2017 Volume 81 Issue 6 Pages 794-798
Background: The characteristics and prognosis of implanted pacemaker-identified new-onset atrial fibrillation (AF) in Japanese people has not been well evaluated.
Methods and Results: A total of 395 consecutive patients with newly implanted pacemakers were retrospectively analyzed between January 2010 and December 2015 at Yokohama City University Hospital. Patients with a prior history of AF, VVI mode pacemaker, congenital heart disease, severe valvular heart disease, and cardiovascular surgery before pacemaker implantation were excluded. Among the remaining patients, 44 (21.3%) developed new AF during follow-up (mean follow-up, 1,115±651 days; range, 9–2,176 days). Patients with new-onset AF had a significantly higher CHADS2 score (2.09±1.27 vs. 1.31±1.08, P<0.001) and CHA2DS2-VASc score (3.00±1.39 vs. 2.26±1.19, P<0.001) compared with those without new-onset AF. On Cox regression analysis only age at implantation was significantly correlated with new-onset AF. Interestingly, the incidence of hospitalization due to heart failure was significantly higher in the new-onset AF than in the without new-onset AF group.
Conclusions: A total of 21.3% of pacemaker-implanted patients with high CHADS2 and CHA2DS2-VASc scores developed new-onset AF during a mean follow-up of 3.1 years; and pacemaker-identified AF was associated with an increased risk of worsening heart failure.
Atrial fibrillation (AF) frequently occurs in the general population and is responsible for considerable morbidity and mortality related to ischemic stroke. Approximately 0.6% of the Japanese population has AF, and its prevalence will continue to increase with the aging of the population.1 According to the Japan Multicenter Stroke Investigators’ Collaboration, AF identified on regular electrocardiogram (ECG) screening is a predictive factor for severe stroke and early death in acute ischemic stroke.2
Early diagnosis of AF has clinical importance for preventing thrombotic stroke, but detection of AF at the early phase is difficult, especially in asymptomatic AF. In a previous study, 36.8% of patients with ischemic stroke or transient ischemic attack (TIA) with AF were not prescribed anticoagulant drug prior to the diagnosis of stroke or TIA, mainly because of a lack of identification of AF.3 ECG is a limited method of detecting AF because of the limited duration of monitoring.
Cardiac implanted electronic devices (CIED), including pacemakers, implantable cardiac defibrillators, and cardiac resynchronization therapy, can automatically record and store atrial high-rate episodes (AHRE) according to a programmable detection algorithm. Frequency, duration, time, date, and interatrial interval during such episodes can be validated using CIED with appropriate algorithms. Therefore, new-onset AF can be diagnosed earlier by analyzing stored CIED information than using regular ECG monitoring.
Glotzer et al showed that patients with pacemaker-identified AF (AHRE >5 min) with sick sinus syndrome (SSS) are more than twice as likely to die or have a stroke compared with patients without AF.4 The correlation, however, between pacemaker-identified AF and prognosis in Japanese people is not well known. The aim of this retrospective study was therefore to determine the incidence rate, characteristics, and prognosis of pacemaker-identified new-onset AF in Japanese people.
This study was a single-center, retrospective, observational study that assessed the characteristics and prognosis of new-onset AF in patients with an implanted pacemaker.
A total of 395 consecutive patients with a newly implanted pacemaker were analyzed between January 2010 and December 2015 at Yokohama City University Hospital. Patients with a prior history of AF, VVI mode pacemaker, congenital heart disease, severe valvular heart disease, and prior history of cardiovascular surgery before pacemaker implantation were excluded. A total of 150 patients, including 100 with a prior history of AF, 34 with a VVI mode pacemaker, 11 with congenital heart disease, and 5 with a prior history of cardiovascular surgery, were excluded (no patients had severe valvular heart disease). Thirty-eight patients dropped out from the study during follow-up, leaving a total of 207 patients for analysis. The indication for pacemaker included SSS (n=76), atrioventricular block (AVB; n=128), and neurally mediated syncope (n=3). Selection of pacemaker manufacturer and atrial sensitivity setting were at the discretion of the implanting physician. Mean atrial sensitivity setting was 0.481±0.169 mV. Fifty-nine patients had a Medtronic (Minneapolis, MN, USA), 66 had a St. Jude Medical (Saint Paul, MN, USA), 36 had a Boston Scientific (Marlborough, MA, USA), 14 had a Biotronik (Berlin, Germany), and 32 had an Ela Medical (Montrouge, France) pacemaker. All of the patients provided written informed consent.
