Symptomatic bradycardia may complicate atrial fibrillation (AF) and necessitate a permanent pacemaker. Identifying patients at increased risk for symptomatic bradycardia may reduce associated morbidities and health care costs. The aim of this study was to investigate predictors for developing bradycardia requiring a permanent pacemaker in patients with AF. The records of all patients treated for AF or atrial flutter in an academic hospital’s emergency department from August 1, 2005, to July 31, 2008, were reviewed. Survival and the presence of a pacemaker as of November 1, 2011, were determined. Cases were defined as patients with pacemakers placed for bradycardia after their AF diagnoses. Patients without pacemakers who were followed constituted the control group. Variables for the logistic regression analysis were identified a priori. A post hoc model was fit adjusting for AF type and atrioventricular nodal blocker use. Of the 362 patients in the cohort, 119 cases had permanent pacemakers implanted for bradycardia after AF diagnosis, and 243 controls were alive without pacemakers. The median follow-up time was 4.5 years (interquartile range 3.8 to 5.4). Odds ratios were determined for age at the time of AF diagnosis (1.02, 95% confidence interval [CI] 1 to 1.04), female gender (1.58, 95% CI 0.95 to 2.63), previous heart failure (2.72, 95% CI 1.47 to 5.01), and African American race (0.33, 95% CI 0.12 to 0.94). The post hoc model identified permanent AF (odds ratio 2.99, 95% CI 1.61 to 5.57) and atrioventricular nodal blocker use (odds ratio 1.43, 95% CI 0.85 to 2.4). In conclusion, in patients with AF, heart failure and permanent AF each nearly triple the odds of developing bradycardia requiring a permanent pacemaker; although not statistically significant, our results suggest that women are more likely and African Americans less likely to develop bradycardia requiring pacemaker implantation.
Symptomatic bradycardia, often due to sinus node dysfunction, complicates atrial fibrillation (AF) management, often requiring permanent pacemaker implantation. Bradycardia may be a manifestation of general conduction system disease or iatrogenic, due to medications used for AF rate control. Studies examining outcomes in patients who received permanent pacemakers for sinus node dysfunction have reported that 36% to 76% had preexisting paroxysmal or persistent AF. Previous investigations have focused on the prevention of incident AF in patients with pacemakers for previously diagnosed sinus node dysfunction. The identification of patients at increased risk for symptomatic bradycardia is important for preventing associated morbidity and the related high health care costs. Improved risk stratification for symptomatic bradycardia may lead to earlier placement of a permanent pacemaker to avert associated complications such as syncope and traumatic hemorrhage in anticoagulated patients. Preventing bradycardia-related syncope in patients with AF is increasingly important given the escalating use of new oral anticoagulants without validated reversal agents and whose therapeutic levels are less easily measured in the acute setting. We investigated potential predictors for developing symptomatic bradycardia requiring a permanent pacemaker in patients with preexisting AF. We hypothesized that specific patient characteristics predispose patients with AF to the development of symptomatic bradycardia and the need for a permanent pacemaker.
Methods
We performed a secondary analysis of a retrospective, observational cohort composed of consecutive patients aged ≥18 years with primary or supporting International Classification of Diseases, Ninth Revision, emergency department discharge diagnoses of AF from August 1, 2005, to July 31, 2008. Our hospital is a university-affiliated, regional cardiac referral center that implanted 412 permanent pacemakers in 2011. The detailed method for the original investigation has been previously reported. Briefly, 2 investigators systematically reviewed the electronic medical record for corresponding data, adhering to strict chart review method guidelines. In November 2011, we reexamined the cohort’s medical records and documented additional data pertinent to this investigation, including patient survival status and the presence or absence of a pacemaker as of November 1, 2011. Patients whose pacemaker status could not be determined and those who died without having undergone the implantation of a permanent pacemaker before November 1, 2011, were excluded. We classified patients as cases if, after their AF diagnosis, they experienced symptomatic bradycardia that required an implantable permanent pacemaker. Patients without pacemakers who were still followed at our hospital constituted the control group for the primary multivariate analysis. Our medical center’s institutional review board approved this study.
The primary outcome measure was the occurrence of symptomatic bradycardia necessitating the placement of an implantable permanent pacemaker after the patient’s AF diagnosis. We defined symptomatic bradycardia as the primary cause if the reported indication for pacemaker implantation was 1 of the following: sinus node dysfunction, sick sinus syndrome, or tachycardia-bradycardia syndrome. Patients whose pacemakers were placed for other indications, including pacemakers or internal cardiac defibrillators after sudden cardiac death and biventricular pacemakers for heart failure, were excluded. We reviewed supporting clinical documentation, including cardiology and arrhythmia inpatient and outpatient consultations, electrocardiograms, cardiology clinic notes, emergency department records, and hospital admission and discharge summaries, to determine the acute symptoms resulting in evaluation for permanent pacemaker placement. We determined whether symptomatic bradycardia was the primary indication for permanent pacemaker implantation by reviewing the operative procedure note.
