Risk factors for the development of incident atrial fibrillation in patients with cardiac implantable electronic devices

European Journal of Internal Medicine xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

European Journal of Internal Medicine journal homepage: www.elsevier.com/locate/ejim

Original Article

Risk factors for the development of incident atrial fibrillation in patients with cardiac implantable electronic devices ⁎

Kazuo Miyazawaa,b, , Yusuke Kondoc, Miyo Nakanob, María Asunción Esteve-Pastora,d, José Miguel Rivera-Caravacaa,e, Keitaro Senooc,f, Yoshio Kobayashib, Gregory Y.H. Lipa,g a

Institute of Cardiovascular Sciences, University of Birmingham, United Kingdom Department of Cardiovascular Medicine, Chiba University Graduate School of Medicine, Chiba, Japan c Department of Advanced Cardiovascular Therapeutics, Chiba University Graduate School of Medicine, Chiba, Japan d Department of Cardiology, Hospital Clínico Universitario Virgen de la Arrixaca, Instituto Murciano de Investigación Biosanitaria (IMIB-Arrixaca), CIBERCV, Murcia, Spain e Department of Hematology and Clinical Oncology, Hospital General Universitario Morales Meseguer, Instituto Murciano de Investigación Biosanitaria (IMIB-Arrixaca), Murcia, Spain f Department of Arrhythmia, Koseika Takeda Hospital, Kyoto, Japan g Thrombosis Research Unit, Department of Clinical Medicine, Aalborg University, Aalborg, Denmark b

A R T I C L E I N F O

A B S T R A C T

Keywords: Atrial fibrillation Cardiac implanted electronic device Risk score

Introduction: Cardiac implantable electronic devices (CIEDs) can detect atrial fibrillation (AF) early and accurately. Risk factors for the development of new-onset AF in patients with CIEDs remains uncertain. Methods: Patients with CIEDs who visited Chiba University Hospital between January 2016 and December 2016 were enrolled. We only included patients without single chamber CIEDs or a known history of AF. Results: Of 371 patients with CIEDs, 78 (21.0%; median age 61.0 years, 65.5% male) developed new-onset AF. Multivariate analysis demonstrated that independent predictors for the development of new or incident AF were age ≥65 years (odd ratio [OR] 2.76, 95% confidence interval [CI] 1.54–4.96, P = 0.001), diabetes mellitus (OR 2.24, 95% CI 1.20–4.19, P = 0.011), congestive heart failure (OR 1.94, 95% CI 1.06–3.54, P = 0.031), and left atrial volume index > 34 ml/m2 (OR 3.51, 95% CI 1.96–6.25, P < 0.001). Based on these 4 clinical factors (age ≥ 65, diabetes mellitus, congestive heart failure, left atrial volume index > 34 ml/m2) there was a good predictive ability for new AF development (AUC 0.728) and clinically usefulness using decision curve analysis. Conclusions: A substantial number of patients with CIEDs develop new-onset AF. Four clinical factors (age ≥ 65, diabetes mellitus, congestive heart failure, left atrial volume index > 34 ml/m2) independently predicted newonset AF and may provide an approach to clinically useful risk assessment for incident AF.

1. Introduction Atrial fibrillation (AF) is associated with an increased risk of stroke, heart failure, and mortality [1]. Therefore, early detection of new-onset incident AF may allow the timely initiation of treatment to prevent not only from progression of AF, but also from the consequences of AF. However, a substantial number of patients has no symptoms regarding AF [2], and are often under-diagnosed by conventional diagnostic methods such as physical examinations, 12-lead electrocardiogram (ECG), and 24-hour Holter ECG [3]. Unfortunately, asymptomatic and short-term AF is sometimes newly diagnosed after admission following an acute stroke or transient ischemic attack (TIA) [4]. Cardiac implantable electronic devices (CIEDs) can automatically

⁎

record all spontaneous episodes of arrhythmia using programmable detection criteria, and continuous ECG monitoring allows the detection of intermittent and short-term AF regardless of the presence of symptoms. Previous studies demonstrated that atrial high rate episodes (AHREs) detected by CIEDs have a high correlation with clinically documented AF [5], and are independently associated with an increased risk of ischemic stroke and systemic embolism [6–8]. Risk factors for the development of new-onset AF in patients with CIEDs remains uncertain. In the present study, we investigated incident AF in a cohort of patients with CIEDs, and determine clinical risk factors that were independently associated with the development of new-onset AF. Second, we tested a risk factor cluster that was associated with a good probability of new-onset AF development amongst CIED patients.