All of the patients underwent implantation of dual chamber pacemaker. The pacemakers had bipolar atrial and ventricular leads, and were programmed to a DDI (R) or DDD (R). A total of 198 atrial leads were placed in the right atrial appendage and the remaining 9 atrial leads were placed in the atrial septum. All ventricular leads were placed in the right ventricular apex. Detection of atrial tachycardia was programmed “on” and the detection rate for AHRE ECG storage was recorded for each patient. Patients were followed up in the outpatient unit every 6 months. In each of the visits, standard physical status was recorded with blood pressure measurements, as well as 12-lead ECG recording (ECG was recorded at standard 25-mm/s speed) and pacemaker interrogation. At each follow-up visit for device interrogation, summary information on AHRE and specific episodes was collected. AHRE summary information included the date, frequency, and duration of each episode. In this study, AF was defined as AHRE >5 min with a cut-off rate of 180 beats/min. Five-min was chosen based on a previous report, which showed that implantable devices can detect AF appropriately, particularly when the cut-off point for duration of AHRE is >5 min.5 Far-field oversensing and sinus tachycardia are likely to be mistaken for AF. Such inadequate data were excluded by checking intracardiac ECG of AHRE that were stored in the pacemakers.
Patients with AHRE >5 min during follow-up were defined as new AF(+), and those with no AHRE or AHRE <5 min as new AF(−). After the mean follow-up period of the new AF patients was determined, the new AF(+) and new AF(−) groups were compared. At the time of implantation, each patient’s chart was reviewed to determine CHADS2 score and CHA2DS2-VASc score,6,7 prior history of AF before pacemaker implantation, dyslipidemia, smoking habit, and use of renin-angiotensin system inhibitor (RASI), β-blocker, diuretic, and calcium channel blocker. Second, we evaluated the incidence of all-cause death, stroke, and hospitalization for heart failure (HF) in the new AF(+) and new AF(−) groups. The diagnosis for worsening HF and decision for hospitalization was made by the attending physician on physical examination, blood test, chest X-ray, and echocardiogram.
Statistical AnalysisAll continuous variables are reported as mean±SD and all categorical variables are represented as frequency (percentage). Unpaired t-test was used to compare continuous variables and chi-squared test was used to test the difference in qualitative variables between groups. Cox regression analysis was used to determine the predictive factor at enrollment for new-onset AF. Kaplan-Meier analysis and log-rank test were used to estimate survival curves and stroke and HF hospitalization event-free curves. For all comparisons, P<0.05 was considered statistically significant. SPSS Statistic ver. 21 (IBM, New York, NY, USA) was used for analysis.
During a mean follow-up of 1,115±651 days (range, 9–2,176 days), 44 patients (21.3%) developed pacemaker-identified new-onset AF at a mean of 570±578 days (range, 7–1,986 days) after pacemaker implantation. Table 1 lists the subject baseline characteristics. The new AF(+) group had higher CHADS2 score (2.09±1.27 vs. 1.31±1.08, P<0.001) and CHA2DS2-VASc score (3.00±1.39 vs. 2.26±1.19, P<0.001) compared with the new AF(−) group. With regard to CHADS2 and CHA2DS2-VASc scores, the prevalence of age ≥75 years, diabetes, and prior history of stroke in the new AF(+) group was significantly higher than in the new AF(−) group, but only age at enrollment was a significant predictor of new-onset AF on Cox regression analysis (OR, 1.04; P=0.021; Table 2).