We recorded demographics, medical history, presence of pacemaker, operative note indication for pacemaker, symptoms prompting evaluation for pacemaker therapy, and emergent or scheduled procedure. Medical history and medication use classification was based on available information in the medical record at the time of the patient’s visit to the emergency department for symptomatic AF. Death certificates and patient death summaries were examined to determine the cause of death in those patients who had died without pacemaker implantation before November 1, 2011. For the multivariate logistic regression model, we selected a priori candidate predictors on the basis of clinical relevance and previously published associations with incident sinus node dysfunction. We included the following candidate predictors in the multivariate logistic regression: age at time of AF diagnosis, gender, previous heart failure, reported race or ethnicity, previous coronary artery disease, documented hypertension, renal insufficiency, and diabetes. Data were entered directly into IBM SPSS version 20.0 (SPSS, Inc., Chicago, Illinois).
To ensure a reliable and valid regression analysis, we adhered to the accepted formula that there must be 15 events per predictor degree of freedom (i.e., per regression coefficient estimated). We selected 8 candidate variables for our primary multivariate logistic regression analysis in accordance with these standards. We did not match the cases and controls. We adjusted for important baseline characteristics in the multivariate analysis. Descriptive statistics on baseline variables are presented as median (interquartile range) or number of nonmissing values (percentage) as appropriate. We analyzed the association of the a priori selected variables with the primary outcome using multivariate logistic regression. We also fit a second post hoc model that included 2 additional variables: AF type (defined as permanent vs persistent, paroxysmal, or newly diagnosed) and home use of atrioventricular nodal blocking medications (e.g., β-adrenergic blocking medications and nondihydropyridine calcium channel blockers) at the time of index emergency department visit. All analyses were done using the statistical programming language R version 2.14.2 (R Development Core Team, Vienna, Austria).
Results
There were 831 patients in the original cohort, and of these, 619 patients met this study’s inclusion criteria. Figure 1 provides a detailed description of patients included in the analysis. The final cohort included 119 cases who developed symptomatic bradycardia requiring permanent pacemakers. Among the 424 patients without implanted permanent pacemakers, 243 were alive without pacemaker and constituted the control group. The remaining 181 were deceased without pacemakers. Documented causes of death among these 181 patients included metastatic cancer (19%), acute decompensated heart failure (17%), sepsis (16%), acute cardiac arrest (12%), and ischemic or hemorrhagic stroke (12%). Table 1 lists the baseline cohort characteristics and symptoms necessitating pacemaker therapy. Nearly half of pacemakers (49%) were implanted during unplanned hospitalizations.
| Variable | n ‡ | Cases: Patients With Pacemaker Placed for Symptomatic Bradycardia After AF Diagnosis | Controls: Patients With AF Still Alive Without Pacemakers | Combined (n = 362) |
|---|---|---|---|---|
| (n = 119) | (n = 243) | |||
| Age (years) | 362 | 71 (64 to 80) | 62 (51 to 71) | 65 (54 to 76) |
| Women | 362 | 60 (50%) | 93 (38%) | 153 (42%) |
| Caucasian | 110 (93%) | 206 (85%) | 316 (88%) | |
| African American | 5 (4%) | 32 (13%) | 37 (10%) | |
| Latino | 2 (2%) | 3 (1%) | 5 (1%) | |
| Asian | 1 (1%) | 0 | 1 (0%) | |
| Native American | 0 | 1 (0%) | 1 (0%) | |
| Other | 0 | 1 (0%) | 1 (0%) | |
| Weight (kg) | 344 | 82 (70 to 96) | 84 (73 to 103) | 84 (72 to 100) |
