Clinical significance of increased arterial stiffness associated with atrial fibrillation, according to Framingham risk score

Atrial fibrillation (AF) is the most common arrhythmia in the elderly. Arterial stiffness may predict the risk of AF, but this relationship has not been fully evaluated. We assessed the association between arterial stiffness and prevalent AF. All subjects who had electrocardiography performed and a cardio-ankle vascular index (CAVI) calculated during a screening examination between 2010 and 2019 were enrolled. To evaluate the association between increased arterial stiffness and AF, we divided the population according to their Framingham risk score (FRS) into low-, intermediate-, and high-risk groups. A total of 8048 subjects were evaluated. The multivariate analysis revealed that increased arterial stiffness was significantly associated with AF prevalence, even after adjusting cardiovascular risk factors [odds ratio (OR) 1.685, 95% confidence interval (CI) 1.908–2.588, p = 0.017]. When we subcategorized the subjects according to their FRS, increased arterial stiffness was significantly associated with AF in the intermediate- and high-risk groups (OR 3.062, 95% CI 1.39-6.740 and OR3.877, 95% CI 1.142-13.167, respectively, BMI adjusted. High arterial stiffness shows a significant association with AF in those with intermediate or high cardiovascular risk, and can be used for further risk stratification of patients.


Methods
Study population. The current study was performed as a retrospective cross-sectional study. All apparently healthy subjects with a 12-lead electrocardiogram (ECG) and a cardio-ankle vascular index (CAVI) on the same day during a health check-up examination at the Healthcare System Gangnam Center, Seoul National University Hospital, between January 2010 and December 2019, were screened for enrollment in this study. Health examination has become popular in Korea because a thorough medical checkup can be performed in a few hours and the majority of referred hospitals in Korea is now equipped with a healthcare center to provide check-ups for screening purpose. The subjects voluntarily attended for a general health check-up, while others were supported by their employer. Among 8726 patients with a CAVI, those without an ECG (n = 399), with a history of angina (n = 189), or with a previous percutaneous coronary intervention (n = 66) or stroke (n = 24) were excluded, leaving 8048 subjects to be included in the analysis (Fig. 1). Information including past medical history, comorbidities, and medications were obtained using subjectrecorded questionnaires and confirmed during the CAVI measurement. Patients were asked about their past history of cardiovascular disease, cerebrovascular disease, peripheral arterial disease, hypertension, diabetes mellitus, dyslipidemia, gout, and chronic kidney disease. Additionally, patients were asked about their smoking and alcohol use. The risk score for cardiovascular disease was determined using the FRS 22 .
The study protocol followed the Declaration of Helsinki of 1975, as revised in 1983. This study was approved by the Institutional Review Board of Seoul National University Hospital (H-1912-114-1090). Because the current study was performed using a retrospective design (use of a database and medical records), informed consent was waived by the board.
Anthropometric and laboratory parameters. The measurements of anthropometric and laboratory parameters were performed on the day of the health check-up examination. Blood pressure, body weight, and height were measured. The body mass index (BMI) was calculated using height and body weight measured using a digital scale, according to the formula: BMI = weight (kg)/height (m 2 ).
Blood tests were performed after at least 12 h of fasting, which included levels of total cholesterol, triglyceride (TG), high-density lipoprotein (HDL)-cholesterol, low-density lipoprotein (LDL)-cholesterol, fasting blood sugar (FBS), glycosylated hemoglobin (HbA1c), hemoglobin, uric acid, blood urea nitrogen, and serum creatinine. Measurements of LDL-cholesterol were used for analyses. An ECG was also obtained on the same day as the health check-up with the patient in the supine position, using 10 electrodes placed on the four limbs and chest surface, over a duration of 10 s and read by three different cardiologists.
Assessment of arterial stiffness using the CAVI. The CAVI was measured using the VaSera VS-1000 (Fukuda Denshi Co. Ltd, Tokyo, Japan), as described previously 19,[23][24][25] . Each patient was placed in a sitting position for at least 5 min, and the brachial pulse pressure was measured using an automatic cuff oscillometric device. Cuffs were applied to the four extremities (upper arms and ankles), with patients in the supine position. After resting for 10 min, the measurement was performed. The systolic, diastolic, and pulse pressures were measured twice with the patients in the supine position, and an average value was taken. A phonocardiography sensor was placed at the right sternal border in the second intercostal space, and ECG leads were attached to both wrists. The CAVI was determined using the following equation: with Ps and Pd as the systolic and diastolic blood pressures, respectively, PWV as the pulse-wave velocity, ΔP as Ps -Pd, ρ as the blood density, and a and b as constants. For analysis, the mean value of the left and right indices  www.nature.com/scientificreports/ was calculated, and the patients were grouped into a CAVI greater than or equal to 8 or a CAVI less than 8. A CAVI greater than or equal to 8 was defined as increased arterial stiffness, which has been shown to be associated with significant coronary artery stenosis, calcifications, and adverse cardiovascular events in previous studies 24-26 . Statistical analysis. Data are expressed as the mean ± standard deviation for continuous variables and as frequencies for categorical variables. Chi-square tests and Student's t tests were used for categorical and continuous variables, respectively, to compare the differences between groups. The univariate and multivariate logistic regression analyses were performed to assess the association between increased arterial stiffness and AF, and multiple models were evaluated to adjust for possible confounders. Among variables with a p value < 0.05 in univariate analyses, those with clinical importance were subjected to multivariate analyses. Confounding variables with potentially multiple collinearity problems were excluded. To evaluate the association between increased arterial stiffness and AF, according to the FRS, patients were categorized into three groups: low-, intermediate-, and high-risk, and the significance of high arterial stiffness was determined for each group. All statistical analyses were performed using the Statistical Package for Social Science software (IBM SPSS Statistics, version 23), and two-tailed p values less than 0.05 were considered to be statistically significant.

