Chronic Obstructive Pulmonary Disease Increases the Risk of New-onset Atrial Fibrillation and Mortality of Patients with Atrial Fibrillation

Background: Although a few previous studies have analyzed the role of reduced lung function in predicting atrial fibrillation (AF), the relationship between the incidence of AF and comorbid chronic obstructive pulmonary disease (COPD) is unclear. We hypothesized that COPD is associated with the occurrence of new-onset AF and clinical outcomes in AF patients. Methods: We analyzed the development of new-onset AF in 501,668 patients without AF and clinical outcomes in 4,541 patients with AF using Korean National Health Insurance Service-National Sample Cohort (NHIS-NSC). Results: Comorbid COPD was found in 4.8% (11,442 of 501,668) of non-AF patients and 18.6% (820 of 4,541) of AF patients. The incidence of AF in COPD patients was significantly higher compared to non-COPD patients (2.6% vs. 0.6%, p<0.001) over the follow-up period (45.5±14.9 months). In a multivariate Cox regression analysis, COPD predicted higher risk of AF independently from other risk factors (HR: 1.41, 95% CI: 1.25-1.60, p<0.001). The allcause mortality of AF patients with COPD was significantly higher in patients who used b-blockers (20.6% vs 13.1% during follow-up, p<0.008). Multivariate Cox regression analysis showed that COPD is still an independent risk factor for all-cause mortality (HR: 1.25, 95% CI: 1.03-1.51, p=0.022), and stroke (HR: 1.19, 95% CI: 1.001.41, p=0.039). Conclusion: The presence of COPD is an independent risk factor for new-onset AF. COPD is independently associated with all-cause mortality and stroke in AF patients.


Introduction
Atrial fibrillation (AF) is the most common sustained arrhythmia 1 and confers an independent risk of stroke and death. 2 A recent retrospective study showed that of comorbid chronic obstructive pulmonary disease (COPD) was associated with an increased likelihood of arrhythmia. Furthermore, it remained a significant predictor of AF/Atrial flutter (AFL) (p<0.0001) after adjusting for other clinical factors. 3 The prevalence of cardiovascular diseases (CVD) was higher in the COPD group than the control group in one retrospective cohort study. 4 But the risk of stroke or mortality associated with COPD in patients with AF is not well known. We hypothesized that COPD is associated with the occurrence of new-onset AF and clinical outcomes in AF patients.

Methods
The institutional review board of Severance Hospital at Yonsei University College of Medicine in Seoul, Republic of Korea approved this study. The institutional review board waived the requirement to obtain informed consent.
To assess the association of COPD and all-cause mortality, cardiovascular mortality, and stroke in AF patients, we used data derived from the Korean NHIS-NSC cohort consisting of 1,025,340 Koreans from 2002 and followed the subjects until 2013.
The National Health Insurance Service (NHIS) is the single insurer managed by the Korean government and the majority of the Korean population (97.1%) are mandatory subscribers, with the remaining 3% of the population being medical aid subjects. All insured individuals and their dependents are required to undergo a periodic (i.e., mostly biennial) general health examination. The NHIS maintains a large health dataset and provides periodic updates on health-related risk factors and baseline biochemical data. 5 We enrolled patients older than 18 years with health examination data from the NHIS-NSC after 2009. Participants who had missing data concerning comorbidities and laboratory results were excluded. We examined the incidence of AF by dividing non-AF patients according to COPD by cohort enrollment to determine the effects of COPD on the occurrence of AF. We used an operational definition to extract data on COPD patients because pulmonary function test (PFT) results were not available in the NHIS-NSC. We searched for subjects with a diagnosis of COPD based on International Classification of Diseases-Tenth Revision (ICD-10) codes and medications prescribed. COPD patients were defined as subjects who met the following 2 criteria: 1) subjects with a diagnosis of ICD-10 codes for COPD or emphysema (J42, J43 [except for J43.0], J44) and 2) subjects with a prescription of more than 1 of the following COPD drugs at least twice per year: long-acting muscarinic antagonists, longacting beta-2 agonists, inhaled corticosteroids, short-acting muscarinic antagonists, short acting beta-2 agonists, or methylxanthine (>1 months). 6 AF was identified using the International Classification of Disease 10th Revision (ICD-10) codes; I48 (atrial fibrillation and atrial flutter), I48.0 (atrial fibrillation), and I48.1 (atrial flutter). Diagnosis was established based on one inpatient or two outpatient records of ICD-10 codes in the database to ensure accuracy. 7,8 To evaluate the accuracy of our definition of AF, we conducted a validation study in two hospitals with 628 randomly chosen patients with the ICD-10 code I48. Their electrocardiograms (ECGs) were reviewed by two physicians (DHK and JBP). The patients were determined to have AF if it was documented by ECG examinations. The positive predictive value was 94.1%.
The primary clinical outcomes were all-cause mortality, cardiovascular mortality, and ischemic stroke. The dates and causes of death were obtained from the qualification data in the cohort database, which was prepared by Statistics Korea.
Cardiovascular mortality was defined as death from a disease of the circulatory system. Causes of cardiovascular mortality were further categorized as MI, stroke any diagnosis of ischemic stroke with concomitant brain imaging studies, including CT or MRI was defined as incident ischemic stroke. The accuracy of the diagnosis of an ischemic stroke in the NHIS claim data has been previously validated. 5 Propensity-score matching was used to reduce potential selection bias associated with an observational study. The cases were matched (without replacement) with controls with a ratio of 1:1 based on the closest possible value of the propensity score (nearest neighbor matching).
Continuous variables were expressed as mean ± standard deviation and categorical variables as counts and percentages. A student's t-test for continuous variables or Chi-square test for categorical variables was used to determine the significance of differences in variables between 2 groups. All-cause mortality and cardiovascular mortality were analyzed using the Kaplan-Meier survival analysis and multivariable Cox-proportional hazard models were applied to determine if there were independent associations between COPD and clinical outcomes.  Table 1 shows the baseline characteristics of the study population (COPD group: n=11,771, non-COPD group: In this study, 820 COPD patients with AF, and 10,622 COPD patients without AF were included. We investigated the difference of AF occurrence according to patients with or without COPD.

