Article ID: CJ-18-0113
Background: Although the prevalence of both atrial fibrillation (AF) and metabolic syndrome (MetS) has been increasing in East Asia, the association between them is uncertain.
Methods and Results: A total of 24,741 middle-aged Korean men without baseline AF were enrolled in a health screening program from January 2003 to December 2008. Among them, 21,981 subjects were evaluated to determine the risk of AF based on baseline MetS status through December 2016. At every visit, the subjects were evaluated for AF using ECG. MetS was defined using the criteria of the International Diabetes Federation and was present in 2,529 subjects (11.5%). Mean (±standard deviation) age was 45.9±5.3 years. During a mean follow-up of 8.7 years, 168 subjects (0.8%) were diagnosed with AF. The age-adjusted and multivariate-adjusted hazard ratios (HR) for MetS with AF were 1.62 (P=0.02) and 1.57 (P=0.03), respectively. Among the components of MetS, central obesity (age-adjusted HR 1.62, P<0.01) and raised blood pressure (age-adjusted HR 1.43, P=0.02) were associated with an increased risk of AF.
Conclusions: MetS is associated with an increased risk of AF in middle-aged East Asian men. Of the components of MetS, central obesity is the most potent risk factor for the development of AF in this population.
Atrial fibrillation (AF) is the most common arrhythmia requiring treatment in clinical practice, and is associated with increased mortality rates.1,2 Although the prevalence of AF is known to be higher in the West than in Asia, the prevalence in Asia has been increasing with aging populations and a westernized lifestyle. In South Korea, prevalence rates of AF progressively increased 2.69-fold between 2004 and 2013.3
Metabolic syndrome (MetS) is a cluster of characteristics including central obesity (defined as waist circumference [WC] with ethnicity specific values), hypertension (HTN), diabetes mellitus (DM), and dyslipidemia, which are recognized as risk factors for cardiovascular disease; its prevalence has increased globally.4–6 In addition, previous studies suggest that MetS is associated with an increased risk of AF.7–9 However, previous studies were aimed at Western populations, or the definition of MetS was not exact (body mass index [BMI] used in place of WC, for example). BMI has a limitation for replacement of central obesity because Asians have a lower BMI in general at a given central obesity compared with Europeans.10
In this study, we aimed to determine the association of MetS with the risk of new-onset AF in middle-aged East Asian men.
We conducted a retrospective cohort study using data from 21,981 individuals who participated in an annual or biennial comprehensive health screening program at Ulsan University Hospital, Ulsan, Republic of Korea, from January 2003 to December 2008 (Figure 1). The Korean Industrial Safety and Health Law requires working individuals to participate in an annual or biennial health examination. In addition to mandatory health examination, the health screening program of Ulsan University Hospital included a 12-lead ECG. Most of the individuals who participated in the health examination at our center were employees of heavy industries and most of them were men (n=24,800, 97%); Women (n=862) were excluded. In addition, we excluded subjects if they did not have an initial ECG, if they had AF or atrial flutter (AFL) on the initial ECG, if they did not have a follow-up ECG; or if data from questionnaires and the others including anthropometric measurements, blood pressure (BP), and blood tests were unavailable. AF and AFL were diagnosed from the 12-lead ECG recorded at a follow-up visit (annual or biennial). AFL was included as an endpoint because it is closely related to AF, often coexists with AF, and is associated with a similar risk of stroke.11–13 The study protocol was approved by the Institutional Review Board at Ulsan University Hospital (IRB No. 2017-02-023). Subjects’ information was anonymized and de-identified prior to analysis. The requirement for informed consent was waived because of the anonymity of the subjects and the nonintrusive nature of the study.
Flowchart of the study.