| Body mass index (kg/m 2 ) | 195 | 27 (24 to 32) | 28 (24 to 33) | 28 (24 to 32) |
| Current smoker ⁎ | 361 | 7 (6%) | 27 (11%) | 34 (9%) |
| Current alcohol drinker ⁎ | 361 | 4 (3%) | 32 (13%) | 36 (10%) |
| Previous cocaine use ⁎ | 361 | 0 | 7 (3%) | 7 (2%) |
| AF type at index emergency department visit | 362 | |||
| New diagnosis | 13 (11%) | 113 (47%) | 126 (35%) | |
| Paroxysmal/persistent | 67 (56%) | 103 (42%) | 170 (47%) | |
| Permanent | 39 (33%) | 27 (11%) | 66 (18%) | |
| Lone AF | 362 | 0 | 45 (19%) | 45 (12%) |
| Previous pacemaker | 362 | 119 (100%) | 0 | 119 (33%) |
| Patient status as of November 1, 2011 | 362 | |||
| Alive with pacemaker | 83 (70%) | 0 | 83 (23%) | |
| Deceased with pacemaker | 36 (30%) | 0 | 36 (10%) | |
| Alive without pacemaker | 0 | 243 (100%) | 243 (67%) | |
| Age when pacemaker implanted (years) | 119 | 71 (64 to 79) | NA | 71 (64 to 79) |
| Age when AF diagnosed (years) | 354 | 66 (59 to 75) | 58 (48 to 68) | 61 (50 to 71) |
| Interval between pacemaker implantation and AF diagnosis (years) | 119 | 4.2 (1.1 to 7.7) | 4.2 (1.1 to 7.7) | |
| Time from index emergency department visit to November 1, 2011, follow-up (years) | 362 | 4.8 (3.8 to 5.4) | 4.4 (3.8 to 5.3) | 4.5 (3.8 to 5.4) |
| Time from index emergency department visit to pacemaker implantation (days) | 119 | 3 (−504 to 394) | 3 (−504 to 394) | |
| Reported operative note indication for pacemaker | 119 | |||
| Sinus node dysfunction | 94 (79%) | 94 (79%) | ||
| Tachycardia-bradycardia syndrome | 25 (21%) | 25 (21%) | ||
| Clinical symptoms requiring pacemaker evaluation and placement | 119 | |||
| Symptomatic bradycardia | 48 (40%) | 48 (40%) | ||
| Near syncope | 26 (22%) | 26 (22%) | ||
| Syncope | 16 (13%) | 16 (13%) | ||
| General weakness | 9 (8%) | 9 (8%) | ||
| Failed medical rate control management | 5 (4%) | 5 (4%) | ||
| Worsening heart failure | 1 (1%) | 1 (1%) | ||
| Status post cardiac arrest/critical arrhythmia | 1 (1%) | 1 (1%) | ||
| After ablation | 5 (4%) | 5 (4%) | ||
| Symptom not documented | 8 (7%) | 8 (7%) | ||
| Condition under which pacemaker placed | 119 | |||
| Unscheduled emergent/inpatient procedure | 58 (49%) | 58 (49%) | ||
| Scheduled procedure | 45 (38%) | 45 (38%) | ||
| Undetermined | 16 (13%) | 16 (13%) | ||
| Previous myocardial infarction ⁎ | 361 | 21 (18%) | 36 (15%) | 57 (16%) |
| Coronary artery disease ⁎ | 362 | 46 (39%) | 63 (26%) | 109 (30%) |
| Diabetes, insulin dependent ⁎ | 362 | 10 (8%) | 13 (5%) | 23 (6%) |
| Diabetes, non–insulin dependent ⁎ | 362 | 19 (16%) | 31 (13%) | 50 (14%) |
| Chronic obstructive pulmonary disease ⁎ | 359 | 10 (8%) | 25 (10%) | 35 (10%) |
| Hypertension ⁎ | 362 | 87 (73%) | 149 (61%) | 236 (65%) |
| Valvular heart disease ⁎ | 362 | 38 (32%) | 35 (14%) | 73 (20%) |
| Heart failure ⁎ | 362 | 34 (29%) | 29 (12%) | 63 (17%) |
| Left atrial diameter (mm) ⁎ | 295 | 45 (40 to 53) | 41 (36 to 47) | 43 (37 to 49) |
| Left ventricular ejection fraction (percentage) ⁎ | 311 | 55 (50 to 55) | 55 (55 to 55) | 55 (52.5 to 55) |
| Previous cerebrovascular accident ⁎ | 362 | 14 (12%) | 25 (10%) | 39 (11%) |
| Previous transient ischemic attack ⁎ | 362 | 7 (6%) | 6 (2%) | 13 (4%) |
| Renal insufficiency ⁎ | 362 | 13 (11%) | 15 (6%) | 28 (8%) |
| Previous cardiac surgery ⁎ | 362 | 35 (29%) | 38 (16%) | 73 (20%) |
| Previous electrical cardioversion ⁎ | 358 | 26 (22%) | 36 (15%) | 62 (17%) |
| Family history of AF | 360 | 8 (7%) | 22 (9%) | 30 (8%) |
| β-blocker use † | 362 | 65 (55%) | 93 (38%) | 158 (44%) |
| Diltiazem/verapamil use † | 362 | 19 (16%) | 31 (13%) | 50 (14%) |
| Digoxin use † | 362 | 28 (24%) | 24 (10%) | 52 (14%) |
| Warfarin use † | 362 | 64 (54%) | 63 (26%) | 127 (35%) |
| Statin use † | 362 | 51 (43%) | 84 (35%) | 135 (37%) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