Results
Baseline characteristics. As shown in Table 1, patients with high arterial stiffness were older (62 ± 9 years vs. 53 ± 8 years, p < 0.001, CAVI ≥ 8 vs. CAVI < 8, respectively). In patients with a CAVI greater than or equal to 8, hypertension (p = 0.000), diabetes mellitus (p < 0.001), and dyslipidemia (p < 0.001) were more common than in patients with a CAVI less than 8. Systolic and diastolic blood pressures (both p < 0.001) and levels of FBS and HbA1C (p < 0.001 for both) were all higher in patients with a CAVI ≥ 8 than in patients with a CAVI < 8. AF was more frequently found in patients with a CAVI ≥ 8 than in those with a CAVI < 8 (2.2% vs. 1.0%, respectively, p < 0.001).

Associations between different parameters and the prevalent AF.
To evaluate the association between each parameter and the prevalent AF, univariate analysis was performed. An age greater than or equal to 65 years (odds ratio (OR) 2.606, 95% confidence interval (CI) 1.787-3.801, p = 0.000), male gender (OR 3.458,  (Table 2). Multivariate models were evaluated for the adjusted risk of increased arterial stiffness in relation to the prevalent AF (Table 3). Model I included age greater than or equal to 65 years, gender, presence of hypertension and diabetes mellitus, and LDL-cholesterol level greater than or equal to 130 mg/dL. Model II included age greater than or equal to 65 years, gender, BMI greater than or equal to 25 kg/m 2 , smoking, presence of hypertension and diabetes mellitus, and LDL-cholesterol level greater than or equal to 130 mg/dL. Model III included the FRS for adjustment. A CAVI greater than or equal to 8 was the most significant parameter associated with the prevalent AF in all models (adjusted OR 1.642, 95% CI 1.082-2.492, p = 0.020 for Model I; adjusted OR 1.685, 95% CI 1.908-2.588, p = 0.017 for Model II; and adjusted OR 2.064, 95% CI 1.397-3.050, p < 0.001 for Model III).

Differences in significance of increased arterial stiffness according to the FRS. Model III
revealed that even after adjusting for cardiovascular risk using the FRS, a CAVI greater than or equal to 8 was still a significant parameter in the prevalent AF. The study subjects were categorized into three groups-low, intermediate, and high risk of cardiovascular events-according to the FRS, and each group was evaluated separately. High arterial stiffness was not an equally significant parameter related to the prevalent AF in the three groups. A CAVI greater than or equal to 8 was a significant parameter in subjects with an intermediate or high risk of cardiovascular events, according to the FRS (OR 3.094, 95% CI 1.414-6.769, p = 0.005 for the intermediate-risk group; OR 3.690, 95% CI 1.090-12.490, p = 0.036 for the high-risk group).
Since the FRS is derived from an algorithm that includes traditional cardiovascular risk factors (age, gender, blood pressure, cholesterol levels, smoking behavior, and diabetes status), we performed additional adjustments for a BMI greater than or equal to 25 kg/m 2 to evaluate the significance of a CAVI greater than or equal to 8 in the different risk groups (Table 4). A CAVI greater than or equal to 8 still showed a significant association with the prevalent AF in both the intermediate-and high-risk groups (adjusted OR 3.062, 95% CI 1.391-6.740, p = 0.005 for the intermediate-risk group; adjusted OR 3.877, 95% CI 1.142-13.167, p = 0.030 for the high-risk group). In the low-risk group, a CAVI greater than or equal to 8 was not a significant parameter in either the unadjusted or adjusted model (OR 1.465, 95% CI 0.854-2.513, p = 0.166 for the unadjusted model; OR 1.535, 95% CI 0.892-2.640, p = 0.122 for the adjusted model). www.nature.com/scientificreports/

