You-Jung Choi, Eue-Keun Choi, Kyung-Do Han, Jiesuck Park, Inki Moon, Euijae Lee, Won-Seok Choe,So-Ryoung Lee, Myung-Jin Cha, Woo-Hyun Lim, Seil Oh

Abstract BACKGROUND Inflammatory bowel disease (IΒD), a chronic inflammatory disease of the gastrointestinal tract, could play a role in the pathophysiology of atrial fibrillation(AF).AIM To investigate the association between IΒD and AF development.METHODS We performed a population-based cohort study using records in the Korean National Health Insurance Services database between 2010 and 2014. A total of 37696 patients with IΒD (12349 with Crohn's disease and 25397 with ulcerative colitis) were identified. The incidence rate of newly diagnosed AF in patients with IΒD was compared with that in a 3 times larger cohort of 113088 age- and sex-matched controls without IΒD.RESULTS During 4.9 ± 1.3 years of follow-up, 1120 patients newly diagnosed with AF (348 in the IΒD group and 772 in controls) were identified. After adjustments using multivariable Cox proportional hazards, patients with IΒD were at a 36% [95%confidence interval (CI) 20%-54%] higher risk of AF than controls. The association between IΒD and the development of AF was stronger in younger than in olderOpen-Access: This is an openaccess article that was selected by an in-house editor and fully peerreviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC ΒY-NC 4.0)license, which permits others to distribute, remix, adapt, build upon this work non-commercially,and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See：http：//creativecommons.org/licen ses/by-nc/4.0/Manuscript source： Unsolicited manuscript Received: February 2, 2019 Peer-review started: February 6,2019 First decision: February 21, 2019 Revised: April 22, 2019 Accepted: April 29, 2019 Article in press: April 29, 2019 Published online: June 14, 2019 P-Reviewer: Cardoso CRL, Ksel IY,Wittmann T S-Editor: Yan JP L-Editor: A E-Editor: Zhang YL patients. Patients without cardiovascular risk factors showed a higher risk of AF primarily. Additionally, patients receiving immun-omodulators [Hazard ration(HR) 1.46, 95%CI 1.31-1.89], systemic corticosteroids (HR 1.37, 95%CI 1.10-1.71),or biologics agents (HR 2.38, 95%CI 1.51-3.75) were at higher risk of AF than patients without them.

Key words: Atrial fibrillation; Inflammatory bowel disease; Colitis, Ulcerative; Crohn's disease

INTRODUCTION

CONCLUSION

IΒD significantly increased the risk of AF, and the impact of IΒD on developing AF was in patients with moderate to severe disease.

Atrial fibrillation (AF) is the most common arrhythmia observed in clinical practice.Recently, inflammation is being recognized as a pathogenic contributor to the development of AF[7]. Previous research has shown a significant association between serum inflammatory mediators such as CRP, tumor necrosis factor-α, interleukin (IL)-2, IL-6, and IL-8, and the development and persistence of AF[8,9]. Several cardiovascular disorders, notably coronary atherosclerosis, are associated with inflammation, and cytokines are known to affect plaque rupture and thrombus formation resulting in myocardial infarction[10]. However, it is unclear whether autoimmune-mediated inflammation contributes to the development and persistence of AF.

Considering that pathogenesis of AF is being increasingly linked to systemic inflammation, IΒD may a potential risk factor for AF. However, there is limited information regarding the association between the risk of development of AF and the presence of IΒD. Therefore, we performed a nationwide population-based study to investigate the association between IΒD and the development of AF, including in young individuals and those without cardiovascular risk factors.

MATERIALS AND METHODS

Data sources

Most Koreans (97.2%, approximately 50 million individuals) are enrolled in the mandatory National Health Insurance Service (NHIS) provided by the Korean government. This database includes all healthcare utilization-related information,including in-patient and outpatient services, diagnoses, and prescriptions recorded for reimbursement by the Health Insurance and Review Agency.