Corresponding author at: University of Birmingham Institute of Cardiovascular Sciences, City Hospital, Birmingham B18 7QH, England, United Kingdom. E-mail address: [email protected] (K. Miyazawa).

https://doi.org/10.1016/j.ejim.2018.02.019 Received 22 October 2017; Received in revised form 30 January 2018; Accepted 21 February 2018 0953-6205/ © 2018 Published by Elsevier B.V. on behalf of European Federation of Internal Medicine.

Please cite this article as: Miyazawa, K., European Journal of Internal Medicine (2018), https://doi.org/10.1016/j.ejim.2018.02.019

European Journal of Internal Medicine xxx (xxxx) xxx–xxx

K. Miyazawa et al.

2. Material and methods

deviation or median with interquartile range (IQR), and categorical variables as frequency (percentage). Continuous variables were compared using Student's t-test or Mann-Whitney U test, as appropriate. Categorical data were analyzed using the Chi-square test. To assess risk factors for the development of new-onset AF, we used logistic regression model by adding variables that were significant (P value < 0.10) from the univariate analysis. Receiver-operating characteristic (ROC) curve analysis was performed to estimate continuous variables with the risk scores for the development of new AF based on an estimated area under the curve (AUC), which was used as an indicator of predictive value of the risk scores (often referred to as c-indexes). Comparisons of ROC curves were performed according to DeLong et al. [15]. To assess the risk scores for clinical utility, we also performed decision curve analysis, which was established by Vickers and Elkin for evaluating and comparing the clinical net benefit of prediction models [16]. The clinical net benefit is calculated by summing the benefits (true positive) and subtracting the harms (false positive). The result of this analysis is presented with the selected probability threshold plotted on the x-axis and the benefit of the evaluated model on the y-axis. SPSS Statistic ver. 24 (IBM, New York, NY, USA) and STATA 13 (STATA Inc., USA) were used for the analysis. P values < 0.05 were considered statistically significant.

We enrolled the patients receiving pacemakers, implantable cardioverter defibrillator (ICD), and cardiac resynchronization therapy (CRT) with or without defibrillation, who visited Chiba University Hospital between January 2016 and December 2016. Patients were eligible for inclusion if they had at least 1 follow-up visit and device interrogation after CIED implantation. Patients who had a prior history of AF or had single-chamber CIED implanted were excluded. If pacing mode of VVI or AAI was set even in patients with dual-chamber CIEDs, we excluded these patients. Of the total number of CIED patients (n = 504) attending our unit, total of 371 patients (73.6%) were included in the present analysis. The present study was conducted with the approval of the Ethics Committee of Chiba University Hospital. We retrospectively reviewed the patients' medical records, and collected clinical information on age, gender, body surface area, systolic/diastolic blood pressure, alcohol consumption, indication for CIEDs (sick sinus syndrome, atrioventricular block, ventricular tachycardia (VT) or fibrillation (VF), and chronic heart failure), past history of stroke or TIA, underlying heart disease (coronary artery disease, hypertrophic cardiomyopathy, dilated cardiomyopathy, and valvular heart disease), comorbidity (hypertension, diabetes mellitus, peripheral artery disease, and chronic obstructive pulmonary disease), and medication (beta-blocker, ACE inhibitor/ARB, statin, diuretics, and class I and III antiarrhythmic agent) at the CIEDs implantation. Furthermore, data on 12-lead ECG, laboratory data (estimated glomerular filtration rate (eGFR), brain natriuretic peptide (BNP), thyroid stimulating hormone (TSH), and free thyroxine (FT4)), and transthoracic echocardiography (left ventricular ejection fraction (LVEF) and left atrial volume index (LAVI)) were also collected. Echocardiographic images were acquired in the standard parasternal and apical views. LVEF was assessed by Simpson's biplane method of disks, left atrial (LA) volume by the formula; LA volume = π / 6 (D1D2D3); where D1 was the antero-posterior LA dimension in parasternal long axis view, D2 and D3 was shortand long-axis in the apical 4 chamber view. LAVI was also calculated as LA volume / BSA [9]. Prior history of AF was defined as a documented AF on 12-lead ECG or Holter ECG monitoring, and such patients were excluded from our study cohort. Patients attended for follow-up every 3 to 6 months, at which time the device diagnostic information was interrogated and stored. All of the CIEDs were programmed to the nominal setting, which detected any episodes of arrhythmia. We defined the CIEDs-detected AF as the AHREs lasting at least 5 min with atrial rate ≥180 beats/min. AHREs with the longest duration of < 5 min were excluded from the CIEDs-detected AF given that previously published studies suggested that the 5 minute cut-off value excluded most episodes of over-sensing due to mechanical problems and appropriately detected clinical AF [5,10]. Device diagnostic information on AHREs was reviewed by at least 1 experienced electrophysiologist, blinded to clinical outcomes. We calculated the CHADS2 and CHA2DS2-VASc scores, which are well established clinical risk scores for predicting stroke and thromboembolism in patients with AF [11,12]. The HATCH score, which is a risk score for predicting the clinical progression of paroxysmal to persistent AF, was also calculated [13]. One recent study suggested that the HATCH score was useful in estimation and stratification of the development of new AF [14]. The study population was initially categorized into the two groups according to whether the CIEDs-detected AF was recorded or not. The former was defined as the ‘New-onset AF’ group, and the latter as the ‘No AF’ group. Furthermore, a subanalysis was performed to assess the relationship between risk factors for new-onset AF and duration time of AHRE. Duration time of AHRE was divided into 3 groups; 5 min ≤ AHRE < 1 h, 1 h ≤ AHRE < 24 h, and 24 h ≤ AHRE, and was compared with score of risk factors for development of new-onset AF. Continuous variables are presented as the mean ± standard

