in patients presenting with signs and symptoms of cardiac ischaemia with non-obstructive CADs (NOCADs).⁷ A large body of evidence demonstrated the association of CMD with an elevated risk of major adverse cardiovascular events (MACE) including myocardial infarction, congestive heart failure, sudden cardiac death, and a higher rate of hospitalization.^5,6,8–10

Moreover, CMD has been shown to be associated with a wide variety of non-cardiovascular conditions such as diabetes, osteoporosis, autoimmune rheumatic disorders, and chronic kidney disease.^11–13 In a previous study, we showed that peripheral microvascular endothelial dysfunction is associated with an increased risk of solid cancer incidents.¹⁴ However, previous studies showed that there is a poor correlation between peripheral endothelial dysfunction, using arterial tonometry, and centrally measured coronary reactivity test (CRT).^15,16 In this study, we sought to investigate whether the CRT, the gold standard for diagnosing microvascular dysfunction, is associated with the risk of cancer occurrence in a cohort of patients presenting with angina and NOCAD. We also aimed to investigate the association of CMD with the increased risk of cardiovascular events in patients with NOCAD.

Methods

Study design

This is a retrospective single-centre cohort study approved by the Institutional Review Board of the Mayo Clinic (approval reference number: 05-004191). The primary aim of the study is to investigate the association of CMD with cancer incidents during the follow-up of patients with no history of baseline cancer independent of the conventional cardiovascular risk factors; the secondary aim of the study is to investigate whether CMD is associated with increased risk of cardiovascular events in NOCAD patients (Figure 1). All patients provided written informed consent. All investigators conformed to the principles outlined in the Declaration of Helsinki.

Study population

This is a retrospective observational cohort study consisting of all patients with symptoms of angina who were referred to the cardiac catheterization laboratory at Mayo Clinic between 2000 and 2019 for the evaluation of coronary angiography and were found to have non-significant coronary artery stenosis (<50%). Patients then underwent a comprehensive coronary physiology evaluation using the CRT. The exclusion criteria included obstructive CAD (angiographic stenosis ≥50%), unstable angina, myocardial infarction, ejection fraction <45%, cerebrovascular accidents within the preceding 6 months, having radiocontrast agents within 12 h of the procedure, and a previous history of myocardial infarction.

Study protocol

Data regarding demographic, clinical history, laboratory, and diagnostic coronary angiography of patients were collected. Coronary microvascular dysfunction was determined as described in detail previously.^17,18 Briefly, a Doppler guidewire was advanced and positioned in the mid-left anterior descending coronary artery. Intracoronary bolus injections of incremental doses (18–60 μg) of adenosine, an endothelium-independent vasodilator, were administered into the guiding catheter until maximal hyperaemia was achieved. Coronary flow reserve was calculated by dividing the maximal coronary flow velocity after adenosine by the baseline average peak flow velocity.¹⁹ Coronary flow reserve is an integrated measure of flow through both the large epicardial arteries and coronary microcirculation. However, in the absence of obstructive lesions of the epicardial arteries, reduced CFR can be taken as a marker of CMD.²⁰ The cut-off for normal CFR was considered to be >2.5 based on previous studies.⁷

In addition, traditional cardiovascular risk factors data were also collected as covariables including sex (male, female), chronological age, body mass index (BMI), smoking status (had exposure vs. had no exposure), history of documented diabetes mellitus, hypertension, and hyperlipidaemia as well as laboratory data including total cholesterol, estimated glomerular filtration rate (eGFR), and Hemoglobin A1C (HbA1C). The covariables selection was based on clinical relevance and previous evidence regarding the common risk factors of cancer and CVD.^21,22

Study outcomes

Study participants were followed up for the median of 9 (4, 16) with a maximum of 27 years. The time of study entry was considered as the time of undergoing coronary physiology study and the end of follow-up time was a period between study entry and the time of data retrieval (14 July 2020). For those who were diagnosed with cancer and those who died during the time of follow-up, the ‘time to event’ was defined as the period between study entry and the time of cancer diagnosis or time of death, respectively. In the survival analysis, the subjects were censored if there were no events (cancer incident) by the time of follow-up.