Results
Of the 11,771 patients who had COPD at the time of initial enrollment in NHIS-NSC, 303 were newly diagnosed with AF (2.6%), and of the 483,650 patients who had no COPD, 3,113 were newly diagnosed with AF (0.6%). This difference in incidence was statistically significant (p<0.001).  The hazard ratio of AF was higher in patients with COPD who used beta-agonist-containing drugs (Table 3). In the propensity score matched population, the incidence of AF was higher in patients treated with drugs, including beta-agonists, is higher (HR 1.46; p=0.003) compared to that in patients treated with other drugs (HR 1.31, p=0.004).
We investigated the risk of all-cause mortality, cardiovascular

Discussion
In this large, retrospective general population-based cohort study, we found that COPD was associated with higher newonset AF in the general population. Furthermore, COPD was an independent risk factor and associated with the risk for stroke as well as all-cause mortality in AF patients. Associations between COPD and stroke were significant after adjusting for variable clinical factors.
Our results showed that people with COPD have an   approximately 20% increased risk of stroke, which was similar to findings from previous studies. [9][10][11][12] To our knowledge, this is the first time that associations with COPD and AF and the major clinical outcomes were investigated in the general population. COPD is the fourth leading cause of mortality and morbidity in the world. 13  The coexistence of COPD and AF is common and the interplay between the two is complex. COPD patients are at an increased risk of incident AF. [22][23][24] Additionally, the adverse effect of incident AF on the clinical outcomes in COPD has also been proposed. 18,20 However, few studies have evaluated the influence of COPD  25 The QTc interval reflects the atrial expiratory refractory period (AERP), suggesting that the QTc interval may be an atrial risk index associated with cardiovascular risk and cardiovascular-specific treatment strategy evaluation. 26 Hypercapnia has also been implicated in AF occurrence in COPD. Hypercapnia causes a marked and uniform increase in atrial refractoriness and marked slowing of atrial conduction. 27 COPD contributes to ventricular diastolic dysfunction. Left ventricular diastolic dysfunction is associated with the severity of the disease and may provide another possible pathophysiological mechanism for AF initiation and persistence. [28][29][30] Oxidative stress and inflammation represent a major pathophysiological mechanism in COPD, but are also associated with AF initiation and permanent preservation. [31][32][33][34] The use of short-or long-acting beta-agonists may increase the risk of AF in COPD. 35 Beta-agonist drugs can potentially cause arrhythmias because of their effects on chronotropy, depolarization and repolarization mediated by the b-adrenergic receptor. 36 In this cohort, the higher Cox proportional hazards of AF in patients with COPD, particularly those who took beta agonists, is consistent with this.
This study has some limitations. First, this study is subjected to all the limitations inherent in a cohort analysis. Second, the risk of cardiovascular mortality and stroke in COPD patients with AF was calculated for a period of 5 years only. Therefore, increasing age, one of the strongest risk factors for cardiovascular mortality and stroke, could not be accounted for as time progressed. Third, detailed parameters reflecting lung function, such as FEV1, were unavailable. Finally, it is unlikely that all COPD patients were identified although we validated COPD based on disease codes and medication. Therefore, the prevalence rate of COPD was likely to have been underestimated in this study.

Conclusions
COPD was associated with increased incidence of AF, and increased mortality and incidence of stroke in patients with AF.
After multivariate adjustment, COPD was still an independent risk factor for all-cause mortality and stroke, but not a risk factor for cardiovascular mortality.

Sources of Funding
This study was supported by a research grant from the Korean Heart Rhythm Society (KHRS 2017-1).