Atrial fibrillation and risk of sudden cardiac arrest in young adults

Abstract

Aims

Evidence of an association between atrial fibrillation (AF) and sudden cardiac arrest (SCA) in young adults is limited. In this study, we aim to evaluate this association in a general population aged between 20 and 39 years.

Methods and results

Young adults who underwent health check-ups between 2009 and 2012 were screened from a nationwide healthcare database in South Korea. A history of AF diagnosis before the health check-ups was identified based on the relevant International Classification of Diseases, 10th edition codes reported in the database. Associations between an established diagnosis of AF and the risk of SCA during follow-up were examined. A total of 6 345 162 young people were analysed with a mean follow-up duration of 9.4 years. The mean age was 30.9 ± 5.0 years, and 5875 (0.09%) individuals were diagnosed with AF. During follow-up, SCA occurred in 5352 (0.08%) individuals, and the crude incidence was 0.56 and 0.09 events per 1000 person-years for participants with and without AF, respectively. Individuals with AF had a 3.0-fold higher risk in a multivariate model adjusted for age, sex, lifestyle, anthropometric data, and medical comorbidities (adjusted hazard ratio 2.96, 95% confidence interval 1.99–4.41, P < 0.001). Both incident and prevalent AFs were associated with an increased risk of SCA, with no significant differences between the two groups.

Conclusion

Atrial fibrillation was associated with a significantly higher risk of SCA developing in healthy young adults. Whether the rate or rhythm control influences the risk of SCA in young patients with AF remains to be examined.

Graphical Abstract

AF, atrial fibrillation; CI, confidence interval; SCA, sudden cardiac arrest.

Open in new tabDownload slide

Introduction

Arrhythmic crosstalk between the atrium and the ventricle has been reported in previous studies. Kim et al.¹ reported that atrial fibrillation (AF) is associated with a significantly increased risk of ventricular tachycardia or fibrillation. On the other hand, premature ventricular contraction, the most common arrhythmia originating from the ventricle, is associated with a significantly increased risk of AF.² Although AF is reported to be associated with increased overall mortality, the epidemiological association between AF and sudden cardiac arrest (SCA) has not been fully demonstrated; especially in younger populations aged <40 years.^3–6 Rapid ventricular response commonly observed in patients with AF is one of the mechanisms linking AF and SCA.⁷ Continuously varying R–R intervals might provoke an R-on-T phenomenon that can lead to ventricular fibrillation in patients with AF.^8,9 Epidemiological evidence of an association between AF and SCA is scarce, since atherosclerotic cardiovascular disease (ASCVD) and heart failure are closely associated with both AF and SCA. In other words, whether the increased risk of SCA in patients with AF is due to AF itself or is attributable to the accompanying heart failure or ASCVD remains unclear. Another obstacle is the rarity of SCA events, which makes it difficult to achieve a statistically powered analysis.

Sudden cardiac arrest in young adults has grave effects on both patients and society.^10,11 Yet, the risk factors for SCA in young adults are not fully understood. Identifying risk factors for SCA and defining high-risk groups, especially in young people, is important. Based on a nationwide cohort of young adults aged between 20 and 39 years old, we aimed to investigate the association between AF and SCA. The prevalence of ASCVD and heart failure is relatively low in young adults, thereby minimizing confounding effects, and the low rate of SCA events can be overcome by analysing cohorts with significantly large sample sizes. Atrial fibrillation in young adults is more likely to produce a rapid ventricular response, as well as genetic variations responsible for the disease, enabling direct evaluation of the association between AF and SCA, which is better than that of older adults whose conditions could be complicated by confounding comorbidities, such as ASCVD, chronic kidney disease, or heart failure.