Data were collected by a clinical data warehouse platform in conjunction with electronic medical records in Ulsan University Hospital. Anthropometric measurements were made by well-trained examiners while individuals were wearing light clothing and without shoes. Height was measured to the nearest 0.1 cm and weight to the nearest 0.1 kg. BMI values were calculated by dividing weight (kg) by height squared (m2). WC was measured in the standing position, midway between the lowest rib and the iliac crest with a measuring tape. Seated BP was measured by well-trained nurses using a mercury sphygmomanometer after subjects rested for ≥10 min. Following an overnight fast (≥8 h), blood samples were collected and analyzed in the same core clinical laboratory, which has been accredited and participates annually in inspections and surveys by the Korean Association of Quality Assurance for Clinical Laboratories. Blood tests included fasting glucose, hemoglobin A1c (HbA1c), fasting plasma lipid profile including total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol, and triglycerides (TG), serum creatinine, and blood urea nitrogen. The estimated glomerular filtration rate (eGFR) was calculated with the Modification of Diet in Renal Disease equation (eGFR=175×(serum creatinine)−1.154×age−0.203). Data on smoking status (never, former, or current), alcohol drinking status (frequency, amount), frequency of exercise, and medical history were collected from self-administered questionnaires. HTN was defined as a systolic BP (SBP) ≥140 mmHg or diastolic BP (DBP) ≥90 mmHg or current use of antihypertensive medication. DM was defined as fasting plasma glucose (FPG) value ≥126 mg/dL, HbA1c ≥6.5% or the use of blood glucose-lowering agents. Chronic kidney disease (CKD) was defined as eGFR <60 mL/min/1.73 m2.
Definition of MetSMetS was defined according to the criteria of the International Diabetes Federation.14 Central obesity (WC ≥90 cm) had to be present and any 2 of the following 4 components were required: (1) raised BP (SBP ≥130 mmHg or DBP ≥85 mmHg or treatment of previously diagnosed HTN); (2) raised FPG (≥100 mg/dL or diagnosed type 2 DM); (3) raised TG (≥150 mg/dL or drug treatment for high TG); and (4) reduced HDL-C (<40 mg/dL or drug treatment for lowering HDL-C).
Statistical AnalysisAll baseline patient characteristics were summarized as mean±standard deviation (SD) or frequency (percentage) for continuous or categorical variables. The baseline characteristics by baseline MetS status were compared by chi-square test and Student’s t test for categorical and continuous variables, respectively. Follow-up years were computed from the baseline examination until a first AF or AFL diagnosis, loss to follow-up, or the end date of December 2016. We estimated the cumulative incidence of AF based on the Kaplan-Meier method and we compared the cumulative incidence rates curves by log-rank test. Overall and age-adjusted incidence rates for AF by the baseline MetS status were also calculated. Age-adjusted and multivariate-adjusted hazard ratios (HRs) of MetS were estimated by Cox proportional hazard regression model. Multivariate-adjustments were made for age, smoking status, regular exercise (≥once/week), alcohol drinking status, and CKD. HRs for individual components of MetS after additional adjustment for the other components were also evaluated by Cox model. Results of the Cox proportional hazard model are presented as the HR and 95% confidence interval (95% CI). Analyses were performed with R software, version 3.4.1 (R Foundation for Statistical Computing, Vienna, Austria). All reported P-values are 2-sided, and P<0.05 was considered to indicate statistical significance.
A total of 21,981 men were enrolled in the cohort analysis. The mean age was 45.9±5.3 years. At baseline, MetS was present in 2,529 subjects (11.5%). Table 1 shows the baseline characteristics of this study population by baseline MetS status. Subjects with MetS were older, and had higher values for BMI, WC, SBP, DBP, pulse pressure, and had more comorbidities (HTN, DM, dyslipidemia, and impaired renal function) than subjects without MetS. Table 2 shows the prevalence of MetS and of the individual components of MetS.