Additionally, a registration program for Rare Intractable Diseases (RID), a part of the NHIS, offers special support to patients with RID including IΒD, since 2006. The diagnostic accuracy of IΒD based on the RID registry was 95.5%-100% and 92.5%-95.5% in terms of sensitivity and specificity, respectively, as assessed by a previous epidemiological study[11]. This study used data from the NHIS and the RID program database, which stores diagnoses in the form of the International Classification of Disease, Tenth Revision, Clinical Modification (ICD-10-CM) codes.

43. Not worthy of a king: The mothers in the Princess and the Pea and de Villeneuve s version of Beauty and the Beast are also concerned about the royal heritage and perceived worthiness of their sons wives, although they do not resort to wicked ways to get rid of the prospective brides.Return to place in story.

Study population

We identified patients diagnosed with IΒD between January 1, 2010 and December 31,2014 from the NHIS database (ICD-10-CM codes： K50.0-50.9 for CD and K51.0-51.9 for UC) and confirmed these diagnoses based on special codes (V130 for CD and V131 for UC) provided by the RID program. The special codes of the RID program are assigned to a patient with IΒD who meets the following diagnostic criteria determined by certified physicians： (1) Diarrhea and abdominal pain lasting more than 16 wk; (2)Endoscopic features corresponding to IΒD; and (3) Histopathological confirmation based on endoscopic biopsy. We excluded patients who had been diagnosed with AF prior to or within a month from the initial diagnosis of IΒD.

Endpoint

The primary endpoint was new-onset non-valvular AF, that occurred during the follow-up period. AF was defined using ICD-10-CM codes (I480-I484, I489), which were registered by the physicians responsible for diagnosis. To ensure diagnostic accuracy and exclude patients with transient AF, patients with AF were defined as those with AF, which have been diagnosed with discharge, or had been confirmed more than twice in an outpatient clinic. In addition, to include only non-valvular AF,patients diagnosed with mitral stenosis (I050, I052, I059) or those with mechanical heart valves (Z952-Z954) were and excluded from the analysis. The definition of AF was validated in previous nationwide studies[12-14].

Covariates

Βaseline comorbidities and medication history were obtained from the medical claims data in the NHIS database by ICD-10-CM and prescription codes. We investigated the comorbidities of patients when they were first enrolled in the RID registry due to IΒD.Βaseline comorbidities included hypertension, diabetes mellitus (DM), dyslipidemia,ischemic heart disease, venous thromboembolism, end-stage renal disease (ESRD) and hyperthyroidism. Definitions of covariates were verified from previous papers and have been summarized in S1 Table[12,13]. Low income was defined as participation in the medical aid program or a monthly income corresponding to the lowest 25% of the total population. Cities with a population ＞ 1 million were considered urban areas.

Of note, 5-aminosalicylic acid (5-ASA) is the preferred primary drug used to treat mild-to-moderate IΒD in Asians, while drugs such as immunomodulators (azathioprine, methotrexate, cyclosporine, and tacrolimus) or biological agents(infliximab, adalimumab, and golimumab) are recommended as first- or second-line therapy for moderate-to severe IΒD[15]. Therefore, we estimated the severity of IΒD based on a previous report of describing treatment of IΒD; those receiving systemic corticosteroids in combination with 5-ASA, immunomodulators (azathioprine,methotrexate, cyclosporine, and tacrolimus) or biological agents (infliximab,adalimumab, and golimumab) were classified as having moderate-to-severe IΒD in this study[16].

Statistical analysis

Demographic and baseline clinical characteristics are reported as means with standard deviation (SD) for normally distributed continuous variables and as numbers with percentages for categorical variables. The independent samples t-test was used for continuous variables and the chi-square test for categorical variables for intergroup comparisons.

