The progression of atrial fibrillation (AF) to a more sustained form is associated with increased symptoms and morbidity. The aims of the REgistry on Cardiac Rhythm DisORDers Assessing the Control of Atrial Fibrillation (RecordAF)–United States (US) cohort study were to identify the risk factors of AF progression and the effects of management approaches. RecordAF is the first worldwide, 1-year observational study of the treatment of community-based patients with recent-onset AF. We assessed AF progression at 12 months in the US cohort. AF progression was defined as a change of AF to a more sustained form (either paroxysmal becoming persistent or permanent, or persistent becoming permanent). The US cohort included 955 patients, with mean age of 68.9 years; 56.8% were men and 88.8% were white. At entry, 59.6% of patients were selected for rate-control and 40.4% for rhythm-control therapy. At 12 months, the management strategy was unchanged for 68.2% of the patients in the rate- and 77.7% of the patients in the rhythm-control groups. Overall, AF progression had occurred in 18.6% of patients at 12 months. The progression rate was significantly greater in the rate-control (27.6%) than in the rhythm-control (5.8%) group (p <0.001). Progression to permanent AF occurred in 16.4% of patients. In addition to a rate-control strategy, older age, AF rhythm at entry, persistent AF at baseline, and a history of stroke or transient ischemic attack independently predicted AF progression. Rate control was associated with AF progression, with a propensity score adjusted odds ratio of 2.67 (p <0.001). In conclusion, rate control was the preferred treatment of recent-onset AF in the US but was associated with more AF progression than rhythm control.
Atrial fibrillation (AF) is a progressive disease in which arrhythmia-induced electrical and structural remodeling facilitates evolution of AF from paroxysmal to persistent and permanent AF. More sustained forms of AF are associated with more symptoms and an increased risk of heart failure and stroke. The REgistry on Cardiac Rhythm DisORDers Assessing the Control of Atrial Fibrillation (RecordAF) was the first worldwide, 1-year, observational cohort study of the real-life treatment of patients with recently diagnosed paroxysmal or persistent AF. In the original report of the RecordAF, persistent AF, AF at baseline, AF duration (>3 months), age >75 years, heart failure, and rate-control strategy predicted progression to permanent AF. Although the results provided a global perspective, they were inevitably confounded by significant regional differences in management strategy. In particular, cardiologists in the United States (US) use rate-control therapy more often than in many other countries. The RecordAF–US cohort offered an opportunity to further define the risk factors affecting AF progression and the patterns of current AF management in the US.
Methods
A detailed description of the methods, data collection, validation, and first results of RecordAF has been previously published. The primary objectives of the RecordAF were to prospectively assess the therapeutic success with rhythm- and rate-control strategies. The physicians were randomly selected from an exhaustive global list of office- or hospital-based cardiologists. Consecutive patients were considered for enrollment if they presented with AF or had a history of AF (≤1 year from diagnosis, including patients with first-detected AF, irrespective of whether the AF was treated and regardless of the rhythm at inclusion) and were eligible for pharmacologic treatment of AF by rhythm- or rate-control agents. The exclusion criteria included permanent AF or a transient/reversible cause of AF. All patients signed an informed consent form. Rate- or rhythm-control therapy and the medications used were chosen by the patient and physician according to the physician’s standard practice. Data were collected at the baseline and 6- and 12-month visits.
The US cohort consisted of 955 patients at baseline, with 759 patients (79.5%) completing the study at 12 months. The withdrawal rate (20.5%) was higher than that in the overall RecordAF registry (7.7%). A total of 30 all-cause deaths occurred, 22 (3.9%) in the rate-control and 8 (2.1%) in the rhythm-control groups (p = 0.13). The reasons for withdrawal were patient choice (n = 40), lost to follow-up (n = 55), adverse event related to AF treatment (n = 1), other (n = 37), and missing data (n = 33). We analyzed the progression of AF in patients with AF present at both baseline and the 12-month visit. AF progression was determined by clinical assessment and was defined as a change of AF to a more sustained form at 12 months (either paroxysmal becoming persistent or permanent or persistent becoming permanent).
The baseline characteristics were summarized by descriptive statistics (mean ± SD) according to the baseline AF treatment strategy. The baseline characteristics were compared between AF treatment strategy groups using a chi-square test for categorical variables and a t test for quantitative variables. To identify the baseline factors associated with AF progression, the association between AF progression and each individual factor was assessed using Fisher’s exact test. To adjust for covariates, stepwise logistic regression analysis was performed with AF progression at 12 months as the response variable and AF treatment strategy as the primary explanatory factor (included in every model during selection). Additional factors were included or deleted according to p = 0.05 as the selection threshold. Odds ratios (ORs) and their 95% confidence intervals (CIs) were computed for the AF treatment strategy and explanatory variables retained in the model after logistic regression analysis. Missing values of a dichotomous variable were combined with the category not listed in Table 1 . For example, the missing value of a medical history (yes/no) was combined with “no” in the analysis. Missing data for the body mass index, heart rate, and CHADS 2 score were not imputed and were excluded from analysis.
