-
PDF
- Split View
-
Views
-
Cite
Cite
Daniel J Friedman, Peter Liu, Adam S Barnett, Kristen Bova Campbell, Kevin P Jackson, Tristram D Bahnson, James P Daubert, Jonathan P Piccini, Obstructive sleep apnea is associated with increased rotor burden in patients undergoing focal impulse and rotor modification guided atrial fibrillation ablation, EP Europace, Volume 20, Issue FI_3, November 2018, Pages f337–f342, https://doi.org/10.1093/europace/eux248
- Share Icon Share
Abstract
To assess whether obstructive sleep apnea (OSA) was associated with increased rotor burden among atrial fibrillation (AF) patients.
We studied 33 consecutive patients who were scheduled for focal impulse and rotor modulation (FIRM) ablation at our institution to describe the mapping, ablation, and outcomes, among patients with and without OSA. Patients underwent biatrial FIRM mapping in AF with ablation of stable rotors in addition to conventional ablation lesion sets. Differences between groups were tested with student’s t-tests and Fisher’s exact tests, as appropriate. Survival analyses were performed using the Kaplan–Meier method. Twelve of the 33 (36%) patients had OSA and 8 (66%) used continuous positive airway pressure ventilation (CPAP). Obstructive sleep apnea patients had a higher body mass index (BMI) (33.6 vs. 28.8 kg/m2, P = 0.01) and were more commonly on beta blockers (67% vs. 29%, P = 0.03) but were otherwise similar regarding baseline characteristics, medication use, and prior AF treatments, including antiarrhythmic drugs and prior ablation. Focal impulse and rotor modulation mapping demonstrated increased rotor burden in the OSA patients (2.6 ± 0.9 vs. 2.0 ± 1.0, P =0.03). The increased rotor burden was more evident in the right atrium (RA) (1.0 ± 0.7 vs. 0.5 ± 0.7, P =0.04 compared with left atrium (1.7 ± 0.8 vs. 1.4 ± 0.7, P = 0.15). There was no correlation between BMI and total number of rotors (r = 0.0961, P = 0.59). Among the population of patients with OSA, CPAP therapy was associated with a lower number of RA rotors (0.8 ± 0.7 vs. 1.5 ± 0.6, P = 0.05) but no significant difference in overall rotors (P = 0.33).
Obstructive sleep apnea patients demonstrate increased rotor prevalence, driven predominantly by an increase in RA rotors. CPAP therapy was associated with fewer RA rotors.
Atrial fibrillation (AF) patients with obstructive sleep apnea (OSA) have increased rotor burden compared with those without OSA.
The increased rotor burden among OSA patients is largely driven by an increase in right atrium (RA) rotors.
The increase in RA rotors among OSA patients may explain higher rates of AF recurrence and AF progression among this important subgroup of AF patients.
Introduction
Untreated obstructive sleep apnea (OSA) is an established risk factor for new onset atrial fibrillation (AF),1 progression from paroxysmal AF to persistent AF,2 and AF recurrence after catheter ablation.3,4,5 The relationship between OSA and AF is complex and is mediated by indirect effects due to associated comorbidities (diabetes, obesity, and hypertension) and the direct effects of OSA on the atrial substrate and arrhythmia triggers.1 However, the relationship between OSA and electrophysiologic mechanisms of AF maintenance remain unclear.
Recent advances in AF mapping have identified rotors as an important mechanism driving the maintenance of human AF.6–8 The Topera mapping system (Abbott Electrophysiology, Menlo Park, CA, USA) generates phase maps from the unipolar signals from a 64-pole basket catheter and creates AF isopotential movies that allow rotors to be identified based on the presence of rotational activity. The role of rotors as AF drivers has been supported by numerous studies demonstrating that Focal Impulse and Rotor Modulation (FIRM) guided rotor ablation is associated with acute termination of 9,10 and long-term reduction in AF.11 Rotor ablation has also been performed using non-proprietary mapping algorithms.7,8 In order to determine whether OSA was associated with differences in rotor burden and clinical outcomes, we analysed consecutive patients who underwent FIRM guided rotor mapping and ablation. We hypothesized that OSA would be associated with higher rotor density and higher risk of recurrent AF.