| All (n=207) |
New AF(+) (n=44) |
New AF(−) (n=163) |
P value | |
|---|---|---|---|---|
| Male | 108 (52.2) | 27 (61.4) | 81 (49.7) | 0.148 |
| Age (years) | 70.9±14.4 | 77.6±9.66 | 69.0±15.0 | <0.001 |
| SSS | 76 (36.7) | 18 (40.9) | 58 (35.6) | 0.571 |
| AVB | 128 (61.8) | 26 (59.1) | 102 (62.6) | 0.707 |
| Atrial sensitivity (mV) | 0.481±0.169 | 0.466±0.172 | 0.483±0.169 | 0.546 |
| Hypertension | 117 (56.5) | 30 (68.2) | 87 (53.4) | 0.079 |
| Age ≥75 years | 99 (47.8) | 29 (65.9) | 70 (42.9) | <0.01 |
| Diabetes | 39 (18.8) | 14 (31.8) | 25 (15.3) | 0.013 |
| Prior history of stroke | 13 (6.28) | 6 (13.6) | 7 (4.29) | 0.023 |
| Vascular disease | 38 (18.4) | 12 (27.3) | 26 (16.0) | 0.085 |
| 64<Age<75 years | 57 (27.5) | 11 (25.0) | 46 (28.2) | 0.671 |
| CHADS2 score | 1.48±1.16 | 2.09±1.27 | 1.31±1.08 | <0.001 |
| CHA2DS2-VASc score | 2.43±1.26 | 3.00±1.39 | 2.26±1.19 | <0.001 |
| Hyperlipidemia | 49 (23.7) | 12 (27.3) | 37 (22.7) | 0.527 |
| Smoking | 86 (41.5) | 25 (56.8) | 61 (37.4) | 0.019 |
| Echocardiographic parameters | ||||
| LVEF (%) | 69.5±9.05 | 67.7±7.65 | 70.0±9.34 | 0.188 |
| LVDd (mm) | 4.62±0.58 | 4.72±0.57 | 4.59±0.59 | 0.261 |
| LADs (cm) | 3.73±0.67 | 3.87±0.83 | 3.69±0.62 | 0.151 |
| Laboratory findings | ||||
| BNP (pg/mL) | 139±225 | 179±255 | 129±216 | 0.205 |
| Creatinine (mg/dL) | 1.08±1.36 | 1.12±0.71 | 1.07±1.47 | 0.821 |
| eGFR (mL/min/1.73 m2) | 68.3±52.6 | 57.3±20.1 | 71.3±57.9 | 0.122 |
| CRP (mg/dL) | 0.44±0.94 | 0.55±1.01 | 0.41±0.91 | 0.383 |
| Medication | ||||
| β-blocker | 31 (15.0) | 10 (22.7) | 21 (12.9) | 0.108 |
| RASI | 69 (33.3) | 20 (45.5) | 49 (30.1) | 0.058 |
| CCB | 68 (32.9) | 20 (45.5) | 48 (29.4) | 0.042 |
| Diuretics | 31 (15.0) | 12 (27.3) | 19 (11.7) | <0.01 |
Data given as n (%) or mean±SD. AF, atrial fibrillation; AVB, atrioventricular block; BNP, brain natriuretic peptide; CCB, calcium channel blocker; CHF, congestive heart failure; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; LAD, left atrial dimension; LVDd, left ventricular diastolic dimension; LVEF, left ventricular ejection fraction; RASI, renin-angiotensin system inhibitor; SSS, sick sinus syndrome.