The incidence rates (IR) of AF are presented per 1000 person-years and were calculated as the number of patients with AF divided by the total population for each year because the total population changes every year owing to death and emigration.The cumulative IR of AF in patients with IΒD and matched control group were compared using Kaplan-Meier censoring estimates and the log-rank test. Univariate and multivariate Cox proportional hazards regression models were used to evaluate the association between IΒD and the risk of development of AF. Adjusted hazard ratios (HR) have been presented with a 95% confidence interval (CI) and adjusted for known risk factors of AF and IΒD, including age, sex, place, income, DM,hypertension, dyslipidemia, stroke, ischemic heart disease, venous thromboembolism,and ESRD. A two-sided P value ＜ 0.05 was considered statistically significant. All statistical analyses were performed using the SAS software version 9.3 (SAS Institute,Cary, NC, United States) and the SPSS software version 23.0 (SPSS Inc., Chicago, IL,United States).

Ethical considerations

The NHIS database is an open database with patient information anonymized to protect privacy. This information is accessible to all researchers whose study protocols have been approved by the official review committee. Our study was performed in accordance with the Declaration of Helsinki.

RESULTS

Baseline characteristics

We enrolled 37696 patients with IΒD (12349 patients with CD and 25347 patients with UC). We selected a 3-fold larger group of age and sex-matched individuals without IΒD (total 113088 patients： 37047 matched with patients with CD and 76041 matched with patients with UC) as the control group.

Βaseline clinical characteristics of patients have been summarized in Table 1. The percentage of patients with IΒD was greater in urban than in rural areas (50.4 vs 46.8%, respectively, P ＜ 0.001). Additionally, the percentage of patients with IΒD showed a higher income level (23.3 vs 19.8% respectively, P ＜ 0.001). Prevalence of DM, hypertension, and dyslipidemia was lower in patients with IΒD than in the control group, whereas prevalence of ischemic heart disease, venous thromboembolism, ESRD, and hyperthyroidism was higher in those with IΒD than in the control group (P ＜ 0.001 for all comparisons except dyslipidemia). Among patients with IΒD, those with CD were younger than those with UC (30.4 ± 14.2 years vs 43.8 ±15.6 years), and the former group members more frequently received immunomodulators (55.8 vs 13.0%), systemic corticosteroid (58.64 vs 56.67%) and biological agents (13.3 vs 1.7%).

Primary endpoint: New onset AF

During the mean follow-up of 4.87 ± 1.28 years, we identified 1120 cases of incident AF (IΒD group： 348; control group： 772), with a crude IR of 1.90 per 1000 person-years in patients with IΒD and 1.40 in matched controls. The incidence of AF was higher in the IΒD group than in the control group (P ＜ 0.001 by the log-rank test, Figure 1).After adjustment for potential confounding factors, IΒD was significantly associated with the incidence of AF (HR 1.36, 95%CI 1.20-1.55, Table 2). Within the IΒD group,AF was significantly associated with both CD and UC compared to matched non-IΒD subgroups (CD： HR 1.91, 95%CI 1.47-2.49; UC： HR 1.24, 95%CI 1.07-1.43, Table 2).

Additionally, patients who received medications (immunomodulators, systemic corticosteroid, and biological agents) for moderate-to-severe IΒD showed a higher risk of AF than those who did not receive such medications (Table 3). The highest risk of AF was observed in the subgroup receiving biological agents (HR 2.37, 95%CI 1.50-3.74) compared with those receiving immunomodulators (HR 1.45, 95%CI 1.13-1.88)and systemic corticosteroids (HR 1.37, 95%CI 1.10-1.71).

Subgroup analysis performed to assess the risk of atrial fibrillation in patients with inflammatory bowel disease

Βoth patients with UC and CD subgroups showed an increased risk of AF (Figure 2).Patients with IΒD aged below 45 years showed a higher risk of AF than those belonging to the control group, and both UC and CD groups showed similar trend in age subgroups. In contrast, no significant difference was observed in those aged more than or equal to 45 years. Also, the risk of AF development was not significantly different between men and women. The association between IΒD and AF was weaker in subgroups with well-known cardiovascular risk factors (hypertension, DM, and dyslipidemia) than in subgroups without such risk factors (P ＜ 0.001). These differences were statistically more significant in patients with CD (P ＜ 0.001) than in those with UC (P = 0.07).

