The detection of undiagnosed atrial tachycardia/atrial fibrillation (AT/AF) among patients with stroke risk factors could be useful for primary stroke prevention. We analyzed newly detected AT/AF (NDAF) using continuous monitoring in patients with stroke risk factors but without previous stroke or evidence of AT/AF. NDAF (AT/AF >5 minutes on any day) was determined in patients with implantable cardiac rhythm devices and ≥1 stroke risk factors (congestive heart failure, hypertension, age ≥75 years, or diabetes). All devices were capable of continuously monitoring the daily cumulative time in AT/AF. Of 1,368 eligible patients, NDAF was identified in 416 (30%) during a follow-up of 1.1 ± 0.7 years and was unrelated to the CHADS 2 score (congestive heart failure, hypertension [blood pressure consistently >140/90 mm Hg or hypertension treated with medication], age ≥75 years, diabetes mellitus, previous stroke or transient ischemic attack). The presence of AT/AF >6 hours on ≥1 day increased significantly with increased CHADS 2 scores and was present in 158 (54%) of 294 patients with NDAF and a CHADS 2 score of ≥2. NDAF was sporadic, and 78% of patients with a CHADS 2 score of ≥2 with NDAF experienced AT/AF on <10% of the follow-up days. The median interval to NDAF detection in these higher risk patients was 72 days (interquartile range 13 to 177). In conclusion, continuous monitoring identified NDAF in 30% of patients with stroke risk factors. In patients with NDAF, AT/AF occurred sporadically, highlighting the difficulty in detecting paroxysmal AT/AF using traditional monitoring methods. However, AT/AF also persisted for >6 hours on ≥1 days in most patients with NDAF and multiple stroke risk factors. Whether patients with CHADS 2 risk factors but without a history of AF might benefit from implantable monitors for the selection and administration of anticoagulation for primary stroke prevention merits additional investigation.
We have previously reported on the incidence of newly detected atrial tachycardia (AT)/atrial fibrillation (NDAF) among patients with a history of stroke or transient ischemic attack (TIA). This information could be useful in strategies for the secondary prevention of stroke or TIA; however, primary stroke prevention strategies could have an even greater affect on both patients and the healthcare system. Detecting undiagnosed atrial fibrillation (AF) might be particularly warranted in those patients who have stroke risk factors but have not yet been diagnosed with AF (i.e., a “virtual” CHADS 2 [congestive heart failure, hypertension (blood pressure consistently >140/90 mm Hg or hypertension treated with medication), age ≥75 years, diabetes mellitus, previous stroke or transient ischemic attack] score). The recent advent of novel anticoagulants, which offer the promise of high efficacy, coupled with ease of use and lower risk of intracranial bleeding than warfarin, might shift the “tipping point” in favor of more aggressive treatment of patients with stroke risk factors on the diagnosis of AF. In addition, an earlier diagnosis of AF in these patients would permit earlier treatment with antiarrhythmic medications and upstream therapies, which might help to maintain sinus rhythm and slow the progression of the disease. The purpose of the present study was to measure, through implanted cardiac devices, the incidence and pattern of NDAF in patients with stroke risk factors, but without previous stroke. We hypothesized that continuous electrocardiographic monitoring would identify incident AF in a high proportion of patients, some of whom could be candidates for a primary stroke prevention strategy using anticoagulation therapy.
Methods
The present analysis was based on a subset of patients who were enrolled in the TRENDS study. The design and primary results of the TRENDS study have been previously published. In brief, TRENDS was a prospective, observational study that enrolled patients who had recently received a pacemaker, implantable cardioverter defibrillator, or cardiac resynchronization therapy device for a class I-II indication and who also had ≥1 of the following stroke risk factors: congestive heart failure, hypertension, age ≥65 years, diabetes, or previous stroke/TIA. A history of AF was not required for entry into the study, and warfarin use was prescribed at the discretion of the patient’s managing physician. The TRENDS study assessed the temporal relation between atrial tachycardia/AF (AT/AF) burden and the subsequent risk of stroke.