Data regarding the history of a cancer diagnosis before or after coronary angiography, type of cancer, date of cancer diagnosis, MACE (all-cause death, myocardial infarction, cerebrovascular event, and heart failure hospitalization) were gathered by using questionnaires and detailed chart review by an investigator blinded to the CFR results. Given that certain somatic gene mutations such as Tet2 have been identified as a shared risk factor in both haematopoietic malignancy and coronary atherosclerosis, we excluded the haematological cancers from the further analysis.²³

Statistical analysis

Normally and skewed distributed continuous variables were presented as the mean ± standard deviation and median with interquartile range (IQR), respectively. The between-group comparison was performed via Student’s t-test and with Mann–Whitney U test for normally and skewed distributed variables, respectively. The comparison of categorical variables was performed with Pearson’s χ² test. Kaplan–Meier approach with the log-rank method was used to determine if the cancer-free survival differs between patients with CMD compared with non-CMD patients. In addition, to prevent the risk of guarantee-time bias, Kaplan–Meier analysis was performed using the landmark approaches. We chose the landmark time points of 1, 3, and 5 years based on the clinical judgment and visual inspection of the Kaplan–Meier curve. Univariate and multivariate Cox proportional hazard analysis was performed to determine the factors that are associated with cancer incidents during follow-up.

The association of CMD in predicting the MACE during the follow-up period was assessed by multivariable binary logistic regression. The regression was adjusted to the conventional cardiovascular risk factors including age, sex, BMI, smoking, hypertension, diabetes, and hypolipidaemia. The analysis was conducted in patients with a positive history of cancer at the baseline and patients without previous history of cancer, separately.

Results

Baseline characteristics

A total of 1042 patients were recruited (67.4% were female and the average age was 52.4 ± 12.2 years). Hypertension was present in 42.2%, hyperlipidaemia in 56%, diabetes in 11.7%, and history of smoking in 46.5%. The mean BMI was 29.5 ± 6.5 kg/m² (Table 1).

Mean CFR for all patients was 2.9 ± 0.7 with 25.2% (262 patients) having CFR ≤ 2.5. The following risk factors significantly differed between those with vs. without CMD included: age (56.3 ± 11.8 vs. 50.6 ± 11.9 years old, P < 0.0001), sex (female = 79.8 vs. 63.3%, P < 0.0001), prior history of diabetes (16 vs. 10.3%, P = 0.012), hyperlipidaemia (62.9 vs. 53.6%, P = 0.005), prior cancer history (16.4 vs. 10.5%, P = 0.01), lower eGFR (74.7 ± 18.6 vs. 78.8 ± 17.1 mL/min/1.73 m², P < 0.0001), higher leucocytes count (6.9 ± 2.4 vs. 6.7 ± 2 × 10³/μL, P = 0.028), and higher HbA1c concentration (5.7 ± 1 vs. 5.4 ± 0.8%, P < 0.0001) (Table 1).

One hundred and twenty-five patients (77.60% females) had a baseline cancer history at the time of the CRT. The most common cancers included breast (26.2%), colon (11.4%), and cervical cancer (8.2%). The following variables were significantly different between patients with a history of cancer to those without cancer: the older age (60.1 ± 9.6 vs. 51.4 ± 12.2 years old, P < 0.0001), females (77.6 vs. 66.1%, P = 0.009), lower eGFR (70.9 ± 17.5 vs. 79.5 ± 17.5 mL/min/1.73 m², P < 0.0001), higher prevalence of hypertension (52.5 vs. 40.9%, P = 0.018), and diabetes (17.6 vs. 10.9%, P = 0.029). Coronary microvascular dysfunction was also more prevalent among patients with a history of cancer than those without cancer (34.4 vs. 24%, P = 0.011) and CFR was lower in patients with cancer (2.78 ± 0.7 vs. 2.95 ± 0.7, P = 0.015) (Table 1).

Risk factors for cancer incidence in follow-up

Patients with baseline cancer history (125 patients) at the time of the coronary angiography and haematological cancers (N = 8) were excluded from the follow-up analysis. The median follow-up time was 9 (4, 16) years. The total number of hospital visits during the follow-up was a median of 8 (4, 30) visits, regardless of the type of services attended. There was no significant difference between the number of visits between CMD compared with non-CMD patients [9 (IQR, 6, 30) compared with 8 (IQR, 3, 31), P = 0.4]. During follow-up, 15.5% (141 patients) were newly diagnosed with cancer. The three most common cancers were breast, prostate, and skin cancer (24.1, 11.3, and 10.6%, respectively) (see Supplementary material online, Table S1). In males, prostate, colon, and chronic myeloid leukaemia (31, 8.1, and 8.1%, respectively), and in females, breast, melanoma, and lung cancer (34, 9, and 8%, respectively) were the most incident cancers. The median time from undergoing CRT to developing cancer was 7 (4, 10) years. The diagnosis of cancer was significantly more frequent in those with CMD compared with those without (20.5 vs. 12.8%, P = 0.006). The average CFR was significantly lower in cancer-positive group (2.7 ± 0.6 with 31.2% having CMD) than in the cancer-negative groups (3 ± 0.8 with 22.4% having CMD) (Table 2).