DISCUSSION

Table 1 Baseline characteristics of the study population

This study is the first large population-based cohort study to investigate the association between IΒD and the incidence of AF in Asians. The primary findings of our large population-based cohort study are： (1) IΒD was associated with an increased risk of AF development; (2) both CD and UC increase the risk of AF, with a higher risk in patients with CD than UC; (3) patients receiving immunomodulators, systemic corticosteroids, and/or biological agents showed a higher risk of AF; and (4) the relative risk of IΒD for the development of AF was particularly high in younger patients (aged below 45 years) and in those without cardiovascular risk factors.

Several previous studies have demonstrated the relationship between IΒD and AF[17,18]. A small-scale study comprising 141 IΒD cases showed that the prevalence of AF was higher across all age groups in patients with IΒD than in the control group[17].In a Danish nationwide study, 24499 patients with new-onset IΒD were identified and compared with age- and sex-matched controls[18]. Overall IΒD-associated risk of AF was increased by 26% (95%CI, 16%-36%). We observed that the risk of AF was higher in patients with IΒD, which was in agreement with previous studies. Interestingly, the absolute risk of AF in patients with CD was higher than that in patients with UC (HR for CD and UC being 1.92 and 1.23, respectively).

There is a reasonable body of evidence to support the pathophysiological features between IΒD and the development of AF. Systemic inflammation is known to be a significant contributor to the development of AF[7-9,19]. Recent studies performed in patients with IΒD have demonstrated that atrial electromechanical conduction time is prolonged and that structural and electrophysiological changes occurring in atrial tissues affect left atrial electromechanical function[20,21]. Several previous studies have suggested that systemic inflammation is linked to various pathological processes such as oxidative stress, apoptosis, and fibrosis of cardiomyocytes, all of which lead to structural and electrical remodeling of the atria, promoting the development and persistence of AF[19,22,23]. Moreover, increases in the level of serum inflammatory markers, such as CRP and IL-6, were observed in patient with AF and IΒD, especially CD[8,24]. Chronic autoimmune and inflammatory diseases such as rheumatoid arthritis,ankylosing spondylitis, and psoriasis are known to increase the risk of development of AF[25-27].

Figure 1 Kaplan-Meier survival curves for cumulative incidence of atrial fibrillation. Cumulative incidence was defined as the probability of new-onset atrial fibrillation during follow-up. A: Comparison between inflammatory bowel disease (IBD) group and control group; B: Comparison between Crohn’s disease group and control group; C: Comparison between ulcerative colitis group and non-IBD group. CD: Crohn’s disease; IBD: Inflammatory bowel disease; UC: Ulcerative colitis; AF:Atrial fibrillation.

The positive relationship between the severity of IΒD and the incidence of AF accounts for the association between the two diseases. In a Danish nationwide study,IΒD was classified as flare-up, persistent disease, and remission based on disease activity, and the association between disease activity of IΒD and the development of AF were evaluated[18]. Incidence of AF increased during IΒD flare-ups and periods of persistent disease, but not during periods of remission[18]. Particular medication such as immunomodulators, systemic corticosteroid, and biological agents are recommended as first or second-line therapy for moderate-to-severe IΒD, and 5-ASA is commonly used for the initial treatment of mild IΒD[28-30]. Therefore, in large population-based studies with difficulty in obtaining clinical symptoms and sign, it is preferable to assess the severity of IΒD using medical history instead of a clinical scoring system. In this study, patients with IΒD who received these aforementioned agents may have a moderate-to-severe disease, and have shown a higher risk of AF.This suggests that patients with moderate-to-severe IΒD have a higher risk of AF than those with mild IΒD.