| OR | P value | 95% CI | |
|---|---|---|---|
| Female sex | 1.16 | 0.740 | 0.476−2.85 |
| Age | 1.04 | 0.021 | 1.007−1.084 |
| SSS | 1.49 | 0.309 | 0.691−3.22 |
| Heart failure | 1.37 | 0.525 | 0.520−3.61 |
| Hypertension | 0.539 | 0.185 | 0.216−1.34 |
| Diabetes | 2.13 | 0.071 | 0.938−4.85 |
| Prior history of stroke | 1.89 | 0.288 | 0.584−6.14 |
| Vascular disease | 1.15 | 0.748 | 0.496−2.65 |
| Hyperlipidemia | 1.05 | 0.925 | 0.423−2.58 |
| Smoking | 1.78 | 0.197 | 0.742−4.24 |
| BNP | 1.001 | 0.194 | 0.999−1.003 |
| EF | 0.979 | 0.192 | 0.941−1.01 |
| LAD | 0.934 | 0.818 | 0.523−1.67 |
| eGFR | 0.998 | 0.564 | 0.991−1.005 |
| CRP | 0.925 | 0.766 | 0.552−1.55 |
| β-blocker | 0.877 | 0.794 | 0.329−2.34 |
| RASI | 1.29 | 0.571 | 0.535−3.11 |
EF, ejection fraction. Other abbreviations as in Table 1.
The number of patients with CHADS2 score ≥1 in the new AF(−) and new AF(+) groups was 116 (71.2%) and 40 (90.9%), respectively (Figure 1). In the new AF(+) group, 1 patient was prescribed an anticoagulant before pacemaker implantation because of deep vein thrombosis, and 13 were prescribed anticoagulants after pacemaker implantation due to identification of AF.
Distribution of (A,B) CHADS2 and (C,D) CHA2DS2-VASc scores vs. presence of new-onset atrial fibrillation (new AF: with, (+); without, (–)).
In the new AF(+) group, during follow-up, 7 patients died, 3 developed stroke, and 6 needed to be admitted to hospital because of worsening HF. On Kaplan-Meier method and log-rank test, the incidence of admission due to worsening HF in the new AF(+) group was significantly higher than in the new AF(−) group (HR, 8.87; 95% CI: 2.22–35.5; P=0.002). In contrast, the incidence of all-cause death and stroke in the new AF(+) group was not significantly different to the new AF(−) group (Figure 2). The HR for stroke was not available because none of the patients developed stroke during follow-up in the new AF(−) group.
Kaplan-Meier estimates for (A) admission for heart failure (HF), (B) all-cause death, and (C) stroke in patients with (+) and without (–) atrial fibrillation (new AF). HR, hazard ratio; NA, not available.
A total of 3 patients who developed stroke during follow-up in new AF(+) group, were not prescribed anticoagulant drugs before development of stroke. Two of these 3 patients had low CHADS2 score (1 or 2 points). Forty-one patients in the new AF(+) group did not develop stroke, 14 of whom were prescribed anticoagulant drugs. In the remaining 27 patients, 16 patients had low CHADS2 score. Comparing these 2 groups of low CHADS2 score (n=2 vs. n=16), median longest duration of AF was significantly different between the 2 groups (224 h vs. 8.2 h, P<0.001).
In addition, the onset patterns of AF were evaluated in this study. A total of 123 intracardiac ECG describing only the onset of AF were available in the new AF(+) group. In these 123 AF episodes, 63 AF episodes (51.2%) were followed by premature atrial complexes (PAC).
In the present study, we characterized the incidence and predictors of AF in patients with an implanted pacemaker under various cardiovascular indications and validated the prognosis. The principal findings are as follows: (1) AF, which was defined as AHRE >5 min, was detected during follow-up in 21.3% of patients with no history of AF at pacemaker implantation; (2) patients with new-onset AF had higher CHADS2 and CHA2DS2-VASc scores than those without new-onset AF because of a higher prevalence of age ≥75 years, diabetes, and a prior history of stroke; (3) age was a significant predictive factor for new-onset AF; and (4) AF identified by pacemaker at Yokohama City University Hospital was associated with an increased risk of worsening HF.