Discussion
In this study, the prevalence of AF was higher in patients with high arterial stiffness, as determined by the CAVI. The significance of a CAVI greater than or equal to 8 differed according to the relative risk of cardiovascular events, as measured by the FRS. High arterial stiffness showed a significant association with the prevalent AF in those with an intermediate or high cardiovascular risk, whereas it was not a significant parameter in the lowrisk group.
One of the strengths of our study is that our patient population represents an apparently healthy general adult population of a homogeneous ethnicity. All measured parameters were used for screening purposes in healthy subjects without known cardiovascular disease; thus, our data can be used for primary prevention in the general population. An interpretation of high arterial stiffness without known cardiovascular risk factors or any cardiac symptoms can be a challenge for physicians during medical screening, especially in those with a low or intermediate FRS. Our previous studies have tried to identify those individuals at relatively high risk for coronary Table 3. Multivariate models for association between arterial stiffness and the prevalent AF. BMI body mass index, CAVI cardio-ankle vascular index, FRS Framingham risk score, LDL low density lipoprotein.  www.nature.com/scientificreports/ atherosclerosis using the CAVI 24,25,27 , and we have tried to develop a more tailored and comprehensive approach for managing each person's cardiovascular risk. In this study, we evaluated the usefulness and significance of a CAVI greater than or equal to 8, according to the FRS, in relation to the prevalent AF. There is a need for an accurate and reliable estimate of absolute cardiovascular disease risk in order to identify those in need of aggressive medical management. Using multiple variables, the FRS plays an important role in the identification of at-risk patients and can be used as the foundation for modifying risk score models. Although widely used, the FRS has well-known limitations: Shaw et al. showed that 45% of a group at intermediate risk for cardiovascular events were shifted to the high-risk group when the FRS was re-calculated using coronary artery calcium-adjusted age 28 . Thus, cardiovascular risk cannot be solely evaluated using the FRS, especially in those at low or intermediate risk for cardiovascular events, and subjects with a relatively high risk need to be re-stratified for follow-up or appropriate medical service. As shown in our study, subjects with an intermediate or high FRS must be monitored and screened for not only coronary artery disease but also for the prevalent AF because if underdiagnosed, AF may present as cerebrovascular disease and cause significant morbidity, mortality, and an extremely poor quality of life.
Current evidence suggests that arterial stiffness reflects the integrated effects of traditional cardiovascular risk factors on vessels, as endothelial dysfunction is an early event in the development and progression of atherosclerotic disease. Vascular stiffening is also associated with risk factors of AF, and high pulse pressure is known to be a risk factor for incident AF 3,29 . Moreover, arterial stiffness is known to affect recurrence rates of AF after elective cardioversion 30 . Impaired vascular function has shown improvement after restoration of sinus rhythm in patients with AF 31 . Among various parameters used to measure arterial stiffness, the CAVI is a simple, effective, and reliable parameter used for screening arterial stiffness at many medical screening centers in Korea: It reflects the stiffness of whole arterial segments from the aorta to the tibial artery and is known to be a superior index of arterial stiffness compared to traditionally established parameters, such as brachial ankle pulse-wave velocity 32 . The CAVI is also independent of blood pressure, unlike other parameters 33 . The reliable measurement of arterial stiffness in AF has been evaluated in another study by Caluwe et al., and although that study did not use the CAVI, its results showed similar to our study 34 .
Although our study does not focus on anticoagulation therapy in patients diagnosed with AF, the clinical significance of high arterial stiffness in association with the prevalent AF provides an additional method for further risk stratification among those who may have been categorized into a low-risk group according to the CHA 2 DS 2 -VASc score. Anticoagulation therapy for stroke prevention in patients with AF is based on risk stratification using the CHA 2 DS 2 -VASc score, but the indications for anticoagulation therapy in patients with low scores are still vague. The 2016 guidelines from the European Society of Cardiology recommend an individualized approach for the weighting of risk factors 35 . The guidelines also suggest that biomarker-based risk scores may help better stratify patients, and, as shown in our study, arterial stiffness measured by the CAVI may serve as a marker to identify those who need early anticoagulation treatment, when more prognostic data is available.
Limitations. Our study has some limitations. First, as the study was designed as a retrospective review of medical records, the causal relationship between increased arterial stiffness and incident AF cannot be evaluated. Second, we also do not have data regarding future cardiovascular events in study subjects, so the prognostic value of our findings can only be elucidated in future studies. From a previous study, high pulse-wave velocity was associated with increased cardiovascular events and improved the prediction of adverse cardiovascular events in patients with AF 36 . As we have suggested above, we have found that the association between AF and high arterial stiffness differs according to the relative cardiovascular risk of each patient; thus, arterial stiffness can help improve risk stratification. Although we cannot provide guidelines for treatment measures to prevent cardiovascular events or incident AF, we provide a tool to define those who should be monitored and managed for cardiovascular risk factors more aggressively, even within the same FRS group. Future studies with serial CAVI and ECG data may help identify the incidence and associated parameters of AF. Third, it is possible that there may be a selection bias that the majority of participants of the study population interested in health care were included in the study. Lastly, there are potential missed AF patients for whom no AF episodes were recorded during the ECG evaluation, such as paroxysmal AF or those with sinus conversion after treatment.

Conclusion
The clinical significance of a high CAVI in relation to AF differs according to the FRS. These findings imply that arterial stiffness can be used for risk stratification of patients with AF, especially those with an intermediate or high risk for cardiovascular events.