The purpose of the present analysis was to quantify the incidence and duration of NDAF in the TRENDS study population without previous known AF or stroke/TIA through the use of the continuous monitoring features available within their implanted cardiac devices (Medtronic, Minneapolis, Minnesota). The present analysis included those patients who had ≥1 established risk factor for stroke (including congestive heart failure, hypertension, age ≥75 years, or diabetes) but did not have a history of AF, stroke, or TIA. We also investigated the incidence of NDAF and the number of days with >6 hours of AT/AF among patients with intermediate (virtual CHADS 2 score of 1), high (virtual CHADS 2 score of 2), or very high (virtual CHADS 2 score of ≥3) risk of stroke. The 6-hour threshold was chosen because our main trial results showed that a similar amount of AT/AF on any day within the past 30 days conferred approximately twice the risk of thromboembolism compared to periods without AT/AF. The patients with a documented history of AF were excluded from these analyses. To exclude patients with possibly undocumented, but previously diagnosed AF, we also excluded those patients who were taking warfarin or antiarrhythmic drugs. A minimum of ≥1 day of device-stored data was required for inclusion in all analyses.
The AT/AF burden was defined as the aggregate duration of AT/AF episodes detected by the device on a given day. The AT/AF burden could be composed of multiple episodes on a single day or a portion of a single episode that spanned multiple days. The implanted devices were capable of continuously monitoring the daily AT/AF burden for up to 14 months. The AT/AF detection algorithms used in the present study have been previously shown to have high sensitivities and specificities for both the detection of AT/AF episodes and the quantification of AT/AF burden.
NDAF was defined as device-detected atrial arrhythmias lasting ≥5 minutes on any day of the study. This duration threshold was selected because previous studies have shown it to exclude most episodes of atrial oversensing that can lead to false-positive recordings of AF. To be widely applicable clinically, we used a pragmatic approach by basing our analysis on the readily available diagnostic data within the device memory. We did not distinguish between episodes of AT, atrial flutter, or AF and did not examine the intracardiac electrograms, which were not available for every episode owing to device memory limitations.
The demographic data were compared among all patients with stroke risk factors who had NDAF versus those without NDAF. We quantified the percentage of patients with ≥1 established stroke risk factors, no documented history of AF, no previous stroke/TIA, and no warfarin or antiarrhythmic drug use who experienced NDAF. The incidence of NDAF and the number of days with >6 hours of AT/AF were also compared across patients with intermediate, high, and very high stroke risk factors. Additional analyses were performed for the high and very high risk patients in whom anticoagulation has been strongly recommended in the presence of AF. The percentage of patients having >5 minutes, >1 hour, >3 hours, >6 hours, >12 hours, or 24 hours of AT/AF burden on a single day and 2 or 7 consecutive days of continuous AT/AF were tabulated for patients with CHADS 2 scores of ≥2. The interval to the identification of NDAF was evaluated using Kaplan-Meier survival estimates for patients with high and very high CHADS 2 scores. We also modeled the ability of several external rhythm monitoring regimens (24- and 48-hour and 7-, 21-, and 30-day monitoring) to identify NDAF among patients with high and very high stroke risk factors. Modeling was simulated by only considering data from the implantable device for the initial 30 days of the study and assumed 100% patient compliance.
Continuous variables are presented as the mean ± SD or median (interquartile range), as appropriate, and categorical variables are presented as percentages. Continuous variables were compared using a t test and categorical variables using the chi-square test. A significance level of p = 0.05 was used, and all statistical analyses were performed using SAS, version 9.2 (SAS Institute, Cary, North Carolina).
Results
We enrolled 2,813 patients from 116 centers in the TRENDS study. The patients with previous stroke or TIA (n = 353), an absence of CHADS 2 stroke risk factors (n = 60), a documented history of AF (n = 480), warfarin and/or antiarrhythmic drug use (n = 380), or device data unavailable (n = 172) were excluded from the present analysis ( Figure 1 ). The remaining 1,368 patients were the focus of the present analysis and were followed up for a mean of 1.1 ± 0.7 years. Of the 1,368 patients, 416 (30%, 95% confidence interval 28% to 33%) had NDAF identified by the implantable device during follow-up. The factors associated with a greater likelihood of NDAF included age >75 years and the absence of diabetes at baseline. Additional demographic information for this patient cohort is listed in Table 1 . The percentage of ventricular pacing did not alter the incidence of NDAF among the patients with pacemakers, implantable cardioverter defibrillators, or cardiac resynchronization therapy devices ( Table 2 ).