3. Results Baseline clinical characteristics of the ‘New-onset AF’ and the ‘No AF’ group are shown in Table 1. Mean age of the included patients was 61.0 ± 14.9 years old, and 243 (65.5%) were male. Overall median follow-up period was 55.0 (IQR 29.0–90.0) months. Of the 371 patients with CIED, 35.0% had pacemakers, 47.2% ICD, and 17.8% CRT. Indication for CIEDs included sick sinus syndrome (7,3%), atrioventricular block (29.1%), and VT/VF or chronic heart failure (63.6%). Seventy-eight patients (21.0%) developed new-onset CIEDs-detected AF during the follow-up period (New-onset AF group), and 293 (79.0%) had no CIEDs-detected AF (No AF group). Compared to the No AF group, the New-onset AF patients were older with more prevalent hypertension, diabetes mellitus, and heart failure. The eGFR was significantly lower and left atrial diameter was significantly higher in the New-onset AF group, compared to the No AF group. We assessed cut-off value of eGFR ≤65 ml/min/1.73 m2 and left atrial volume index > 34 ml/m2 using ROC curve analysis. The proportion of pacemaker, ICD, and CRT was not significantly different between New-onset AF group and No AF group. In addition, atrial and ventricular pacing rate was also not significantly different between two groups. Using a multivariate logistic regression analysis (Table 2), adjusting for age ≥65, hypertension, diabetes mellitus, congestive heart failure, chronic obstructive pulmonary disease, hypertrophic cardiomyopathy, eGFR ≤65 ml/min/1.73 m2, left atrial volume index > 34 ml/m2, and Class I and III antiarrhythmic agent, independent predictors for the development of new-onset AF were age ≥65 (odd ratio [OR] 2.76, 95% confidence interval [CI] 1.54–4.96, P = 0.001), diabetes mellitus (OR 2.24, 95% CI 1.20–4.19, P = 0.011), congestive heart failure (OR 1.94, 95% CI 1.06–3.54, P = 0.031), and left atrial volume index > 34 ml/m2 (OR 3.51, 95% CI 1.96–6.25, P < 0.001). Based on these 4 clinical factors (age ≥ 65, diabetes mellitus, congestive heart failure, left atrial volume index > 34 ml/m2), each component of the score was assigned 1 point, giving a score range from 0 to 4 points; mean number of clinical factors was higher in the New-onset AF group (2.2 vs. 1.3, P < 0.001). The mean CHADS2, CHA2DS2-VASc and HATCH scores were also higher in the New-onset AF group than the No AF group (Table 3). ROC curve analysis showed a good predictive ability of our 4 risk factor cluster (age ≥ 65, diabetes mellitus, congestive heart failure, left atrial volume index > 34 ml/m2) for the development of new-onset AF, with an AUC of 0.728 (95% CI; 0.680–0.773, P < 0.001), whereby ≥2 risk factors had the best predictive value with 74.4% sensitivity and 58.7% specificity. 2

European Journal of Internal Medicine xxx (xxxx) xxx–xxx

K. Miyazawa et al.

Table 1 Baseline characteristics in patients with CIEDs.

Age, years Age ≥65, n (%) Male, n (%) BSA, m2 Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Alcohol consumption, n (%) Underlying disease Hypertension, n (%) Diabetes mellitus, n (%) Congestive heart failure, n (%) Stroke or TIA, n (%) COPD, n (%) Peripheral artery disease, n (%) Sick sinus syndrome, n (%) AV block, n (%) Coronary artery disease, n (%) Hypertrophic cardiomyopathy, n (%) Dilated cardiomyopathy, n (%) Valvular heart disease, n (%) Laboratory data eGFR, ml/min/1.73 m2 (IQR) eGFR ≤65, n (%) BNP, pg/ml (IQR) TSH, μIU/ml (IQR) FT4, ng/ml (IQR) Echocardiographic data Left ventricular ejection fraction, % (IQR) Left atrial volume index, ml/m2 (IQR) Left atrial volume index > 34, n (%) Medication Beta-blocker, n (%) ACE inhibitor or ARB, n (%) Statin, n (%) Diuretic, n (%) Class I and III antiarrhythmic agent, n (%) Duration of follow-up, month (IQR) Type of CIED Pacemaker, n (%) ICD, n (%) CRT, n (%) Pacing rate Atrial pacing rate, % Ventricular pacing rate, %