In Europe and the USA showed approximately 20–50% of patients with an implanted pacemaker develop new-onset AF.8–10 In the present study the incidence of AF in Japanese people with implanted pacemakers was also within in this range. The incidence of new AF in the general population was 0.3–0.5% in the Framingham Heart study11 and 0.3% in the Hisayama study in Japan.12 The difference in incidence of AF between these studies is probably due to the method of diagnosing AF, despite the fact that the background of these studies was different. AF is generally diagnosed on ECG or Holter monitoring, but these methods are not useful in early detection of AF because of the short duration of monitoring. CIED, including pacemakers, can continually check atrial arrhythmia and help in the early diagnosis of AF.
In the present study a high percentage (90.9%) of patients with pacemaker-identified new-onset AF had indications for anticoagulants, according to Japanese Circulation Society guidelines.12 As noted here, AHRE >5 min is associated with all-cause death and stroke.4 Another study showed that atrial tachyarrhythmia >6 min in patients with implanted pacemaker or implantable cardiac defibrillator was also associated with ischemic stroke or systemic embolism.13 The correlation between the appropriate time for starting anticoagulant therapy and the duration of AF among Japanese patients, however, especially with low CHADS2 score, is not known. In the present study, in patients with low CHADS2 score in the new AF(+) group, median longest duration of AF was significantly different between patients with and without development of stroke (224 h vs. 8.2 h, P<0.001). The number of subjects was too small to enable statistical analysis, but increase of longest AF duration may correlate with increase of stroke, even in patients with low CHADS2 score. Use of a home monitoring system may be helpful for detecting increase of longest AF duration and prevention of stroke.14,15
We have shown that pacemaker-identified new-onset AF is significantly associated with worsening HF at Yokohama City University Hospital. AF, with its loss of atrioventricular synchrony and irregularity of the ventricular cycle, results in a reduction in cardiac output. In particular, new-onset AF can cause an acute change in hemodynamics, resulting in a worse prognosis.16 Yamauchi et al reported that new AF diagnosed on ECG was significantly associated with increased incidence of worsening HF in Japanese patients with chronic HF.17 We could reveal that pacemaker-identified AF is also associated with worsening HF.
In the present study aging was correlated with new-onset pacemaker-detected AF on Cox regression analysis. Common risk factors, such as aging, hypertension, diabetes, smoking, obesity, renal failure, and sleep apnea syndrome, are associated with AF.18 Among these factors, advanced aging is one of the most important predictive factors of AF.19,20 In particular, in a society with progressive aging, such as Japan, CIED, including pacemakers, play a crucial role in detecting the early stage of AF.
In a previous study of the onset patterns of AF, PAC was the most prevalent AF onset scenario.21 And in the present study, according to the intracardiac ECG for onset of AF, PAC was also the major onset scenario for AF in Japanese patients.
Pacemaker manufacturers were not identical because this was a retrospective study. The algorithms vary with manufacturer and in complexity, but a difference between manufacturers is likely to occur if the AHRE is late and the duration is short. In the present study, the AHRE criteria used to define AF were heart rate >180 beats/min and duration >5 min. Furthermore, the sensing configuration of atrial leads was not significantly different between the new AF(+) and new AF(−) groups. Therefore, varying manufacturer algorithm for detecting AF should not have affected AF detection in this study.
There were several limitations in this study. First, we excluded patients with prior history of AF before pacemaker implantation, but subclinical AF is difficult to identify. For this reason, there is a possibility that patients in the new AF(+) group had already developed AF before pacemaker implantation, but this limitation may occur in every study on new-onset AF. Second, given that this was a single-center study, the number of analyzed patients was relatively small. Third, medication dose and patient compliance were not analyzed. For more precise validation, a prospective study with a larger number of patients is required.
A total of 21.3% of pacemaker-implanted patients with high CHADS2 and CHA2DS2-VASc scores developed new-onset AF during a mean follow-up of 3.1 years. Patients with pacemaker-identified AF at Yokohama City University Hospital were associated with an increased risk of worsening HF.
We sincerely thank the members of the Departments of Medical Science and Cardiorenal Medicine of Yokohama City University Hospital.
None.