A comparison of cancer-free survival between patients with vs. without CMD using Kaplan–Meier analysis demonstrated that patients with CMD had significantly lower cancer-free survival compared with those without. To reduce the reverse causality bias, the Kaplan–Meier analysis was done using the landmark time of 1, 3, and 5 years following the study start time. A significant difference in cancer-free survival was observed between patients with CMD and non-CMD groups in all three time points. (At baseline: log-rank χ², 7.29, χ²P = 0.005; after 1 year: log-rank χ², 6.36, P = 0.016; after 3 years: log-rank χ²: 5.23, P = 0.028; after 5 years: log-rank χ², 7.025, P = 0.012.) Additionally, the multivariate Cox proportional hazard model exhibited that CMD was independently associated with cancer incidence after adjusting for covariables including age, sex, BMI, hypertension, diabetes, hyperlipidaemia, eGFR, and smoking [hazard ratio, 1.45; 95% confidence interval (CI), 1.01–2.09; P = 0.04 for CMD] (Table 3 and Figure 2).

Association of coronary microvascular dysfunction with developing major adverse cardiovascular events during follow-up

Among 125 patients with a positive history of cancer prior to coronary angiography, all-cause mortality was 13% (68.7% were female). The overall incidence of MACE as a composite outcome was significantly higher in cancer patients who had CMD compared with non-CMD (53.5 vs. 34.1%, P = 0.005). Moreover, binary logistic regression analysis revealed that CMD was associated with cardiovascular events after adjusting to the traditional cardiovascular events such as age, sex, history of hypertension, diabetes, and smoking [odds ratio (OR), 2.5; 95% CI, 1.003–6.2; P = 0.04].

Additionally, among 917 patients without previous history of cancer, the incidence of all-cause mortality and MACE was 8.3 and 29.9%, respectively. Patients who were diagnosed with CMD have a significantly higher incidence of MACE compared with non-CMD patients (34.7% compared with 28.4%, P = 0.04). Multivariable logistic regression including conventional CVD risk factors revealed the independent association between CMD and cardiovascular events (OR, 1.4; 95% CI, 1.01–1.9; P = 0.044) (Table 4).

Discussion

Our study demonstrated that coronary microvascular function is associated with cancer incidence in patients presenting with NOCAD who had no baseline cancer history. Despite mounting evidence regarding the association between CVD and cancer, our study is the first report that specifically focused on the role of CMD in the absence of obstructive CAD to predict cancer incidence. The current study may support a common mechanism for CMD and cancer in humans. In addition, we observed CMD is associated with the increased risk of MACE in patients with or without previous history of cancer. This finding added evidence of the substantial role of microvascular dysfunction in CVDs.

The underlying mechanisms of overlap may account for the relationship between CMD and the increased risk of cancer. The microvasculature plays an important role in the homeostasis of the heart and other organs. The impaired function of vessels to vasodilate in response to stress or vasodilators subsequently leads to ischaemia, inflammation, and oxidative stress. The two most important causes of CMD are (i) abnormal coronary microvascular structure, such as intimal hyperplasia, smooth muscle cell proliferation, and low capillary density and (ii) abnormal coronary vasomotor function, such as endothelial dysfunction.²⁴ Therefore, microvascular function impairment may be an index of a range of underlying vascular abnormalities. Microvascular dysfunction can occur as a consequence of an imbalance between vasodilators [nitric oxide (NO) activity, prostaglandin I₂, etc.] and vasoconstrictors mediators (endothelin-1, superoxide, hydrogen peroxide, and thromboxane).²⁵ Reduction of NO bioavailability per se or subsequent to oxidative stress is one of the key mechanisms of developing CMD.^26,27 On the other hand, oxidative stress also plays a crucial role in carcinogenesis by DNA methylation and genetic instability.^28,29

Moreover, a growing body of evidence highlighted the substantial role of pro-inflammatory cytokines including interleukin (IL)-1, IL-8, and tumour necrosis factor-a in the pathophysiologic process of chronic conditions such as malignancy and chronic kidney disease as well as CAD.³⁰ For instance, a previous study suggested that chronic inflammation is responsible for nearly 25% of all cancer cases.^31,32 Hence, these common mechanisms explain in part the promising effects of medications such as statins, aspirin, and angiotensin-converting enzyme inhibitors in the prevention of both cancer and CVD.^33,34 In line with the literature, we also observed an increased value of innate immune cells in patients with CMD compared with those without, supporting the role of inflammation in developing CMD. However, further investigations on the role of novel risk factors, particularly pro-inflammatory markers to better understand the association of CMD with increased risk of cancer and other inflammatory conditions are warranted.