Methods
Study population
We studied consecutive patients who underwent FIRM guided AF ablation at Duke University Medical Center between 16 October 2012 and 16 June 2015. All patients failed ≥1 antiarrhythmic drug or had a compelling reason (e.g. contraindication) to avoid antiarrhythmic drug therapy. Patients had AF documented by either 12 lead ECG or device interrogation. De novo and redo AF ablations were included. Patients underwent pulmonary vein isolation (or redo pulmonary vein isolation) with adjunctive FIRM ablation due to the presence of one or more risk factors for AF, including recurrent AF despite prior AF ablation, non-paroxysmal AF, left atrial diameter >5 cm, or the presence of cardiomyopathy. Patients who underwent redo ablations underwent adjunctive FIRM and, as well as pulmonary vein isolation if there was any evidence of recurrent pulmonary venous conduction. Focal impulse and rotor modulation mapping and rotor ablation were performed before pulmonary vein isolation (or re-isolation).
Obstructive sleep apnea was defined as polysomnography diagnosed OSA that was of sufficient severity to merit treatment with continuous positive airway pressure (CPAP, apnea hypopnea index of ≥15 or apnea hypopnea index of 5–14 with symptoms).12 Since AF is considered a symptom of OSA, all patients in our study with an apnea hypopnea index of ≥5 were considered to have OSA.
All patients underwent a pre-procedure transoesophageal echocardiogram to exclude left atrial and left atrial appendage thrombus. Pre-procedure cardiac imaging was obtained via magnetic resonance imaging or computed tomography. The study was approved by the Duke University Institutional Review Board.
Electrophysiology study
All patients underwent electrophysiologic study and ablation under general anaesthesia. Heparin was administered before or at the time of trans-septal puncture and activated clotting times were maintained between 300 and 400 s. Left atrium (LA) and right atrium (RA) pressures were measured via the trans-septal needle and sheaths and compared to confirm successful trans-septal puncture. Ablation was performed with open-irrigated catheters and the use of an electroanatomic mapping system (CARTO, NavX, or Rhythmia). Intracardiac ultrasound was used in all patients. Pulmonary vein isolation (or redo pulmonary vein isolation) was performed using a circumferential approach. Entrance and exit block was confirmed with a circular multipolar catheter.
Focal impulse and rotor modulation mapping and ablation
Focal impulse and rotor modulation mapping in AF has been described previously.11 Patients who presented in normal sinus rhythm underwent AF induction using rapid atrial pacing (under isoproterenol if needed) and remained in AF for at least 5 min prior to mapping to ensure rotor stability. A 64-electrode basket catheter (Constellation, Boston Scientific, Natick, MA, USA or FIRMmap, Abbott Electrophysiology, Menlo Park, CA, USA) was selected based on atrial size from the pre-procedural magnetic resonance imaging (MRI) or computed tomography (CT) and inserted into the RA. The basket catheter was repositioned as necessary to ensure spline separation and atrial contact and coverage using fluoroscopy and ultrasound. The unipolar electrogram (EGMs) from the basket catheter electrodes were filtered at 0.05–500 Hz and recorded at 1 kHz sampling frequency with a notch filter at 60 Hz. Continuous recordings of AF were exported into the computational FIRM mapping system (RhythmView™, Abbott Electrophysiology, Menlo Park, CA, USA) allowing for the generation of isochronal projections of AF propagation. Rotors were defined as stable areas of circular rotational activity around a central location that demonstrated limited precession. Each rotor was specifically targeted for ablation (total 1.5–2 cm2) unless located prohibitively close to a critical structure [e.g. atrioventricular (AV) node or phrenic nerve] or within a planned empiric lesion set. After each rotor was ablated, repeat mapping was performed to confirm successful ablation. The basket catheter was repositioned as needed to ensure complete mapping of the entire chamber. Subsequently, the process was repeated in the LA, followed by additional lesion sets (e.g. pulmonary vein isolation) as needed.
Clinical and procedural endpoints
Acute procedural success was defined by AF cycle length slowing of 10% or more, conversion from AF to an organized atrial tachycardia (AT), or conversion from AF to normal sinus rhythm (NSR), with FIRM ablation. AF cycle length was defined as the average of 10 consecutive cycles as measured from the coronary sinus catheter bipole with the sharpest deflections.