Methods

Database

We used data from the Korean National Health Insurance Service (K-NHIS) database for this study. The K-NHIS mandates subscription by all citizens of the Republic of Korea. Individuals who cannot pay subscription fees are offered separate medical aid. Therefore, all citizens of the Republic of Korea are subject to a government-managed healthcare system, and their medical data are stored for their lifetime. Individuals who die or emigrate to different nations are automatically reported to the K-NHIS and they lose their subscriber status, indicating that there are no uncensored losses to follow-up. In addition, the K-NHIS offers a biennial nationwide health-screening programme for its subscribers, which includes (i) self-reports of lifestyle factors such as alcohol consumption, smoking, and physical exercise level; (ii) physical examinations including body weight, height, waist circumference, and blood pressure; and (iii) laboratory tests such as complete blood cell count, renal function, fasting blood glucose level, and lipid profiles. These unique features of the K-NHIS make it valuable for medical research, particularly epidemiological research.

The K-NHIS permits medical researchers to use data with the approval of both the official review committee of the K-NHIS (https://nhiss.nhis.or.kr/) and the researchers’ local institutional review board. This study was approved by the Institutional Review Board of the Korea University Anam Hospital and the official review committee of the K-NHIS (IRB no. 2021AN0185). The requirement for written informed consent was waived by the Institutional Review Board of the Korea University Medicine Anam Hospital because of the retrospective nature of the study. The principles of the 2013 Declaration of Helsinki and legal regulations of the Republic of Korea were strictly adhered to throughout this study.

Study population

We included people who were aged ≥20 and <40 years and underwent nationwide health screening between 2009 and 2012. Korean adults are recommended to receive nationwide health screening biennially. The South Korean government has reported that 75–80% of adults aged between 20 and 39 years who are assigned to national health screening in a specific year actually undergo health screening. Clinical follow-up data until December 2020 were available. Patients with International Classification of Diseases, 10th edition (ICD-10) codes for SCA before undergoing health screening (from 2009 to 2012) were excluded. Individuals with reports of ICD-10 codes for SCA or death within 1 year of health screening were excluded because an outcome that occurred within such a short time interval may have little association with the independent variable, and AF may be just a phenomenon rather than a risk factor in this circumstance. Also, SCA or death within 1 year of health screening could be confounded by repeated reports of prior SCA events that could have occurred before health screening. For example, SCA survivors who experienced aborted SCA before health screening could have visited the emergency department for other diseases such as pneumonia, but ICD-10 codes for SCA could have been repeatedly used at the initial visit. This event would not be a true SCA event that occurred after the start of clinical follow-up, but a simple repeat report.

Outcome measurement and definition of variables

Sudden cardiac arrest was the primary outcome of this study, which included both aborted and non-aborted events. The secondary outcome was all-cause death. The ICD-10 codes in the emergency department were used to define SCA. Codes to diagnose SCA included ‘cardiac arrest with successful resuscitation (I46.0)’, ‘sudden cardiac arrest (I46.1)’, ‘cardiac arrest, cause unspecified (I46.9)’, ‘ventricular fibrillation and flutter (I49.0)’, ‘instantaneous death (R96.0)’, and ‘death occurring less than 24 h from symptom onset (R96.1)’. Cardiopulmonary resuscitation in the emergency department was also identified as an event associated with SCA. If the aforementioned ICD-10 codes were preceded by reported events of a potential non-cardiac cause of cardiovascular collapse, such as haemorrhagic or ischaemic stroke, asphyxia, suffocation, drowning, anaphylaxis, gastrointestinal bleeding, major trauma, sepsis, lightning, electric shock, or burn within 6 months, the event was not considered SCA.

The presence of AF was determined based on the claim of the ICD-10 codes for AF before the enrolment period (health screening from 2009 to 2012). The K-NHIS offers medical data from 2002 onward. Therefore, a screening period >7 years was obtained for baseline demographics, including AF. Two outpatient claims and one inpatient claim for AF were required for the diagnosis of AF. Participants who were diagnosed with AF <1 year before the health examination were defined as having incident AF. Those who were diagnosed with AF ≥1 year before the health examination were defined as having prevalent AF.

Our strategy for defining SCA, AF, and other medical conditions has been validated in prior studies.^12–14 The ICD-10 codes used in this study are summarized in Supplementary material online, Table S1. The definitions of various conditions, such as smoking status, hypertension, chronic kidney disease, and diabetes mellitus, are provided in Supplementary material online, Table S2.