| Variable | No MetS (n=19,452) |
MetS (n=2,529) |
P value |
|---|---|---|---|
| Age (years) | 45.9±5.2 | 46.0±5.7 | 0.03 |
| Body weight (kg) | 66.6±7.9 | 79.8±8.1 | <0.01 |
| BMI (kg/m2) | 23.2±2.3 | 27.0±2.1 | <0.01 |
| WC (cm) | 82.2±6.1 | 94.1±4.0 | <0.01 |
| Systolic BP (mmHg) | 119.5±13.9 | 128.7±14.5 | <0.01 |
| Diastolic BP (mmHg) | 77.0±9.6 | 83.3±10.0 | <0.01 |
| Pulse pressure (mmHg) | 42.6±8.1 | 45.4±8.7 | <0.01 |
| HTN (%) | 3,554 (18.3) | 1,096 (43.3) | <0.01 |
| DM (%) | 1,648 (8.5) | 422 (16.7) | <0.01 |
| Stroke (%) | 65 (0.3) | 5 (0.2) | 0.35 |
| Smoking status (%) | 0.02 | ||
| Never smoker | 6,300 (32.7) | 752 (30) | |
| Former smoker | 4,898 (25.4) | 644 (25.7) | |
| Current smoker | 8,083 (41.9) | 1,113 (44.4) | |
| Regular exercise (%)a | 12,972 (67.0) | 1,755 (69.7) | 0.01 |
| Alcohol drinking status (%) | <0.01 | ||
| <40 g/day | 18,732 (96.5) | 2,387 (94.5) | |
| ≥40 g/day | 670 (3.5) | 138 (5.5) | |
| Glucose (mg/dL) | 100.1±15.6 | 108.5±20.3 | <0.01 |
| HbA1c (%) | 5.29±0.62 | 5.56±0.76 | <0.01 |
| Total cholesterol (mg/dL) | 194.6±32.5 | 204.8±34.0 | <0.01 |
| TG (mg/dL) | 124.1±73.2 | 197.0±104.5 | <0.01 |
| LDL-C (mg/dL) | 121.7±29.9 | 124.2±33.1 | <0.01 |
| HDL-C (mg/dL) | 48.0±11.1 | 41.3±8.8 | <0.01 |
| BUN (mg/dL) | 15.7±3.9 | 15.5±3.6 | 0.05 |
| Creatinine (mg/dL) | 1.09±0.16 | 1.12±0.14 | <0.01 |
| GFR (mL/min/1.73 m2) | 74.1±9.9 | 72.0±10.3 | <0.01 |
| CKD | 975 (5.0) | 233 (9.2) | <0.01 |
Data are reported as mean±SD or as number (%). a≥Once/week. BMI, body mass index; BP, blood pressure; BUN, blood urea nitrogen; CKD, chronic kidney disease; DM, diabetes mellitus; GFR, glomerular filtration rate; HbA1c, hemoglobin A1c; HDL-C, high-density lipoprotein cholesterol; HTN, hypertension; LDL-C, low-density lipoprotein cholesterol; MetS, metabolic syndrome; TG, triglycerides; WC, waist circumference.
| n (%) | |
|---|---|
| MetS | 2,529 (11.5) |
| Components | |
| Central obesity | 4,018 (18.3) |
| Raised BP | 7,789 (35.4) |
| Raised FPG | 10,503 (47.7) |
| Raised TG | 6,524 (29.7) |
| Reduced HDL-C | 5,653 (25.7) |
FPG, fasting plasma glucose. Other abbreviations as in Table 1.
During a mean follow-up of 8.7 years, AF (including AFL) occurred in 168 subjects (0.8%, AF=166, AFL=2). Figure 2 depicts the cumulative incidence rates of AF by baseline MetS status. Although overall AF incidence rates were low (8.81/10,000 person-years), AF incidence rates were higher in subjects with MetS (13.32/10,000 person-years) than in subjects without MetS (8.25/10,000 person-years). This trend was consistently observed after age-adjustment. The age-adjusted AF incidence rates were 5.94 and 4.29/10,000 person-years in subjects with and without MetS, respectively (Table 3).