Advancing age is a well-known risk factor for AF[31]. Surprisingly, however, in our study, the association between IΒD to the development of AF gradually weakened with increasing age. Additionally, the association between IΒD and AF was more significant in patients without cardiovascular risk factors including DM, hypertension, and dyslipidemia.

Several studies performed in Western population cohorts have reported a bimodal distribution of the incidence of IΒD with a peak at a younger age (20-30 years) for CD,a later peak (30-40 years) for UC, and a lower second peak at an older age (60-70 years) for both diseases[32]. In contrast, in Asians including Koreans, the incidence of CD showed a peak in patients aged below 20 years and steadily decreased after that,whereas the incidence of UC was observed to plateau in patients aged 25-69 years[11].In the present study, the prevalence of AF showed a peak at a younger age, and the association between IΒD and the development of AF gradually became weaker with increasing age. These findings suggested that systemic inflammatory disease is one of the mechanisms contributing to the development of AF different from classic risk factors.

The incidence of IΒD in Korea (2.9-3.6 per 100000 for CD and 4.2-5.1 per 100000 for UC) was the highest in East Asia, followed by that in Japan, China, and Hong Kong from 2006 to 2012[33,34]. The prevalence of IΒD in Korea is steadily rising, and IΒDrelated healthcare costs are showing a significant increase[35]. This study is the first to demonstrate that IΒD is an independent risk factor for the development of AF in Asians.

As well as having the increased IR with age, Asians have a generally higher incidence of AF-related risks of stroke and intracranial bleeding, compared with non-Asians[36,37]. The incidence rate ratio (IRR) of AF in patients with IΒD in this Korean population-based cohort was not significantly different to that in patients with IΒD in the previously described Danish population-based cohort (Korea： IRR 1.36, 95%CI 1.20-1.55; Denmark： IRR 1.26, 95%CI 1.16-1.36) after fully adjusting for risk factors of AF[18]. The lower prevalence of AF in Asians (approximately 1%) than that observed in non-Asians (approximately 2%) suggests that the risk of incident AF in patients with IΒD is similar in Asians and non-Asians[12,38]. Our findings add new evidence that IΒD,a systemic inflammatory disease, is associated with an increased prevalence of AF in Asians.

Table 2 Risk of atrial fibrillation in patients with inflammatory bowel disease

Limitations

The primary limitation of this study is the observational population-based retrospective design, which does not permit specific determination of the severity of IΒD. Therefore, detailed information regarding the severity, extent, location, and activity of the IΒD, laboratory data, and duration of treatment modalities were not available in our study. Although we used certain prescriptions as a surrogate measure of moderate-to-severe IΒD, there is still the limitation in discriminating between those with moderate-to-severe disease and those with remission.

Additionally, the role of confounders including general condition, presence of malnutrition, smoking, and other significant comorbidities such as cardiomyopathy,heart failure, and cancer could not be accurately evaluated in patients. However, this study used the certified database to identify patients diagnosed with IΒD and to retrieve their medication history. The database of the RID program includes the complete medical history of patients, which is required to provide special financial support, thereby minimizing the possibility of inaccuracies in prescription data.

Secondly, the definitions of diseases were determined using ICD-10-CM codes of claims data from the NHIS database. Misclassification is a potential bias when using diagnostic codes. Also, when determining cases of IΒD, the ascertainment bias could be induced depending on the patient's residences. Additionally, the detail information on AF subtypes such as paroxysmal, persistent, permanent, and perioperative AF could not be accurately identified in our database.

Moreover, we could not obtain accurate information regarding the time interval between the diagnosis of IΒD and the onset of AF. However, patients were followedup for more than 4 years, which is sufficient to detect AF caused by IΒD. We excluded patients demonstrating new-onset AF within a year of being diagnosed with IΒD to reduce the chance of coincidental occurrence of the disease.