Statistical analysis

Student’s t-test or Mann–Whitney U test (for non-normally distributed data) was used to compare continuous variables, which are expressed as mean ± standard deviation. Categorical variables are expressed as numbers and percentages and were compared using the χ² test. The incidence of SCA was calculated as event numbers per 1000 person-years of follow-up, and the cumulative incidence was depicted using Kaplan–Meier survival curve analysis with the log-rank t-test used for comparison. The Fine–Gray competing risk model was used to adjust for competing outcomes (non-sudden cardiac death). Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated with Cox-regression analysis using seven models—demographics, social habits, or comorbid conditions that could be associated with AF were included in multivariate models.¹⁵ Model 1 was the non-adjusted model; Model 2 was adjusted for age and sex; Model 3 was adjusted for age, sex, body mass index, income, alcohol intake, smoking, regular exercise, diabetes mellitus, hypertension, chronic kidney disease, heart failure, depression, and thyroid disease (hyper- or hypothyroidism); Model 4 was adjusted for age, sex, waist circumference, income, alcohol intake, smoking, regular exercise, fasting blood glucose, systolic blood pressure, LDL, estimated glomerular filtration rate, heart failure, depression, and thyroid disease; Model 5 was adjusted for dilated cardiomyopathy, hypertrophic cardiomyopathy, cardiac arrhythmia, congenital heart disease (atrial or ventricular septal defects and anomalies of great vessels, valves, or cardiac chambers), ischaemic stroke, pacemaker, implantable cardioverter defibrillator, catheter ablation in addition to Model 3; and Model 6 was adjusted for non-sudden death as a competing risk in addition to the variables used in Model 5. Participants who were newly diagnosed with AF after the start of follow-up were censored in Model 7 (adjusted variables were identical to those of Model 6). In order to decrease the heterogeneity of SCA defined by ICD-10 codes, additional analysis was performed excluding cardiac arrest survivors diagnosed with ‘cardiac arrest, cause unspecified (I46.9)’. The ICD-10 codes for each disease are summarized in Supplementary material online, Table S1. All tests were two-tailed, and P-values <0.05 were considered statistically significant. All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC, USA).

Results

Baseline characteristics

A total of 6 891 400 individuals aged 20–39 years underwent health-screening examinations between 2009 and 2012 (Figure 1). Individuals were excluded if they had a prior diagnosis of SCA (n = 186), missing data (n = 543 540), or died or experienced SCA within 1 year after health-screening examinations (n = 2512). The participants were followed up until December 2020. The mean duration of follow-up was 9.4 ± 1.2 years.

In the total cohort (n = 6 345 162), the mean age was 30.9 ± 5.0 years and 3 775 535 (59.5%) were men (Table 1). Anthropometric data and laboratory markers, such as body mass index, waist circumference, fasting blood glucose, blood pressure, lipid profiles, and liver function test results, were within the normal range. The prevalence of various medical diseases, including hypertension (7.4%), diabetes mellitus (1.9%), dyslipidaemia (6.9%), chronic kidney disease (2.7%), heart failure (0.1%), and AF (0.1%), was relatively low. Participants with AF had a markedly higher prevalence of comorbidities, including hypertension, diabetes mellitus, ischaemic stroke, and thyroid disease (Table 1). Relevant cardiovascular diseases, including heart failure, dilated cardiomyopathy, hypertrophic cardiomyopathy, cardiac arrhythmia, congenital heart disease, and pre-excitation syndrome, were also more prevalent among patients with AF. Although the proportion of heavy drinkers was higher among patients with AF than those without AF (9.4 vs. 8.8%), non-drinkers were more prevalent among patients with AF (41.2 vs. 37.8%, P < 0.001).

Among the participants with AF, 1462 (24.9%) were classified as incident AF and 4413 (75.1%) as prevalent AF. Participants with incident AF were mostly men, non-drinkers and had more regular exercise habits, chronic kidney disease, and thyroid disease (see Supplementary material online, Table S3). Participants with prevalent AF were largely non-smokers and had diabetes mellitus and dyslipidaemia. No clinically significant differences in other baseline characteristics such as age or prevalence of heart failure were observed.