Complications and arrhythmia free survival were determined retrospectively based on chart review. Arrhythmia recurrence was defined as atrial tachycardia, atrial flutter, or atrial fibrillation (AT/AF/AFL) on a 12-lead electrocardiogram or AT/AF/AFL ≥30 s on a continuous monitor or implantable device. A 90-days blanking period was employed per guideline recommendations.13
Statistical analysis
We tested the hypothesis that patients with OSA had a greater number of total, right atrial, and/or left atrial rotors. Continuous variables were described using means ± standard deviations and categorical variables were described using proportions. Student’s t-tests were used to compare differences among continuous variables and Fisher’s exact tests or χ2 tests were used to compare differences among categorical variables. Survival analyses were performed using the Kaplan–Meier method and differences were assessed using Log Rank testing. Correlations were calculated using the Spearman’s Rho test. Given the exploratory nature of the analyses and the physiologic basis for the study hypothesis, all tests were one sided. A P-value of <0.05 was considered significant. All analyses were performed using JMP Pro 12 (SAS Institutes, Cary, NC, USA).
Results
Thirty-three patients were scheduled for FIRM mapping and ablation at Duke University Medical Center between 16 October 2012 and 16 June 2015 (Table 1). The average age was 62 ± 11 years, 27% of patients were female, the mean LA diameter was 4.4 ± 0.7 cm and the average body mass index (BMI) was 30.5 ± 5.8 km/m2. The majority of patients (58%) had persistent AF, 30% had paroxysmal AF, and 12% had longstanding persistent AF. Fourteen patients had undergone prior AF ablation and 58% had failed ≥2 antiarrhythmic drugs. Twelve of the 33 patients (36%) have OSA and 8 of these patients used CPAP.
Baseline characteristics of entire population and after stratification by presence of comorbid OSA
. | All (n = 33) . | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|---|
Age (years) | 62 (11) | 61.3 (11.2) | 63.0 (11.8) | 0.32 |
Female (%) | 27 (9) | 25 (3) | 29 (6) | 0.73 |
AF type | 0.42 | |||
Paroxysmal (%) | 30 (10) | 17 (2) | 38 (8) | |
Persistent (%) | 58 (19) | 67 (8) | 52 (11) | |
Longstanding persistent (%) | 12 (4) | 17 (2) | 10 (2) | |
BMI (kg/m2) | 30.5 (5.8) | 33.6 (4.6) | 28.8 (5.7) | 0.01 |
CHA2DS2VASC | 2.2 (1.9) | 2.5 (2.0) | 2.0 (1.9) | 0.24 |
LVEF | 54 (9) | 52 (10) | 55 (9) | 0.62 |
LA diameter (cm) | 4.4 (0.7) | 4.7 (0.6) | 4.3 (0.8) | 0.08 |
Moderate or severe LAE (%) | 42 (14) | 58 (7) | 33 (7) | 0.16 |
Moderate or severe RAE (%) | 27 (9) | 33 (4) | 24 (5) | 0.42 |
RV dysfunctiona (%) | 7 (2) | 20 (2) | 0 (0) | 0.11 |
Moderate or severe TR (%) | 6 (2) | 18 (2) | 0 (0) | 0.11 |
Prior AF ablation (%) | 39 (13) | 42 (5) | 38 (8) | 0.56 |
CHF (%) | 36 (8) | 33 (4) | 19 (4) | 0.30 |
HTN (%) | 61 (20) | 75 (9) | 52 (11) | 0.20 |
DM (%) | 12 (4) | 8 (1) | 14 (3) | |
Vascular disease (%) | 12 (4) | 25 (3) | 5 (1) | 0.13 |
CAD (%) | 30 (10) | 42 (5) | 24 (5) | 0.85 |
CPAP (%) | 67 (8) | N/A | N/A | |
Previously failed >1AAD (%) | 58 (19) | 58 (7) | 57 (12) | 0.95 |