Overall (n = 371)

No AF (n = 293)

New-onset AF (n = 78)

P value

61.0 ± 14.9 175 (47.2%) 243 (65.5%) 1.64 ± 0.20 116 ± 24 67 ± 12 52 (14.0%)

59.7 ± 15.2 124 (42.3%) 189 (64.5%) 1.63 ± 0.21 119 ± 23 68 ± 13 41 (14.0%)

66.1 ± 12.5 51 (65.4%) 54 (69.2%) 1.67 ± 0.18 120 ± 23 67 ± 12 11 (14.1%)

< 0.001 < 0.001 0.435 0.136 0.831 0.724 0.980

137 (36.9%) 81 (21.8%) 198 (53.4%) 32 (8.6%) 34 (9.2%) 77 (20.8%) 27 (7.3%) 108 (29.1%) 47 (12.7%) 44 (11.9%) 66 (17.8%) 25 (6.7%)

99 (33.8%) 54 (18.4%) 146 (49.8%) 25 (8.5%) 23 (7.8%) 61 (20.8%) 23 (7.8%) 80 (27.3%) 36 (12.3%) 30 (10.2%) 54 (18.4%) 17 (5.8%)

38 (48.7%) 27 (34.6%) 52 (66.7%) 7 (9.0%) 11 (14.1%) 16 (20.5%) 4 (5.1%) 28 (35.9%) 11 (14.1%) 14 (17.9%) 12 (15.4%) 8 (10.3%)

0.015 0.002 0.008 0.902 0.089 0.953 0.411 0.138 0.668 0.061 0.532 0.163

69.0 (63.0–73.0) 132 (35.6%) 67.8 (24.2–126.0) 1.72 (1.05–2.71) 1.20 (1.07–1.31)

69.0 (63.0–73.5) 97 (33.1%) 64.3 (22.4–125.5) 1.80 (1.09–2.94) 1.20 (1.07–1.31)

67.0 (60.8–72.0) 35 (44.9%) 83.0 (38.6–126.3) 1.40 (0.83–2.33) 1.16 (1.08–1.28)

0.023 0.054 0.477 0.386 0.815

59.0 (34.0–65.0) 30.0 (25.0–47.4) 108 (34.0%)

59.0 (35.0–65.0) 29.0 (23.8–39.7) 65 (26.5%)

60.0 (32.0–66.0) 46.8 (27.6–59.9) 43 (58.9%)

0.568 < 0.001 < 0.001

185 (50.4%) 200 (54.5%) 138 (37.6%) 146 (39.8%) 62 (16.9%) 55.0 (29.0–90.0)

149 (51.2%) 164 (56.4%) 106 (36.4%) 112 (38.5%) 55 (18.9%) 51.0 (26.0–86.0)

36 (47.4%) 36 (47.4%) 32 (42.1%) 34 (44.7%) 7 (9.2%) 64.0 (36.0–93.0)

0.552 0.161 0.363 0.322 0.045 0.061

130 (35.0%) 175 (47.2%) 66 (17.8%)

100 (34.1%) 145 (49.5%) 48 (16.4%)

30 (38.5%) 30 (38.5%) 18 (23.1%)

0.476 0.083 0.170

25.0 ± 31.2 47.9 ± 47.9

25.2 ± 31.5 47.2 ± 48.5

24.1 ± 29.9 50.7 ± 45.8

0.791 0.452

ACE; angiotensin converting enzyme, AF; atrial fibrillation, ARB; angiotensin II receptor blocker, AV block; atrial-ventricular block, BNP; brain natriuretic peptide, BSA; body surface area, CIED; cardiac implantable electronic device, COPD; chronic obstructive pulmonary disease, CRT; cardiac resynchronization therapy, eGFR; estimated glomerular filtration rate, FT4; free thyroxine, ICD; implantable cardioverter defibrillator, IQR; interquartile range, TIA; transient ischemic attack, TSH; thyroid stimulating hormone.

duration time of AHREs between patients with low and moderate risk, while the number of patients with high risk being significantly higher each duration time of AHREs except for the 5 min ≤ AHRE < 1 h category than that of those with low risk (Table 4).