Occurrence of sudden cardiac arrest

During follow-up, SCA occurred in 5352 patients (0.08%; Supplementary material online, Figure S1 and Table S4). The participants with SCA were mostly older men who were current smokers and heavy drinkers (see Supplementary material online, Table S5). Comorbidities, including hypertension, diabetes mellitus, dyslipidaemia, chronic kidney disease, heart failure, cardiomyopathy, and congenital heart disease, were more prevalent among patients with SCA. Markers of metabolic syndrome were also elevated in people who experienced SCA, such as body mass index, waist circumference, systolic blood pressure, diastolic blood pressure, fasting blood glucose, total cholesterol, triglyceride, LDL, and liver function tests.

Of the 59 374 293 person-years of follow-up, 5322 SCA events occurred in the non-AF group. In the AF group, 30 SCA events occurred during the 53 845 person-years of follow-up. Participants with AF had a 6.26-fold higher incidence of SCA (crude incidence, 0.56 per 1000 person-years) than those without AF (crude incidence, 0.09 per 1000 person-years; Table 2 and Figure 2A). Adjustment for non-sudden deaths revealed consistent findings (see Supplementary material online, Figure S2). Demographics, lifestyle habits, and comorbidities were adjusted, and people with AF had a 4.07-fold increased risk of SCA (95% CI 2.83–5.86; P < 0.001; Table 2). After adjusting for various cardiovascular conditions such as cardiomyopathies or congenital heart diseases, the presence of AF was associated with a 3-fold increased risk of SCA (adjusted HR 2.96, 95% CI 1.99–4.41; P < 0.001; Table 2 and Supplementary material online, Figure S3). The incidence and risk of SCA were stratified according to AF duration (incident vs. prevalent AF). There were no significant differences in the incidence or risk of SCA between the incident and prevalent AF groups (Figure 2B and Supplementary material online, Table S6). Further analysis in refined population of SCA (excluding 2849 cases coded with ‘cardiac arrest, cause unspecified [I46.9]’), the association remained consistent, resulting >3-fold higher risk of SCA in patients with AF (see Supplementary material online, Table S7).

Patients with AF had a higher incidence of all-cause mortality (crude incidence, 1.52 vs. 0.57, P < 0.001; Supplementary material online, Figure S4 and Table S8). Adjustment of covariates resulted in 58% increased risk of all-cause death in people with AF compared with that in those without AF (95% CI 1.26–1.98; P < 0.001).

Subgroup analysis

The association between AF and SCA risk was evaluated in various subgroups and a significant association was consistently observed (Figure 3 and Supplementary material online, Table S9). No significant interactions were found between the subgroups except for smoking status. Individuals who were not current smokers showed a stronger association between AF and risk of SCA.

Discussion

In this study, we investigated the association between AF and SCA in young adults. Data were derived from a nationwide cohort of 6.3 million young people in the 20 and 30 s age group who had undergone a nationwide health-screening programme in South Korea. Atrial fibrillation was rare in this cohort. Overall, the incidence of SCA was very low in young adults; however, the presence of AF was associated with a >2-fold increased risk of SCA. The risk of SCA did not differ according to the AF duration, both incident and prevalent. The incidence of SCA in this study as well as previous studies was shown to be low, making statistical analysis difficult.^11,16,17 The major strength of our study was the inclusion of >6.3 million healthy young adults with both anthropometric and laboratory measurements, which was further strengthened by a mean follow-up duration of 9.4 years. If a sudden collapse was to occur in young adults, it is highly likely that they would be transferred to an emergency department. All reported data from all emergency departments across the Republic of Korea would be stored in the K-NHIS, a single exclusive healthcare provider in the nation, making undetected SCA events unlikely. The absence of uncensored loss to follow-up due to the intrinsic nature of the K-NHIS database is another important strength of this study.