| Variable | Unadjusted Data | Data Adjusted Using Inverse Propensity Score Weighting | ||||
|---|---|---|---|---|---|---|
| Rate Control (n = 569) | Rhythm Control (n = 386) | p Value | Rate Control (n = 569) | Rhythm Control (n = 386) | p Value | |
| Male gender | 317 (55.7%) | 225 (58.3%) | 0.430 | 275 (57.9%) | 283 (59.0%) | 0.734 |
| Age (yrs) | 69.2 ± 12.6 | 68.6 ± 12.4 | 0.486 | 68.7 ± 12.1 | 68.5 ± 14.0 | 0.778 |
| BMI (kg/m 2 ) | 30.0 ± 6.9 | 29.7 ± 6.7 | 0.476 | 30.0 ± 6.5 | 29.8 ± 7.6 | 0.920 |
| Ethnicity | 0.239 | 0.453 | ||||
| White | 497 (87.4%) | 351 (90.9%) | 423.6 (89.2%) | 433.1 (90.3%) | ||
| Black | 50 (8.8%) | 22 (5.7%) | 36.2 (7.6%) | 29.2 (6.1%) | ||
| Asian | 10 (1.8%) | 8 (2.1%) | 7.4 (1.57%) | 12.4 (2.6%) | ||
| Other | 12 (2.1%) | 5 (1.3%) | 7.9 (1.7%) | 5.2 (1.1%) | ||
| Heart rate (beats/min) | 74.6 ± 15.6 | 69.8 ± 15.5 | <0.001 | 72.5 (13.8%) | 73.5 (21.6%) | 0.381 |
| Coronary artery disease ∗ | 138 (24.3%) | 78 (20.2%) | 0.142 | 105.1 (22.1%) | 104.2 (21.7%) | 0.884 |
| Myocardial infarction | 52 (9.1%) | 34 (8.8%) | 0.861 | 41.5 (8.7%) | 38.2 (8.0%) | 0.671 |
| Stroke or TIA | 47 (8.3%) | 38 (9.8%) | 0.399 | 41.5 (8.7%) | 44.5 (9.3%) | 0.775 |
| Stroke | 31 (5.5%) | 19 (4.9%) | 0.720 | 24.7 (5.2%) | 19.0 (4.0%) | 0.358 |
| TIA | 26 (4.6%) | 23 (6.0%) | 0.340 | 24.8 (5.2%) | 30.0 (6.3%) | 0.488 |
| Arterial hypertension | 400 (70.3%) | 269 (69.7%) | 0.840 | 335.1 (70.5%) | 331.6 (69.1%) | 0.634 |
| Peripheral arterial disease | 22 (3.9%) | 18 (4.7%) | 0.546 | 19.3 (4.1%) | 18.9 (3.9%) | 0.922 |
| Heart failure | 91 (16.0%) | 52 (13.5%) | 0.284 | 66.3 (14.0%) | 60.5 (12.6%) | 0.541 |
| Dyslipidemia † | 315 (55.4%) | 215 (55.7%) | 0.918 | 267.1 (56.2%) | 269.5 (56.2%) | 0.987 |
| Diabetes | 112 (19.7%) | 75 (19.4%) | 0.923 | 90.4 (19.0%) | 87.0 (18.1%) | 0.721 |
| Valvular heart disease ‡ | 149 (26.2%) | 98 (25.4%) | 0.782 | 122.6 (25.8%) | 125.6 (26.2%) | 0.894 |
| Peripheral embolic events | 10 (1.7%) | 12 (3.1%) | 0.172 | 9.6 (2.0%) | 10.1 (2.1%) | 0.929 |
| Atrial flutter | 38 (6.7%) | 28 (7.3%) | 0.731 | 34.7 (7.3%) | 37.9 (7.9%) | 0.730 |
| Thyroid disease | 79 (13.9%) | 55 (14.3%) | 0.873 | 69.6 (14.7%) | 69.9 (14.6%) | 0.973 |
| Renal disease | 37 (6.5%) | 22 (5.7%) | 0.613 | 29.3 (6.2%) | 27.2 (5.7%) | 0.741 |
| Family history of AF | 69 (12.1%) | 38 (9.8%) | 0.273 | 54.2 (11.4%) | 51.3 (10.7%) | 0.722 |
| AF duration ≥3 mo | 246 (43.2%) | 194 (50.3%) | 0.031 | 223.8 (47.1%) | 220.1 (45.9%) | 0.703 |
| CHADS 2 score | 0.566 | 0.677 | ||||
| 0 | 105 (18.5%) | 67 (17.4%) | 88.5 (18.6%) | 89.6 (18.7%) | ||
| 1 | 180 (31.6%) | 137 (35.5%) | 158.5 (33.4%) | 175.9 (36.7%) | ||
| 2 | 177 (31.1%) | 108 (27.9%) | 147.3 (31.0%) | 133.9 (27.9%) | ||
| 3–6 | 101 (17.8%) | 72 (18.7%) | 80.9 (17.0%) | 80.5 (16.8%) | ||
| LVEF >50% | 383 (67.3%) | 263 (68.1%) | 0.491 | 332.1 (69.9%) | 306.9 (64.0%) | 0.023 |
| Smoking status | 0.614 | 0.445 | ||||
| Never | 248 (43.6%) | 176 (45.6%) | 205.1 (43.2%) | 200.8 (41.8%) | ||
| Current | 51 (9.0%) | 40 (10.4%) | 44.3 (9.3%) | 52.7 (11.0%) | ||
| Former | 247 (43.4%) | 159 (41.2%) | 210.1 (44.2%) | 202.4 (42.2%) | ||