We divided the risk score into 3 categories; low risk with score of 0 or 1, moderate risk with 2, high risk with 3 or 4. Fig. 1 showed the distribution of patients in relation to each risk of the score and each duration time of AHREs. There were no significant differences in

Table 2 Logistic regression analysis for development of new-onset AF. Multivariatea

Univariate

Age ≥65 Hypertension Diabetes mellitus Congestive heart failure COPD Hypertrophic cardiomyopathy eGFR ≤65 Left atrial volume index > 34 Class I and III antiarrhythmic agent

OR (95% CI)

P value

OR (95% CI)

P value

2.56 1.86 2.34 2.01 1.93 1.92 1.65 3.97 0.44

< 0.001 0.015 0.002 0.008 0.089 0.061 0.054 < 0.001 0.045

2.76 (1.54–4.96)

0.001

2.24 (1.20–4.19) 1.94 (1.06–3.54)

0.011 0.031

3.51 (1.96–6.25)

< 0.001

(1.52–4.31) (1.12–3.08) (1.35–4.07) (1.19–3.40) (0.90–4.15) (0.96–3.83) (0.99–2.73) (2.30–6.85) (0.19–0.99)

a Adjusted covariates included age ≥65, hypertension, diabetes mellitus, congestive heart failure, COPD, hypertrophic cardiomyopathy, eGFR ≤65 ml/min/1.73 m2, left atrial diameter ≥43 mm, left atrial volume index > 34 ml/m2, and class I and III antiarrhythmic agent.

3

European Journal of Internal Medicine xxx (xxxx) xxx–xxx

K. Miyazawa et al.

Table 3 Risk scores for predicting newly developed AF, in comparison to the 4 risk factor cluster (congestive heart failure, age ≥ 65, diabetes mellitus, left atrial volume index > 34).

CHADS2 score CHA2DS2-VASc score HATCH score 4 risk factor cluster (age ≥ 65, diabetes mellitus, congestive heart failure, left atrial volume index > 34)

Overall (n = 371)

No AF (n = 293)

New-onset AF (n = 78)

P value

1.5 2.5 1.9 1.5

1.3 2.3 1.7 1.3

1.9 3.1 2.4 2.2

< 0.001 < 0.001 < 0.001 < 0.001

± ± ± ±

1.2 1.5 1.4 1.0

± ± ± ±

1.1 1.4 1.3 0.9

± ± ± ±

1.3 1.6 1.5 1.1

Fig. 2. Receiver-operating characteristic of the risk scores for predicting the development of new-onset AF in patients with CIEDs.

Table 5 ROC curves comparison.

CHADS2 score CHA2DS2-VASc score HATCH score 4 risk factor cluster a

c-Index

95% CI

Z scorea

P valuea

0.633 0.637 0.632 0.728

0.582–0.682 0.586–0.686 0.581–0.681 0.680–0.773

3.209 3.189 3.165 Reference

0.001 0.001 0.002 Reference

For comparison of c-index.

Fig. 1. Distribution of the patients who developed AHREs in relation to each risk of the score and each duration time of AHREs.

The CHADS2, CHA2DS2-VASc, and HATCH scores also showed modest predictive ability for new-onset AF (AUC; 0.633, 95% CI; 0.582–0.682, P < 0.001, AUC; 0.637, 95% CI; 0.586–0.686, P < 0.001, AUC; 0.632, 95% CI; 0.581–0.681, P < 0.001, respectively). The predictive performance of these 3 scores (CHADS2, CHA2DS2-VASc, HATCH) was significantly lower compared to the 4 risk factor cluster (all P < 0.05) (Fig. 2, Table 5). Decision curve analysis (DCA) demonstrated that our 4 risk factor cluster had superior clinical usefulness and net benefit compared with other established risk scores to predict the development of new AF in patients with CIEDs during a wide threshold of probabilities (Fig. 3). 4. Discussion Fig. 3. Decision curve analysis for predicting the development of new-onset AF.

The main findings of the present study were that a substantial proportion (1 in 5) of patients with CIEDs experienced new-onset AF. Second, four clinical factors (age ≥ 65, diabetes mellitus, congestive heart failure, left atrial volume index > 34 ml/m2) independently predicted new-onset AF and may provide an approach to clinically Table 4 The relationship between risk for development of new-onset AF and duration time of AHRE. 4 risk factor cluster

5 min ≤ AHRE < 1 h, n (%) 1 h ≤ AHRE < 24 h, n (%) 24 h ≤ AHRE, n (%) a

Low risk

Moderate risk

n = 192

n = 117

12 (6.7%) 5 (1.7%) 3 (1.6%)

11 (9.4%) 7 (6.0%) 6 (5.1%)

P valuea

High risk

P valuea

n = 62 0.256 0.136 0.076

Low risk as a reference.