Atrial fibrillation and sudden cardiac arrest in the young

Although the incidence of SCA is lower in young adults, the outcome of SCA is devastating necessitating comprehensive evaluation of the underlying aetiology that can lead to primary prevention.¹⁸ Investigations on the aetiology of SCA in individuals (1–50 years old) revealed that 55% had abnormal cardiac conditions, with coronary artery disease being the most common cause.¹⁹ However, coronary artery disease was the main factor among people aged between 40 and 50 years, which differed from that among those in their 20 and 30 s, implying that the risk factors that predisposed this age population to SCA were different.¹⁹

There has been consistent evidence on the link between AF and SCA in middle-aged to older populations. A population-based cohort study of 45- to 64-year-old adults revealed that incident AF was associated with a 2.4-fold increased risk of sudden cardiac death.⁵ Analysis of a community-based out-of-hospital cardiac arrest registry in The Netherlands also reported a 3.1-fold elevated risk of ventricular fibrillation in patients with AF compared with that in those without AF.⁶ However, previous evidence of the association between AF and SCA was derived from the older population, which excluded the younger population (<45 years). Regarding the substantial differences in comorbid conditions between young and old people, the results from middle- to old-age populations are not generalizable to SCA in the younger population. Based on a healthy young cohort with a markedly lower prevalence of cerebrovascular disease, we confirmed a consistent association between AF and SCA in young adults.

Potential mechanism: rhythm

A rapid ventricular response is commonly observed in young patients with AF due to their relatively well-functioning atrioventricular nodes. If antegrade bypass tract conduction exists, extreme tachycardia can occur, leading an to increased risk of sudden collapse and potential death.^20–22 Shorter effective refractory period during rapid heart rate per se may increase the risk of ventricular arrhythmias. That is, supraventricular tachycardia, such as atrial tachycardia with 1:1 atrioventricular conduction, increases the ventricular rate, which may be critical to the initiation of re-entry ventricular tachycardia.⁷ Increased adrenergic tone involved in rapid supraventricular rhythm is also suggested to the trigger ventricular arrhythmia.²³ In addition, consistently varying R–R intervals during AF can precipitate the initiation of ventricular arrhythmias. Sweeney et al.²⁴ reported that bradycardia pacing in patients with implantable cardioverter defibrillators can induce short–long–short R–R intervals and ventricular tachycardia. In addition, programmed ventricular stimulation during electrophysiology studies is a common method for inducing various types of ventricular arrhythmias, such as ischaemia-related or fascicular ventricular tachycardia. Whether continuously varying the R–R interval during AF can work in a manner similar to programmed ventricular stimulation to trigger lethal ventricular arrhythmias has not yet been systemically evaluated. A case report demonstrated an irregular ventricular response initiating ventricular fibrillation in a patient with AF with an implantable cardioverter defibrillator.²⁵ However, such association between AF and SCA (or lethal ventricular arrhythmias) should be further explored to establish a cause-and-effect relationship.¹

Potential mechanism: genetics and substrates

Although patients with AF showed a higher prevalence of acquired risk factors such as hypertension, diabetes mellitus, or dyslipidaemia, their alcohol consumption status was not consistent with prior reports. We observed a higher proportion of non-drinkers in the AF group, which implies that other innate conditions (such as genetic mutations) might predispose an individual to the development of AF. For instance, various cardiomyopathies, such as hypertrophic or dilated cardiomyopathy, are closely associated with SCA and could be a predisposing substrate for the development of AF. Because such cardiomyopathies have a genetic background to their pathophysiology, they commonly develop in young individuals. A recent study reported that loss-of-function variants in the TTN gene were associated with early onset AF.²⁶ In another report, genetic variations in MYH7, MHY6, and LMNA were potentially associated with early onset AF.²⁷ The aforementioned genetic mutations are known to cause hypertrophic and dilated cardiomyopathies which are potentially associated with early onset AF. Therefore, AF (especially early onset) and various cardiomyopathies capable of causing SCA may coexist in young individuals. Similarly, familial AF also shares genetic background with inherited arrhythmia such as long QT syndrome. For instance, loss-of-function mutation in KCNQ1 is associated with long QT syndrome, whereas gain-of-function mutation is associated familial AF. A pleiotropic mutation of KCNQ1 could predispose to both long QT syndrome and familial AF, leading to higher risk of SCA.²⁸ Our data provide epidemiological evidence for the association between AF and SCA, and demonstration of a cause-and-effect relationship through functional studies of various genetic variations is required.