| Carotid stenosis | 42 (7.4%) | 14 (3.6%) | 0.015 | 29.4 (6.2%) | 38.2 (8.0%) | 0.281 |
| Antithrombotic treatment | 479 (84.2%) | 335 (86.8%) | 0.265 | 408.7 (86.0%) | 416.3 (86.8%) | 0.736 |
| Class Ia drugs | 2 (0.4%) | 4 (1.0%) | 0.189 | 0.9 (0.2%) | 5.1 (1.1%) | 0.09 |
| Class Ic drugs | 9 (1.6%) | 46 (11.9%) | <0.001 | 8.7 (1.8%) | 55.3 (11.5%) | <0.001 |
| β Blockers except sotalol | 413 (72.6%) | 210 (54.4%) | <0.001 | 349.7 (73.6%) | 247.2 (51.5%) | <0.001 |
| Calcium channel blockers | 160 (28.1%) | 76 (19.7%) | 0.003 | 135.9 (28.6%) | 98.5 (20.5%) | 0.004 |
| Cardiac glycosides | 126 (22.1%) | 68 (17.6%) | 0.088 | 99.1 (20.9%) | 91.4 (19.0%) | 0.481 |
| Class III drugs | 31 (5.5%) | 170 (44.0%) | <0.001 | 27.5 (5.8%) | 191.8 (40.0%) | <0.001 |
| Sinus rhythm | 213 (37.4%) | 266 (68.9%) | <0.001 | 241.5 (50.8%) | 243.2 (50.7%) | 0.964 |
∗ Documented by coronary angiography or nuclear perfusion scan.
† Low-density lipoprotein >155 mg/dl and high-density lipoprotein <40 mg/dl in men and <48 mg/dl in women.
As an alternative approach to adjust between-treatment group differences for baseline characteristics, logistic regression analysis with an inverse probability of treatment-weighted (IPTW) estimator was used to estimate the OR of AF progression between rate control and rhythm control. For IPTW, a propensity score was calculated using logistic regression as the probability that a patient was selected for rate control with the baseline characteristics as explanatory variables. The baseline characteristic variables are listed in Table 1 , with the exception of the AF treatment strategy, which was the primary factor of interest. Analyses were performed using Statistical Analysis Systems statistical software, version 9.2 (SAS Institute, Cary, North Carolina).
Results
The patient characteristics according to management strategy at baseline are listed in Table 1 . At entry, 50% of patients were in sinus rhythm, 69% had paroxysmal AF, and 31% had persistent AF ( Table 2 ). Compared with the entire RecordAF population, the patients in the US cohort were older (69 vs 66 years) and had a greater body mass index (30 vs 28 kg/m 2 ), a more frequent history of smoking (56% vs 46%), more dyslipidemia (55% vs 42%), and less heart failure (15% vs 26%).
| AF Status | Treatment Strategy at Baseline | p Value | |||
|---|---|---|---|---|---|
| Rate-Control (n = 546) | Rhythm-Control (n = 370) | ||||
| Paroxysmal at Baseline | Persistent at Baseline | Paroxysmal at Baseline | Persistent at Baseline | ||
| Patients at baseline (n) | 334 (61.2%) | 212 (38.8%) | 298 (80.5%) | 72 (19.5%) | <0.001 ∗ |
| Patients at 12 mo (n) | 265 | 159 | 238 | 57 | |
| Paroxysmal | 227 (85.7%) | 38 (23.9%) | 231 (97.1%) | 27 (47.4%) | |
| Persistent | 15 (5.7%) | 42 (26.4%) | 1 (0.4%) | 20 (35.1%) | |
| Permanent | 23 (8.7%) | 79 (49.7%) | 6 (2.5%) | 10 (17.5%) | <0.001 † |
| AF progression | 38 (14.3%) | 79 (49.7%) | 7 (2.9%) | 10 (17.5%) | <0.001 ‡ ; <0.001 § |
| Total progression | 117 (27.6%) | 17 (5.8%) | <0.001 ⋮ | ||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