4

7 (10.0%) 10 (20.0%) 17 (50.0%)

0.190 < 0.001 < 0.001

European Journal of Internal Medicine xxx (xxxx) xxx–xxx

K. Miyazawa et al.

[34–36]. These might be explained by the multiple risk factors included in the CHADS2 and CHA2DS2-VASc scores that are associated with AFrelated pathophysiology. Compared with the other components of the CHADS2 and CHA2DS2-VASc scores, these 4 risk factors, especially left atrial enlargement such as left atrial volume index, may be representative factors for initial short and brief AF occurrence as detected by CIEDs. In addition to the individual risk factors, we tested a risk factor cluster that was associated with a good probability of new-onset AF development amongst CIED, Our data suggested that this simple cluster of 4 clinical factors (age ≥ 65, diabetes mellitus, congestive heart failure, left atrial volume index > 34 ml/m2) was useful to stratify our patients with the risk of AF development, providing the clinician with important information to guide daily clinical practice. The patients with multiple risk factors should be followed up regularly as they are more likely to develop new-onset AF. Consequently, treatment of AF could be initiated at the appropriate timing for protection from consequences of this arrhythmia. We have several limitations in the present study. First, the present study is a single-center, retrospective, and observational study with a relatively small number of CIEDs patients. Because of the single-center study design, a relatively small number of patients with coronary artery disease were included in the present study, which may have reduced the detection power and possibly influenced the validity of some interactions. Furthermore, we collected the clinical data at the time of CIEDs implantation, while the data at follow-up was not taken into account. Additionally, there is a possibility that patients already developed asymptomatic AF before CIEDs implantation because asymptomatic AF is undiagnosed by conventional method. Diagnosis and screening of asymptomatic AF is challenging and clinically important. A large multicenter prospective study of patients with CIEDs would be needed, with a view to a formal derivation and validation of a clinical scoring system for new-onset AF following CIED. While we have focused on predicting new-onset AF in patients with CIEDs, our observations may not apply to the general population.

useful risk assessment for incident AF. Previous studies reported that the incidence of CIEDs-detected AF ranges from 15% to 60% [6–8,17–23]. In the present study 21% of the patients with CIEDs experienced new-onset AF, but the incidence of CIEDs-detected AF would strongly depend on the study population. Therefore, the patients with a prior history of AF were excluded given that patients with a history of AF are more likely to develop new-onset AF. Furthermore, to detect AF accurately we excluded the patients with single chamber CIEDs because an atrial lead provides the accurate detection and characterization of individual episodes of rapid atrial rate over long periods [5]. For the same reason, patients with pacing mode of VVI or AAI were excluded even with dual-chamber CIEDs. The present study included a broad range of CIED patients with pacemakers as well as ICDs and CRTs, and demonstrated that the proportion of type of CIEDs was not associated with development of new-onset AF. Heart failure was one of the independent predictors of new-onset AF, while there was no significant difference in patients between with and without CRT. This result indicates that the remaining 3 risk factors have more impact on development of new-onset AF. Furthermore, previous study has reported that right ventricular pacing is associated with a variety of detrimental outcomes including heart failure and AF [24], while the present study showed that pacing rate was not significantly different in development of new-onset AF. It is probably because recent CIEDs have specific function avoiding unnecessary right ventricular pacing, leading to risk reduction of development of new-onset AF [25]. To the best of our knowledge, no previous study has described the relationship between new-onset AF and CIEDs' characteristics. The incidence of new onset AF in general population is lower than in patients with CIEDs [26,27]. Patients with CIEDs often have underlying heart disease that may contribute to the substrate of AF occurrence. Moreover, CIEDs can detect brief episodes of AF regardless of the presence of symptoms, resulting in detection of AF that would be underdiagnosed by conventional diagnostic methods. Recently, new diagnostic methods such as handheld devices have been developed and introduced for the screening of AF, and may be useful for assessment of the accurate incidence of newly developed AF in general population [28–31]. Although these new devices can provide simple and accurate detection of AF, there is still some room for improvements in feasibility, acceptability, and cost-effectiveness. In the present study, age, diabetes mellitus, heart failure, and left atrial volume index were the independent predictors of AF development. Several cohort studies including general population have also demonstrated that these factors are associated with the development of new-onset AF [32,33], while the results were not necessarily consistent amongst all studies due to differences in underlying heart disease and diagnostic methods of AF. Previous cohort studies have demonstrated that obesity is one of the risk factors for incident AF [1]. The present study also showed that patients with new-onset AF tended to have a high BSA compared to those with no AF, although not statistically significant. As CIEDs-detected AF episodes are usually short and brief that cannot be easily detected by conventional diagnostic methods, these 4 risk factors may have an impact on development of initial brief episodes of AF compared with other clinical factors. Our data suggest that these simple clinical factors are associated with a substrate of newonset AF occurrence in patients with CIEDs. There may be a possibility that patients with these factors had already developed brief episodes of AF before CIEDs implantation, which had not been detected by conventional diagnostic methods such as 12-lead ECG and Holter ECG. Therefore, these factors can be the risk factors of undiagnosed AF by conventional methods in addition to predictors of new-onset AF in patients with CIEDs. The cluster of 4 risk factors largely overlaps with components of the CHADS2 and CHA2DS2-VASc scores. In recent years, use of the CHADS2 and CHA2DS2-VASc scores has extended beyond the prediction of thromboembolism to the prediction of left atrial remodeling, AF recurrence after catheter ablation, and development of new-onset AF