| Variable | All Patients (n = 1,368) | NDAF | p Value ⁎ | |
|---|---|---|---|---|
| Yes (n = 416) | No (n = 952) | |||
| Age (years) | 70.2 ± 11.8 | 71.1 ± 12.1 | 69.9 ± 11.7 | 0.08 |
| Men | 905 (66.2%) | 288 (69.2%) | 617 (64.8%) | 0.11 |
| CHADS 2 score | 0.92 | |||
| 1 | 411 (30.0%) | 122 (29.3%) | 289 (30.4%) | |
| 2 | 551 (40.3%) | 170 (40.9%) | 381 (40.0%) | |
| ≥3 | 406 (29.7%) | 124 (29.8%) | 282 (29.6%) | |
| Heart failure | 708 (51.8%) | 226 (54.3%) | 482 (50.6%) | 0.21 |
| Hypertension | 1,058 (77.3%) | 325 (78.1%) | 733 (77.0%) | 0.65 |
| Age >75 years | 561 (41.0%) | 187 (45.0%) | 374 (39.3%) | 0.05 |
| Diabetes mellitus | 461 (33.7%) | 120 (28.8%) | 341 (35.8%) | 0.01 |
| Previous stroke/transient ischemic attack | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | NA |
| Device type | 0.61 | |||
| Pacemaker | 593 (43.3%) | 182 (43.8%) | 411 (43.2%) | |
| Implantable cardioverter-defibrillator | 474 (34.6%) | 137 (32.9%) | 337 (35.4%) | |
| Cardiac resynchronization therapy | 301 (22.0%) | 97 (23.3%) | 204 (21.4%) | |
| Systolic blood pressure (mm Hg) | 133.8 ± 22.7 | 134.2 ± 22.8 | 133.6 ± 22.6 | 0.69 |
| Coronary artery disease † | 882 (64.5%) | 260 (62.5%) | 622 (65.3%) | 0.31 |
| Class I/III antiarrhythmic drugs | 0 | 0 | 0 | NA |
| Warfarin | 0 | 0 | 0 | NA |
| Aspirin | 941 (68.8%) | 277 (66.6%) | 664 (69.7%) | 0.25 |
| Antiplatelets | 323 (23.6%) | 91 (21.9%) | 232 (24.4%) | 0.32 |
| Statins | 787 (57.5%) | 225 (54.1%) | 562 (59.0%) | 0.09 |
| Device Type | Sample Size (n) | Ventricular Pacing (%) | Patients With NDAF (n) |
|---|---|---|---|
| Pacemaker | 593 | 55 (3–98) | 182 (30.7%) |
| Implantable cardioverter-defibrillator | 474 | 2 (0–13) | 137 (28.9%) |
| Cardiac resynchronization therapy | 301 | 99 (97–99) | 97 (32.2%) |
| p Value | <0.001 | 0.60 |
Although our analysis included patients with ≥1 stroke risk factor, we also investigated the incidence of NDAF and the incidence of days with >6 hours of AT/AF separately for patients with intermediate (virtual CHADS 2 score of 1), high (virtual CHADS 2 score of 2), and very high (virtual CHADS 2 score of ≥3) stroke risk factors. Although the incidence of NDAF (>5 minutes of AT/AF on any day) was consistent across patients with intermediate (30%), high (31%), and very high (31%) stroke risk factors (p = 0.92), a significant increase was seen in the proportion of patients with days with >6 hours of AT/AF as the virtual CHADS 2 score increased (12%, 15%, and 18% for intermediate, high, and very high risk, respectively; p = 0.04; Figure 2 ).
AT/AF persisted for durations much longer than the minimum threshold of 5 minutes in many patients with NDAF. Among the patients with NDAF and a CHADS 2 score of ≥2, Figure 3 shows the percentage who exceeded predefined burden thresholds often associated with increased stroke risk. Importantly, among those with NDAF and a CHADS 2 score of ≥2, 54% had ≥1 day with >6 hours of AT/AF and their median AT/AF duration on the day with a maximum AT/AF burden was 7.1 hours (interquartile range 1.3 to 21.6), indicating that most of these patients with NDAF and elevated stroke risk had ≥1 day with a duration shown to be associated with a significantly increased risk of stroke.
