Asymptomatic Atrial Fibrillation in Pacemaker Recipients: Incidence, Progression, and Determinants Based on the Atrial High Rate Trial

Michael V. Orlov, M.D, Ph.D.; Jalal K. Ghali, M.D.; Mohsen Araghi-Niknam, Ph.D.; Lou Sherfesee, Ph.D.; Diane Sahr, R.N., B.S.N.; Douglas A. Hettrick, Ph.D.; for the Atrial High Rate Trial Investigators

Pacing Clin Electrophysiol. 2007;30(3):404-411. ©2007 Blackwell Publishing

Abstract

Background: The epidemiology and clinical implications of asymptomatic atrial tachyarrhythmias (AT) including both atrial fibrillation and flutter in pacemaker recipients with and without arrhythmia history are not well understood. The Atrial High Rate Episodes (A-HIRATE) in Pacemaker Patients Trial was designed to identify and compare the incidence of AT in patients with and without previously diagnosed AT and a standard indication for dual chamber pacing, and to provide useful diagnostic information for clinical management.
Methods: Four hundred twenty-seven patients were implanted with a pacemaker (Kappa 7-900, Medtronic, Inc., Minneapolis, MN, USA) capable of detecting and storing multiple atrial high rate episodes (AHRE) and followed for 2 years. Group I included 331 patients without prior history of AT and Group II included 96 patients with prior AT history.
Results: Pacemaker diagnostics appropriately detected 93% of reviewed AHRE. The rate of occurrence of first AHRE was significantly higher (P < 0.0001) in Group II patients, as was average AHRE burden. The rate of first AHRE occurrence was 88.6% for patients in Group II and 53.8% in Group I at 24 months post-implant. The rate of AHRE occurrence was similar in both groups after the first month post-implant. The majority of stored AHRE were asymptomatic; symptoms did not correspond to an actual AHRE in most patients.
Conclusions: The incidence of AT in pacemaker recipients is high. Most device-detected AHRE are asymptomatic. Prior history of AT is associated with higher arrhythmia burden. AHRE diagnostics have a high positive predictive value for identifying AT events.
Introduction

Atrial tachyarrhythmias (AT), including atrial fibrillation (AF) and flutter, occur frequently in the general population and are responsible for considerable morbidity and mortality, primarily related to stroke.[1-3] According to US Medicare data, the risk of one-year stroke occurrence more than doubled in patients with AT compared to those without AT.[3] AT also accounts for the majority of hospital admissions with arrhythmia as the primary diagnosis.[3,4] However, these reports are based primarily on symptomatic AF and thus allow correlation with subsequent complications. In contrast, the importance of more difficult to detect asymptomatic atrial arrhythmias was underestimated until recently[5,6] and its relevance to morbidity and mortality is less well known. Recently, the introduction of pacemakers with capacity to detect, quantify, and store asymptomatic AT episodes has resulted in new insights into minimally symptomatic AT.[7-10] The epidemiology and clinical implications of this arrhythmia have been studied in pacemaker recipients with prior history of AT.[10] However, despite these recent advances, the importance of asymptomatic AT in clinical practice remains less clear than that of symptomatic AT.

The purpose of the Atrial High Rate Episodes (A-HIRATE) in Pacemaker Patients Trial was to identify the incidence of AT in patients with previously undiagnosed AT and a standard indication for dual chamber pacing, and to provide useful diagnostic information for successful clinical management.
Methods

The A-HIRATE Trial was a multicenter, prospective study assessing the incidence of atrial high rate episodes (AHRE) in pacemaker patients with and without previously diagnosed AT. The trial was conducted in 44 centers in the US, Australia, and France. Each institutional review board approved the study protocol and each participant provided written informed consent. Patients were followed for two years including scheduled follow-up visits at 1, 6, 12, 18, and 24 months. The first patient was enrolled in July of 2000 and the trial was completed in August of 2004. The study consisted of two arms: Group I included patients without previously diagnosed AT and Group II consisted of patients with a prior history of AT as indicated by the enrolling center. Patients in both study arms were required to have an approved indication for pacemaker implantation.[11] The study aimed to determine the incidence of asymptomatic and symptomatic AT, arrhythmia onset characteristics, cumulative AT burden (i.e. cumulative AT episode duration), and to examine the accuracy of AHRE diagnostics. Dual chamber multiprogrammable pacemakers, (Kappa 700 or Kappa 900, Medtronic, Inc., Minneapolis, MN, USA) were implanted in all trial participants. No specific treatments or therapies were required.