5. Conclusion In conclusion, a substantial number of patients (1 in 5) with CIEDs develop new AF. Four clinical factors (age ≥ 65, diabetes mellitus, congestive heart failure, left atrial volume index > 34 ml/m2) independently predicted new-onset AF and may provide an approach to clinically useful risk assessment for incident AF. Conflict of interest No author has a specific conflict of interest in the publication of this study. Acknowledgements There are no acknowledgments. References [1] Vermond RA, Geelhoed B, Verweij N, Tieleman RG, Van der Harst P, Hillege HL, et al. Incidence of atrial fibrillation and relationship with cardiovascular events, heart failure, and mortality: a community-based study from the Netherlands. J Am Coll Cardiol 2015;66:1000–7. [2] Savelieva I, Camm AJ. Clinical relevance of silent atrial fibrillation: prevalence, prognosis, quality of life, and management. J Interv Card Electrophysiol 2000;4:369–82. [3] Gladstone DJ, Spring M, Dorian P, Panzov V, Thorpe KE, Hall J, et al. Atrial fibrillation in patients with cryptogenic stroke. N Engl J Med 2014;370:2467–77. [4] Lin HJ, Wolf PA, Benjamin EJ, Belanger AJ, D'Agostino RB. Newly diagnosed atrial fibrillation and acute stroke. The Framingham Study. Stroke 1995;26:1527–30. [5] Pollak WM, Simmons JD, Interian Jr. A, Atapattu SA, Castellanos A, Myerburg RJ, et al. Clinical utility of intraatrial pacemaker stored electrograms to diagnose atrial fibrillation and flutter. Pacing Clin Electrophysiol 2001;24:424–9.

5

European Journal of Internal Medicine xxx (xxxx) xxx–xxx

K. Miyazawa et al.

[6] Glotzer TV, Hellkamp AS, Zimmerman J, Sweeney MO, Yee R, Marinchak R, et al. Atrial high rate episodes detected by pacemaker diagnostics predict death and stroke: report of the Atrial Diagnostics Ancillary Study of the MOde Selection Trial (MOST). Circulation 2003;107:1614–9. [7] Glotzer TV, Daoud EG, Wyse DG, Singer DE, Ezekowitz MD, Hilker C, et al. The relationship between daily atrial tachyarrhythmia burden from implantable device diagnostics and stroke risk: the TRENDS study. Circ Arrhythm Electrophysiol 2009;2:474–80. [8] Healey JS, Connolly SJ, Gold MR, Israel CW, Van Gelder IC, Capucci A, et al. Subclinical atrial fibrillation and the risk of stroke. N Engl J Med 2012;366:120–9. [9] Pritchett AM, Jacobsen SJ, Mahoney DW, Rodeheffer RJ, Bailey KR, Redfield MM. Left atrial volume as an index of left atrial size: a population-based study. J Am Coll Cardiol 2003;41:1036–43. [10] Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B, et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J 2016;37:2893–962. [11] Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001;285:2864–70. [12] Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest 2010;137:263–72. [13] de Vos CB, Pisters R, Nieuwlaat R, Prins MH, Tieleman RG, Coelen RJ, et al. Progression from paroxysmal to persistent atrial fibrillation clinical correlates and prognosis. J Am Coll Cardiol 2010;55:725–31. [14] Suenari K, Chao TF, Liu CJ, Kihara Y, Chen TJ, Chen SA. Usefulness of HATCH score in the prediction of new-onset atrial fibrillation for Asians. Medicine (Baltimore) 2017;96:e5597. [15] DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837–45. [16] Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making 2006;26:565–74. [17] Botto GL, Padeletti L, Santini M, Capucci A, Gulizia M, Zolezzi F, et al. Presence and duration of atrial fibrillation detected by continuous monitoring: crucial implications for the risk of thromboembolic events. J Cardiovasc Electrophysiol 2009;20:241–8. [18] Capucci A, Santini M, Padeletti L, Gulizia M, Botto G, Boriani G, et al. Monitored atrial fibrillation duration predicts arterial embolic events in patients suffering from bradycardia and atrial fibrillation implanted with antitachycardia pacemakers. J Am Coll Cardiol 2005;46:1913–20. [19] Gillis AM, Morck M. Atrial fibrillation after DDDR pacemaker implantation. J Cardiovasc Electrophysiol 2002;13:542–7. [20] Healey JS, Martin JL, Duncan A, Connolly SJ, Ha AH, Morillo CA, et al. Pacemakerdetected atrial fibrillation in patients with pacemakers: prevalence, predictors, and current use of oral anticoagulation. Can J Cardiol 2013;29:224–8. [21] Swiryn S, Orlov MV, Benditt DG, DiMarco JP, Lloyd-Jones DM, Karst E, et al. Clinical implications of brief device-detected atrial tachyarrhythmias in a cardiac