Tachycardia-induced cardiomyopathy resulting in heart failure may be another mechanism connecting AF and SCA in young patients. Although we adjusted for history of heart failure, patients with AF could develop heart failure during their disease course and experience heart failure–related SCA. Atrial fibrillation and heart failure, especially the reduced ejection fraction type, are deleterious combinations that lead to increased mortality.²⁹ Catheter ablation of AF in these populations has demonstrated a significant benefit in terms of reducing mortality.^30,31 Further research is required to determine whether AF is associated with a decreased risk of SCA.

Limitations

Our study has several limitations. First, the incidence of SCA was low in the healthy young study population. The total number of SCA cases was <0.1%. However, the medical law in South Korea requires hospital visits for the declaration of death and a significant proportion of unwitnessed sudden cardiac deaths should have been included in our analysis. In this study, we were unable to differentiate between witnessed and unwitnessed SCA. A nationwide analysis of sudden cardiac death in Denmark, approximately half of all sudden cardiac death cases were reported to be unwitnessed, but the clinical characteristics between witnessed and unwitnessed cases did not significantly differ.³² Therefore, we consider that the association between AF and SCA would be consistent in both witnessed and unwitnessed SCA. Second, the burden of AF was not assessed in our analysis. Although participants with AF were classified according to the time since the first diagnosis of AF (≥1 vs. <1 year), detailed information, such as pattern (paroxysmal AF or non-paroxysmal AF), frequency, and extent, was not available. Stratification of AF might provide more information on the differential risk of SCA according to the AF burden or extent of atrial cardiomyopathy. Further, it is possible that paroxysmal AF was underdetected, especially when AF burden was low. However, we used a screening period of >7 years before enrolment. If AF was not detected and reported in the K-NHIS system for about 7 years, then the chances of under-detection might be low.

Third, identification of a common genetic background or family history was not possible in our cohort. Recent genetic evidence for early-onset AF is highlighted, but performing genetic testing for early onset AF can be practically limiting due to drawbacks such as lack of insurance coverage. In addition, genetic testing is usually limited to tertiary medical centres, and its methods are variable, making it difficult to collect data from a national database. Further explorations focusing on the genetics of young patients with AF and SCA should be warranted.

Conclusions

Atrial fibrillation was significantly associated with an increased risk of SCA in healthy young adults. Both prevalent and incident AFs were associated with an increased risk of young adults developing SCA. Whether an appropriate rate or rhythm control in this population can reduce the risk of SCA remains unclear. The increased risk of SCA in young adults with AF indicates the need for a close surveillance of rhythm and relevant structural substrates.

Supplementary material

Supplementary material is available at Europace online.

Authors’ contributions

J.-I.C. has full access to all the data in this study and takes responsibility for its integrity and analytical accuracy. The concept and design of the study were developed by Y.G.K., J.H.J., K.-D.H., J.-I.C., and Y.-H.K. Y.G.K., J.H.J, K.-D.H., and J.-I.C. analysed and interpreted the data. Y.G.K., J.H.J., K.-D.H., and J.-I.C. drafted the manuscript. The statistical analyses were performed by Y.G.K, K.-D.H., and J.-I.C. Data were collected by Y.G.K., S.-Y.R., K.-D.H., J.H.J., J.S., H.S.L., Y.Y.C., and J.-I.C.

Funding

This work was supported by a Korea University Grant (J.-I.C.), a grant from Korea University Anam Hospital, Seoul, Republic of Korea (J.-I.C.), a grant from Korean Heart Rhythm Society (no. KHRS2023-3 to Y.G.K.), and in part by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT, Ministry of Science and ICT; no. 2021R1A2C2011325 to J.-I.C.). The funders had no role in data collection, analysis, or interpretation; trial design; patient recruitment; or any other aspect pertinent to the study.

Conflict of interest: none declared.

Data availability

The data used in this study are available in this article. Raw data used in this article cannot be shared publicly because of privacy concerns and legal regulations of the Republic of Korea. Raw data were stored and analysed on the designated server managed by the K-NHIS.

References

Author notes