Patients were required to program the Atrial High Rate Diagnostic and Clinician Selected Atrial High Rate Detail "ON." The Atrial High Rate Diagnostic records both AHRE summary information and specific episode information. The summary information includes AHRE counts, as well as the percentage of time during the collection period that an AHRE was detected (i.e. "burden"). Specific AHRE information in the Atrial High Rate Diagnostic includes the date, time, and duration of each episode for up to 16 episodes as follows: the Kappa 700 recorded data for the most recent 14 episodes plus the first, fastest, and longest episodes, while the Kappa 900 device recorded data from the most recent 13 episodes plus the first, fastest, and longest episodes. The Atrial High Rate Detail collected device marker channel and beat-to-beat data for up to eight atrial episodes, as well as daily activity recorded in the Atrial Arrhythmia Trend including data on pre-onset events, post-onset events, and pre-termination events. Kappa 900 devices also recorded the cumulative daily AHRE burden (hours/day) for up to 175 consecutive days.

Mode Switch was programmed ON with a 60-second delay in order to avoid detection of non-sustained AT episodes. Decreases in the recommended programmed mode switch delay were allowed in the event of symptomatic AT. An AHRE was defined via device programming as an atrial rate >180 bpm. The minimal episode duration for AHRE detection was programmed to 60 seconds since shorter episodes have questionable clinical significance and are also more likely to be inappropriately detected episodes.[12] Persistent AT was defined as at least seven consecutive days with device classified AHRE burden greater than 22 hours/day. Device programming parameters were recorded at the time of implant and at each follow-up visit. Stored device data including electrograms and the Atrial High Rate Diagnostic data were collected at each follow-up visit. AHRE burden was determined by the device as the cumulative duration of all AHRE divided by the follow-up period. Onset of stored AHRE was classified as bradycardia if the atrial rate was less than 80 bpm and classified as tachycardia if the rate was greater than 80 bpm. A subset of sequentially received device-stored episodes including atrial electrogram and marker channel were reviewed according to prospectively determined criteria in order to validate the accuracy of AHRE detection as previously described.[13]

Patients were issued symptom diaries to record AT symptoms including: "heart fluttering," "skipping," "racing," " light-headedness," "dizziness," "fainting," "chest pain/pressure or fullness," "shortness of breath," and "fatigue," respectively. A symptomatic event was defined as a device-detected AHRE occurring on the same day as patient-reported symptoms. The number of symptomatic episodes per patient-month was determined for each patient by dividing the number of symptomatic episodes for the patient by the total device-measured follow-up period.
Statistical Methods

Fisher's exact test was used to compare demographics between the two groups for qualitative measures, while a t-test was used to compare quantitative measures. Time to first device documented AHRE was compared via AT-free survival analysis by the Kaplan-Meier method for each AT arm. The one-year point estimate and corresponding 95% confidence interval were generated for each arm, and a Log rank test was performed to compare the rate of first occurrence or first recurrence between the two arms. A similar analysis was performed comparing time between a patient's first and second device-detected episodes between the two arms. For the analysis of AHRE burden, a bootstrap confidence interval[14] on the difference in mean burden between the two arms was applied because of the observed non-normal distribution of the burden data. A similar method was used in the analysis of symptomatic episodes. A bootstrap confidence interval on the difference in mean symptomatic episodes per patient month between the two arms was then calculated. Predictors of AHRE occurrence were determined using univariate analyses in the group without prior AT history. All quantitative demographic data are reported as mean ± SD unless indicated. A P-value <0.05 was considered significant.
Results

Recorded data from 427 of the 453 enrolled patients were analyzed, including 331 in the group without prior history of AT (Group I) and 96 patients in the prior AT history group (Group II). Twenty-six patients were excluded from the analysis cohort due to the presence of an ICD or active antiarrhythmic drug therapy at enrollment (Group I). The full 24-month follow-up period was completed by 312 of the patients. Demographic data for both groups are compared in Table 1 . The group with AT history included significantly more females than the primary arm. Patients with AT history had a higher prevalence of sinus node dysfunction, lower prevalence of atrioventricular (AV) block, and higher New York Heart Association (NYHA) Class. In patients with a history of AT, structural heart disease, as defined by coronary artery disease, was more common, while ejection fraction did not differ between groups. There were 32 deaths during the study follow-up period (approximately 7.3% in each group).