Cumulative incidence rates of atrial fibrillation (AF) by baseline metabolic syndrome status.
| No MetS (n=19,452) |
MetS (n=2,529) |
|
|---|---|---|
| No. of events | 140 | 28 |
| Person-years | 169,729 | 21,019 |
| Incidence rate (10,000 person-years) | 8.25 | 13.32 |
| Age-adjusted incidence rates (10,000 person-years) | 4.29 | 5.94 |
AF, atrial fibrillation; MetS, metabolic syndrome.
As shown in Table 4, MetS was associated with an increased risk of AF. In the Cox proportional hazard regression model, the age-adjusted HR for AF in subjects with MetS was 1.62 (95% CI 1.08–2.44, P=0.02). In multivariate models adjusted for age, smoking status, regular exercise, drinking status, and CKD, the HR for AF in subjects with MetS was 1.57 (95% CI 1.04–2.38, P=0.03). Among the components of MetS, central obesity (HR 1.62, 95% CI 1.14–2.29, P<0.01) and raised BP (HR 1.43, 95% CI 1.05–1.94, P=0.02) were associated with an increased risk of AF in the age-adjusted model. However, central obesity (HR 1.62, 95% CI 1.13–2.33, P<0.01) was the only statistically significant risk factor after multivariate adjustment (Table 4). Other variables that were adjusted for the multivariate analysis did not have a statistically significant association with AF, except age (HR 1.72, 95% CI 1.39–2.11, P<0.01, for a 1-SD increment). To account for confounding by a change in central obesity over time, we conducted additional analysis using a time-dependent Cox proportional hazard regression model by updating the central obesity annually or biennially to the time just before diagnosis of AF/AFL. After age-adjustment, central obesity (HR 1.69, 95% CI 1.22–2.34, P<0.01) remained significantly associated with an increased risk of AF (Table S1).
| Age-adjusted HR (95% CI) |
P value | Multivariate-adjusted HR (95% CI) |
P value | |
|---|---|---|---|---|
| MetSa | 1.62 (1.08–2.44) | 0.02 | 1.57 (1.04–2.38) | 0.03 |
| Componentsb | ||||
| Central obesity | 1.62 (1.14–2.29) | <0.01 | 1.62 (1.13–2.33) | <0.01 |
| Raised BP | 1.43 (1.05–1.94) | 0.02 | 1.35 (0.99–1.86) | 0.06 |
| Raised FPG | 1.16 (0.85–1.57) | 0.35 | 1.05 (0.77–1.44) | 0.75 |
| Raised TG | 0.84 (0.60–1.19) | 0.33 | 0.71 (0.49–1.02) | 0.07 |
| Reduced HDL-C | 1.04 (0.74–1.47) | 0.81 | 1.06 (0.74–1.53) | 0.74 |
Cox proportional hazard models were used to estimate HR and 95% CI. aMultivariate-adjusted HR=adjusted for age, smoking status, regular exercise, alcohol drinking status, and CKD. bMultivariate-adjusted HR=adjusted for age, smoking status, regular exercise, alcohol drinking status, CKD, and other components of MetS. CI, confidence interval; HR, hazard ratio. Other abbreviations as in Tables 1–3.
In the present study, we have shown that middle-aged East Asian men with MetS had a 57% increased risk for the development of AF during a mean follow-up of 8.7 years. Among the components of MetS, central obesity was the strongest predictor for the development of AF. Other components of MetS were not significantly associated with an increased risk of AF.