In conclusions, the key finding of our study is that IΒD was associated with an increased risk of AF. Furthermore, IΒD-associated AF was more common in younger patients with less comorbidities. Therefore, although IΒD is a common disease in young people, we need to consider screening for AF, as timely intervention can reduce both morbidity and mortality secondary to the complications of AF.

Table 3 Risk of atrial fibrillation in patients with inflammatory bowel disease according to medication

Figure 2 Subgroup analysis. Subgroup analysis to assess the risk of atrial fibrillation in patients with inflammatory bowel disease (IBD) based on age (＜ 30, 30-44,45-59, and ≥ 60 years), sex (men and women), and cardiovascular (CV) risk factors (those with and without CV risk factors). The association between IBD and the risk of development of atrial fibrillation was weaker in older individuals and in those with CV risk factors. 1Incidence rates have been calculated per 1000 patient-years;2Adjusted for age, sex, location, income, diabetes mellitus, hypertension, dyslipidemia, stroke, myocardial infarction, venous thromboembolism, and end-stage renal disease; 3Hypertension, diabetes mellitus, or dyslipidemia. CI: Confidence interval; CV: Cardiovascular; HR: Hazard ratio; IBD: Inflammatory bowel disease.

ARTICLE HIGHLIGHTS

Research background

Systemic inflammatory disease is known to increase the risk of cardiovascular diseases such as stroke, myocardial infarction, and atherosclerosis. Also, it has recently emerged as a risk factor for atrial fibrillation (AF). Inflammatory bowel disease (IΒD) is a chronic progressive inflammatory disease, which can affect the gastrointestinal tract. Considering that pathogenesis of AF is linked to systemic inflammation, IΒD might be a potential risk factor for AF development.

Research motivation

Recently, the incidence and prevalence of IΒD have increased in the young Asian population.The impact of IΒD on cardiovascular disease is an important issue. Therefore, we need to clarify the relationship between IΒD and AF development in the Asian population and young patients

Research objectives

We aimed to investigate the association between IΒD and AF development in Asians.

Research methods

We performed a population-based cohort study using the certified claim database form the Koran National Health Insurance Services between 2010 and 2014. The special codes of the rare and intractable disease program were used to define the patient, which include the complete medical history of patients. A total of 37696 patients with IΒD (12349 with Crohn's disease and 25397 with ulcerative colitis) were identified. The primary endpoint was new-onset AF, which occurred among IΒD patients without a previous history of AF during the follow-up period. AF was defined using the International Classification of Disease, Tenth Revision, Clinical Modification codes (I480-I484, I489).

Research results

During 4.9 ± 1.3 years of follow-up, 1120 patients newly diagnosed with AF (348 in the IΒD group and 772 in controls). IΒD patients had a 36% (95% confidence interval 20%-54%) higher risk of AF than controls. The association between IΒD and the development of AF was stronger in younger than in older patients, and in patients without cardiovascular risk factors.Additionally, moderate-severe IΒD patients, who received immunomodulators, systemic corticosteroids, or biologics agents were at higher risk of AF. These findings supported that systemic inflammatory disease could be an independent risk factor for AF development.However, further studies on the impact of other systemic inflammation on AF need to be performed to provide conclusive evidence.

Research conclusions

IΒD, chronic systemic disease, significantly increase the risk of AF development in Asians, as similar to the Western population. Moreover, the impact of IΒD on AF incident was higher in young patients and those without cardiovascular risk factors. Therefore, our findings supported that systemic inflammatory disease is an independent risk factor for AF, even in patients without classic cardiovascular risk factors.

Research perspectives

It is necessary to pay attention to the occurrence of cardiovascular diseases such as AF in patients with IΒD, even in young age. In particular, a patient receiving immunomodulators, systemic corticosteroids, or biologics agents for moderate-severe disease, had a higher risk of AF development than those without.