The hazard rate of first device-detected AHRE occurrence was significantly higher in patients with history of AT (Figure 1) compared to those without a history of diagnosed AT (P < 0.0001). At 12 and 24 months post-implant, the estimated Kaplan-Meier AHRE-free survival rates for patients with a history of AT were 22.6% and 11.4%, respectively. In the arm without prior AT history, a total of 159 out of 331 patients experienced at least one AHRE; the AHRE-free rates for this patient group were 59.2% and 46.2% at 12 and 24 months post-implant, respectively. Patients in the AT history group were also significantly more likely to experience a second AHRE reoccurrence. The 24-month survival free rate from device-detected AHRE was 42.1% (95% CI: 34.3-51.6%) for the 182 patients with no history of AT and a primary device indication of conduction block and 49.0% (95% CI: 40.3-59.5%) for the 119 patients with no AT history and sinus node dysfunction (P = 0.3). The 24-month AHRE survival rate for the remaining 30 patients with an indication other than SND or AV block was 56.6% (95% CI: 40.4-79.2%). AHRE burden throughout follow-up was significantly higher in the AT history group. The average burden over all follow-up in patients with no history of AT was 0.6 hours per day, compared to 4.4 hours per day among patients with a history of AT (Figure 2). A two-sided bias-corrected 95% bootstrap confidence interval on the average (AT history-no AT history) difference in AHRE burden over all follow-up was determined to be (2.23-5.55 hours per day). Since the interval does not contain 0, it was concluded that patients with AT history have greater AHRE burden than those without. Of the 156 patients in the analysis cohort with Kappa 900 devices, 25 met the prescribed criteria for persistent AT during follow-up, including 20 (26.3%) from the group with AT history and 5 (6.3%) in the group with no AT history.

Discussion

The principal findings of this study are as follows: the incidence of AT in pacemaker recipients is high, as 89% of patients with and 46% of patients without prior AT history experienced a sustained AHRE by 24 months. Prior history of AT is associated with higher arrhythmia burden, which has a non-normal distribution. Most device-detected AHRE are asymptomatic. Symptomatic AT episodes are more common in patients with prior AT but overall incidence of symptomatic AT is quite low. The positive predictive value of the Atrial High Rate Diagnostic is high.

The correlation between AHRE and actual AT events may have important clinical implications. Pollack et al[15] have demonstrated a high correlation (88%) between episodes classified by pacemaker as AHRE and intracardiac electrograms (EGMs). In the ancillary MOST analysis, an excellent correlation between AHRE and AT was provided by Holter monitoring, albeit in a small subset of patients.[16] Other studies have also demonstrated high sensitivity and specificity of pacemaker diagnosis of atrial tachyarrhythmias and close correlation with actual AT documented by ambulatory monitoring.[7,13,15-19] Our findings are in agreement with the above observations with respect to the high positive predictive value of advanced pacemaker diagnostics (90.2%).