MetS and AFMetS is a cluster of risk factors for cardiovascular disease and each individual component has been suggested as having a relationship with an increased risk of AF. However, the exact mechanism of the relationship between MetS and AF is not well established. Several studies have reported that MetS is associated with AF.7–9,15 A prospective, large observational cohort study based on health examinations in a Japanese population reported that MetS, according to the guidelines of the National Cholesterol Education Program Third Adult Treatment Panel,16 was associated with an increased risk of AF (age and sex-adjusted HR 1.88). In addition, components of MetS were also associated with an increased risk of AF (age and sex-adjusted HR of obesity, elevated BP, low HDL-C, and impaired FPG were 1.64, 1.69, 1.52, and 1.4, respectively); the only exception was elevated TG.7 Other studies of Western populations also revealed that MetS was a risk factor for AF (HR 1.20–1.67).8,15
Our results are consistent with those of the previous studies, although the definition of MetS was different. We found that MetS was associated with an increased risk of AF (HR 1.57). We also found an increased risk of AF associated with central obesity, which is known to be a factor for increased risk.8,15 In addition to baseline central obesity, central obesity as a time-dependent covariate was also associated with an increased risk of AF. Several mechanisms have been proposed to explain the association between obesity and AF. There is a direct correlation between obesity and left atrial size, and increased left atrial size is an important precursor of AF caused by left atrial remodeling.17,18 Epicardial adipose tissue, which correlates with WC and degree of intra-abdominal adipose tissue, has been independently associated with AF in previous studies.19,20 In addition, obesity-related oxidative stress, inflammation,21–23 neurohormonal activation,24 and obstructive sleep apnea25,26 can facilitate the development of AF. Central obesity is the essential risk factor for the diagnosis of MetS, and is regarded as an early step in the etiological cascade leading to full MetS.27 Although mean BMI is lower in Asian populations than in non-Asian populations, Asian populations tend to have a higher incidence of central obesity at a given BMI. Therefore, central obesity could provide more predictive information for the development of AF in Asian populations. Previous Asian studies revealed that central obesity is a more significant risk factor for AF than general obesity (by BMI)28 and increases the risk of AF regardless of BMI except for the obese group (BMI >30 kg/m2).29
Unlike previous studies,7,8,15 however, we found that other components of MetS were not associated with the development of AF. HTN is a well-established risk factor for AF.30,31 Raised BP, a component of MetS, is not identical to HTN. In this study, raised BP also included BP values that were lower, as compared with HTN, which could explain why raised BP was not statistically significant after our multivariate-adjusted analysis (HR 1.35, P=0.06). Type 2 DM32–34 and reduced HDL-C35 have also been questioned as risk factors for AF, especially in men. Raised TG was not associated with an increased risk of AF in previous studies.7,8,15,35 The incidence of AF and the prevalence of MetS in our study were lower than in previous studies,7–9 which could be explained by ethnic differences, age, or the occupation of the study populations. Most of the subjects in our study were manual laborers.
Study Strengths and LimitationsThe strengths of our study are the large sample size and the well-organized cohort design characterized by consistent follow-up because of very low turnover rates in employment, the long average length of service (>18 years), and the obligation for health examinations, including ECGs (at least once every 2 years). Another advantage of our study is that AF was identified purely by 12-lead ECG, not by self-reported questionnaire or a code defined by the International Classification of Disease.
However, our study had several limitations. First, it was a retrospective cohort study with inherent limitations. Second, information regarding smoking, exercise, alcohol intake, and medical history were self-administered, thus allowing recall bias. Third, the study population was not representative of the general population because all subjects were working individuals and middle-aged men. However, this was also a strength of our study because there are few data regarding middle-aged men. AF constitutes a significant economic burden worldwide.36,37 From an economic standpoint, prevention of AF in this group is important because middle-aged men are the main agents of economic activity. Finally, the diagnosis of AF was based only on annual or biennial 12-lead ECGs. Thus, the diagnosis of AF, especially paroxysmal AF, may have been underestimated.
This large and long-term follow-up cohort study demonstrated that MetS was associated with an increased risk of AF in middle-aged East Asian men. Among the components of MetS, central obesity was the most potent risk factor for the development of AF. These finding suggested that strategies to reduce the development of MetS, especially central obesity, might reduce the risk of AF in middle-aged East Asian men.
We acknowledge the efforts of the health screening group at Ulsan University Hospital, Ulsan, Republic of Korea.
The authors have no conflicts of interest to declare.
Supplementary File 1
Table S1. Risk of AF by central obesity using time-dependent Cox proportional hazard regression model
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-18-0113