Microvasculature has a crucial role in maintaining haemostasis, through the regulation of vascular tone together with thrombogenicity and inflammation. Coronary microvascular dysfunction has been considered an early manifestation of atherosclerosis.³⁵ The association between microvascular dysfunction and MACE has been well-established in previous literature.^36–38 However, despite the systemic distribution of microvasculature, there is only limited evidence on the role of CMD on systemic non-cardiac disorders. Prasad et al.¹² previously showed that post-menopausal women with CMD had two-folds higher chances of developing osteoporosis after a 7-year follow-up. In a similar study by Reriani et al.,³⁹ CMD was a predictor of erectile dysfunction in men with non-obstructive coronary atherosclerosis, indicating the systemic involvement of microvascular dysfunction. In our previous study on the role of peripheral vasomotor dysfunction in cancer, we demonstrated that patients who had abnormal peripheral microvascular function defined as reactive hyperaemic peripheral arterial tonometry index <2.0 were almost three times more likely to develop cancer in the future compared with healthier patients.¹⁴ Consistent with our previous paper, we observed that CFR ≤ 2.5, as a gold standard to assess CMD, is associated with an increased risk of cancer and that after adjusting for other cardiovascular risk factors, only CFR and age remained independently associated with cancer. In agreement with our findings on the inter-relationship of traditional risk factors and CFR, Rubinshtein et al.¹⁷ showed in the multivariable analysis of 1063 patients with NOCAD, that a higher Framingham risk score was an independent predictor of CFR, indicating the summative contribution of conventional CVD risk factors to the development of CMD. Our findings emphasize the importance of cautious follow-up in patients with CMD not only for predicting MACE but also to detect cancers in their early stages.

Another interesting finding of our study was that in patients who had a cancer diagnosis established prior to the angiographic study, CMD was associated with an increased future risk of MACE. The relationship between CMD and the future rate of MACE has been reviewed previously in specific populations such as myocardial infarction with non-obstructive coronary arteries,⁴⁰ Takotsubo cardiomyopathy,⁴¹ and kidney-transplant patients.⁴² However, herein, we reported the above association in cancer patients. The current study demonstrated that the risk of developing MACE in NOCAD patients is greater in those with cancer history compared with those without cancer. This observation can be explained in part due to the role of cancer treatment on cardiovascular events, which has been extensively investigated before.^43,44 The importance of this observation is to verify the interaction between cancer and MACE pathogenesis, thus recommending a more close follow-up in patients with CMD not only to evaluate the risk of MACE but also to screen for potential malignant diseases, implement preventive strategies such as lifestyle modifications and adopt a therapeutic approach for vascular disease. As previously shown, cancer treatments such as chemotherapeutic agents and radion are associated with an increased risk of cardiovascular events such as thrombo-embolism, cardiomyopathy, heart failure, and CVD-related mortality.^44–46

There are a few limitations concerning our study. Exploring the cancer status was mostly based on a response to a follow-up questionnaire, raising the possibility of undetected cancer and missing values as not all patients responded to the survey. However, we intend to address this limitation by performing a detailed chart review and double-checking the diagnosis performed by a blinded investigator to minimize the possibility of undetected cancer. In addition, due to the inherent limitation of a retrospective observational study design, the current study is underpowered to detect cancer incidence as well as being incapable of deriving a causal relationship between increased risk of cancer and CMD, yet it paves the way for future research focusing on the pathophysiology and assessing the causal relationship. Furthermore, despite the well-identified effects of different medications as well as inflammatory markers in incident cancer, we were not able to do an adjusting analysis based on these variables due to the small number of events in our study sample. Moreover, since our study was mainly focused on patients with angina and NOCAD diagnosis, extrapolating the results to the general population should be performed with caution. Lastly, it is important to consider the possibility of ascertainment bias as patients with the diagnosis of CMD are more likely to receive increased surveillance. Although there was no significant difference in the number of patient visits to our health centre between CMD and non-CMD patients, it cannot completely exclude the risk of possible ascertainment bias.

Conclusion

Coronary microvascular dysfunction is associated with an increased risk of cancer incidence. Moreover, it is associated with a worse prognosis in cancer patients with pre-existing CMD. This study underscores the importance of early detection of CMD not only to predict the risk of MACE but also as a potential prevention or therapeutic target strategy in cancer.

Author contributions

N.R., A.A., and A.L. contributed to the conception or design of the work. N.R. and T.T. contributed to the analysis and interpretation of data for the work. N.R. drafted the manuscript. J.D.S., J.H., L.O.L., and A.L. critically revised the manuscript. All gave final approval and agreed to be accountable for all aspects of work ensuring integrity and accuracy.

Supplementary material

Supplementary material is available at European Journal of Preventive Cardiology.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Data availability

The data underlying this article will be shared upon reasonable request to the corresponding author.

References

Author notes