Asymptomatic AT has been reported to be up to 12 times more frequent than symptomatic AT, even in patients with a history of "symptomatic" AT.[20] Pacemaker diagnostics have recently enabled a closer look at this otherwise difficult to diagnose arrhythmia. High incidence of recurrent AT was documented in pacemaker recipients.[8,10] Recently, Israel et al.[10] demonstrated recurrent AF in 88% of patients by review of stored EGM and in 46% of patients by electrocardiogram (ECG) recordings during follow-up in a study of 110 patients with prior AF and Class I indications for pacing. These observations clearly indicate clinical utility of pacemaker diagnostics, particularly in view of the fact that 38% of AHRE in that study were completely asymptomatic. Another interesting finding was the nonspecific nature of AT symptoms. For example, 40% of patients in that study reported symptoms but did not have any documented arrhythmias. A similarly poor correlation between pacemaker diagnosed AHRE and symptoms was found by the AIDA trial investigators [7] and an earlier transtelephonic study.[5] Similar incidence of asymptomatic AT (31% during at least 2 follow-up visits) as detected by automatic analysis of pacemaker diagnostics in patients without prior AT history was reported by Schuchert et al.[21] Another interesting observation by Israel et al. was that freedom from AT for ≥3 months did not preclude arrhythmia recurrence.[10] Our findings expand these prior observations to patients without prior AT history and confirm that AT is a recurrent arrhythmia with non-normal pattern and a high incidence of asymptomatic episodes. Therapeutic implications of these findings may be extremely important should it be eventually proven that asymptomatic AT episodes are associated with a higher stroke risk or other co-morbidities. The present trial is one of the first to investigate the prevalence of AHRE in pacemaker recipients without prior documented AT. There was no difference in mortality between patients with and without history of AT suggesting that higher AT burden observed in the former group did not result in excess mortality. However, the trial was not powered to test the impact of AHRE on survival and major morbidity such as stroke. Decisions about the use of medications were left to the treating physician. Therefore, no conclusions regarding the influence of AHRE on mortality and morbidity can be drawn from our data and further studies are needed.

AT burden in this study was found to have a non-normal distribution and was therefore analyzed using nonparametric methods such as bootstrap analysis. This finding is in agreement with prior observations on distribution and clustering of AT events.[22] Future AF studies should take into account this non-normal occurrence of AHRE events and use appropriate nonparametric statistical tools for analysis.

The observed association between AT and advanced age is consistent with prior reports.[1,2,23] Also, our analysis shows that the presence of coronary artery disease (CAD) seemed to have a "protective" role, as the incidence of AHRE was actually lower in patients with versus without history of CAD. However, this seemingly paradoxical association was no longer observed when the data were controlled for beta-blocker use by performing separate analyses for patients with beta-blockers prescribed at baseline and patients without beta-blockers prescribed at baseline. These observations suggest the potential antiarrhythmic and cardio-protective benefits of beta-adrenergic antagonists in a high-risk population, as reported previously.[24]

We observed a higher incidence of AHRE that initiated as bradycardia (atrial rate < 80 bpm) in patients without prior AT history while the majority of AHRE started as tachycardia (ventricular rate > 80 bpm) in both groups. There was no difference in AHRE onset in patients who later developed persistent AT and those who did not. Prior studies have demonstrated that the onset patterns of paroxysmal and chronic AF differ in the cycle length (i.e. shorter in patients with persistent AF) and the degree of organization.[25,26] Circadian variation of the cycle length is also present. Kerr et al.[27] have observed higher ventricular response rates characterize patients who are less likely to progress to permanent AF. While the onset characteristics of AF episodes will require further studies, our observations indicate different subtypes of arrhythmia in patients with AF.
Limitations

The number of symptomatic episodes may have been underestimated in patients who experienced a large number of episodes between device interrogations since they could have exceeded the pacemaker memory capacity. This may have partially confounded the correlation between the symptoms and AHRE and the analysis of episodes onset characteristics. However, this is applicable only to a small number of patients in this study whose pacemaker memory capacity was exceeded and thus is unlikely to make a significant impact on the results. An additional analysis comparing symptoms to daily AT activity available in a subset of devices with larger memory capacity and consequently less likely to have the earlier AHRE overwritten by later episodes (Kappa 900) also indicated that most of the symptoms documented by patients were not accompanied by AHREs. Additionally, symptoms were recorded in the diary and not directly in the pacemaker memory allowing for only approximate correlation with AHRE if both occurred on the same day.
Conclusions

The A-HIRATE study demonstrated that atrial tachyarrhythmias are common in the general pacemaker population. Patients with prior history of AT show a higher arrhythmia burden, but the subsequent incidence is quite considerable even in those patients without any history of atrial arrhythmias. The majority of atrial tachyarrhythmias in this patient population are asymptomatic and symptoms do not correspond to an actual arrhythmic episode in most patients. The Atrial High Rate Diagnostic has a high positive predictive value for identifying AT events. The distribution of AT is non-normal and commands the use of nonparametric criteria for the assessment of arrhythmia burden.