[22]

[23]

[24]

[25]

[26] [27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

6

rhythm management device population: results from the registry of atrial tachycardia and atrial fibrillation episodes. Circulation 2016;134:1130–40. Kim BS, Chun KJ, Hwang JK, Park SJ, Park KM, Kim JS, et al. Predictors and longterm clinical outcomes of newly developed atrial fibrillation in patients with cardiac implantable electronic devices. Medicine (Baltimore) 2016;95:e4181. Ogino Y, Ishikawa T, Ishigami T, Matsumoto K, Hosoda J, Iguchi K, et al. Characteristics and prognosis of pacemaker-identified new-onset atrial fibrillation in Japanese people. Circ J 2017;81:794–8. Sweeney MO, Hellkamp AS, Ellenbogen KA, Greenspon AJ, Freedman RA, Lee KL, et al. Adverse effect of ventricular pacing on heart failure and atrial fibrillation among patients with normal baseline QRS duration in a clinical trial of pacemaker therapy for sinus node dysfunction. Circulation 2003;107:2932–7. Quesada A, Botto G, Erdogan A, Kozak M, Lercher P, Nielsen JC, et al. Managed ventricular pacing vs. conventional dual-chamber pacing for elective replacements: the PreFER MVP study: clinical background, rationale, and design. Europace 2008;10:321–6. Kannel WB, Abbott RD, Savage DD, McNamara PM. Coronary heart disease and atrial fibrillation: the Framingham Study. Am Heart J 1983;106:389–96. Psaty BM, Manolio TA, Kuller LH, Kronmal RA, Cushman M, Fried LP, et al. Incidence of and risk factors for atrial fibrillation in older adults. Circulation 1997;96:2455–61. Samol A, Masin M, Gellner R, Otte B, Pavenstadt HJ, Ringelstein EB, et al. Prevalence of unknown atrial fibrillation in patients with risk factors. Europace 2013;15:657–62. Wiesel J, Arbesfeld B, Schechter D. Comparison of the microlife blood pressure monitor with the Omron blood pressure monitor for detecting atrial fibrillation. Am J Cardiol 2014;114:1046–8. Kearley K, Selwood M, Van den Bruel A, Thompson M, Mant D, Hobbs FR, et al. Triage tests for identifying atrial fibrillation in primary care: a diagnostic accuracy study comparing single-lead ECG and modified BP monitors. BMJ Open 2014;4:e004565. McManus DD, Lee J, Maitas O, Esa N, Pidikiti R, Carlucci A, et al. A novel application for the detection of an irregular pulse using an iPhone 4S in patients with atrial fibrillation. Heart Rhythm 2013;10:315–9. Benjamin EJ, Levy D, Vaziri SM, D'Agostino RB, Belanger AJ, Wolf PA. Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA 1994;271:840–4. Qureshi W, Soliman EZ, Solomon SD, Alonso A, Arking DE, Shah A, et al. Risk factors for atrial fibrillation in patients with normal versus dilated left atrium (from the Atherosclerosis Risk in Communities Study). Am J Cardiol 2014;114:1368–72. Li Y, Ding W, Wang H, Song N, Lin L, Wang Z, et al. Relationship of CHA2DS2-VASc and CHADS2 score to left atrial remodeling detected by velocity vector imaging in patients with atrial fibrillation. PLoS One 2013;8:e77653. Chao TF, Lin YJ, Tsao HM, Tsai CF, Lin WS, Chang SL, et al. CHADS(2) and CHA(2) DS(2)-VASc scores in the prediction of clinical outcomes in patients with atrial fibrillation after catheter ablation. J Am Coll Cardiol 2011;58:2380–5. Saliba W, Gronich N, Barnett-Griness O, Rennert G. Usefulness of CHADS2 and CHA2DS2-VASc scores in the prediction of new-onset atrial fibrillation: a population-based study. Am J Med 2016;129:843–9.