ACE/ARB (%) | 39 (13) | 50 (6) | 33 (7) | 0.28 |
Beta blocker (%) | 42 (14) | 67 (8) | 29 (6) | 0.03 |
Spironolactone (%) | 12 (4) | 25 (3) | 5 (1) | 0.13 |
Non-dihydropyridine calcium channel blocker (%) | 30 (10) | 42 (5) | 24 (5) | 0.25 |
Statin (%) | 36 (12) | 42 (5) | 33 (7) | 0.46 |
Implantable device (%) | 6 (2) | 17 (2) | 0 (0) | 0.13 |
. | All (n = 33) . | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|---|
Age (years) | 62 (11) | 61.3 (11.2) | 63.0 (11.8) | 0.32 |
Female (%) | 27 (9) | 25 (3) | 29 (6) | 0.73 |
AF type | 0.42 | |||
Paroxysmal (%) | 30 (10) | 17 (2) | 38 (8) | |
Persistent (%) | 58 (19) | 67 (8) | 52 (11) | |
Longstanding persistent (%) | 12 (4) | 17 (2) | 10 (2) | |
BMI (kg/m2) | 30.5 (5.8) | 33.6 (4.6) | 28.8 (5.7) | 0.01 |
CHA2DS2VASC | 2.2 (1.9) | 2.5 (2.0) | 2.0 (1.9) | 0.24 |
LVEF | 54 (9) | 52 (10) | 55 (9) | 0.62 |
LA diameter (cm) | 4.4 (0.7) | 4.7 (0.6) | 4.3 (0.8) | 0.08 |
Moderate or severe LAE (%) | 42 (14) | 58 (7) | 33 (7) | 0.16 |
Moderate or severe RAE (%) | 27 (9) | 33 (4) | 24 (5) | 0.42 |
RV dysfunctiona (%) | 7 (2) | 20 (2) | 0 (0) | 0.11 |
Moderate or severe TR (%) | 6 (2) | 18 (2) | 0 (0) | 0.11 |
Prior AF ablation (%) | 39 (13) | 42 (5) | 38 (8) | 0.56 |
CHF (%) | 36 (8) | 33 (4) | 19 (4) | 0.30 |
HTN (%) | 61 (20) | 75 (9) | 52 (11) | 0.20 |
DM (%) | 12 (4) | 8 (1) | 14 (3) | |
Vascular disease (%) | 12 (4) | 25 (3) | 5 (1) | 0.13 |
CAD (%) | 30 (10) | 42 (5) | 24 (5) | 0.85 |
CPAP (%) | 67 (8) | N/A | N/A | |
Previously failed >1AAD (%) | 58 (19) | 58 (7) | 57 (12) | 0.95 |
ACE/ARB (%) | 39 (13) | 50 (6) | 33 (7) | 0.28 |
Beta blocker (%) | 42 (14) | 67 (8) | 29 (6) | 0.03 |
Spironolactone (%) | 12 (4) | 25 (3) | 5 (1) | 0.13 |
Non-dihydropyridine calcium channel blocker (%) | 30 (10) | 42 (5) | 24 (5) | 0.25 |
Statin (%) | 36 (12) | 42 (5) | 33 (7) | 0.46 |
Implantable device (%) | 6 (2) | 17 (2) | 0 (0) | 0.13 |
AAD, antiarrhythmic drug; AF, atrial fibrillation; BMI, body mass index; CHF, congestive heart failure; CPAP, continuous positive airway pressure; DM, diabetes mellitus; HTN, hypertension; ICM, ischaemic cardiomyopathy; LA, left atrium; LVEF, left ventricular ejection fraction; OSA, obstructive sleep apnea.
Two patients had mild RV dysfunction and the remaining patients (n = 27) had normal RV function; RV function could not be adequately assessed in four patients.
Baseline characteristics of entire population and after stratification by presence of comorbid OSA
. | All (n = 33) . | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|---|
Age (years) | 62 (11) | 61.3 (11.2) | 63.0 (11.8) | 0.32 |
Female (%) | 27 (9) | 25 (3) | 29 (6) | 0.73 |
AF type | 0.42 | |||
Paroxysmal (%) | 30 (10) | 17 (2) | 38 (8) | |
Persistent (%) | 58 (19) | 67 (8) | 52 (11) | |
Longstanding persistent (%) | 12 (4) | 17 (2) | 10 (2) | |
BMI (kg/m2) | 30.5 (5.8) | 33.6 (4.6) | 28.8 (5.7) | 0.01 |
CHA2DS2VASC | 2.2 (1.9) | 2.5 (2.0) | 2.0 (1.9) | 0.24 |
LVEF | 54 (9) | 52 (10) | 55 (9) | 0.62 |
LA diameter (cm) | 4.4 (0.7) | 4.7 (0.6) | 4.3 (0.8) | 0.08 |
Moderate or severe LAE (%) | 42 (14) | 58 (7) | 33 (7) | 0.16 |
Moderate or severe RAE (%) | 27 (9) | 33 (4) | 24 (5) | 0.42 |
RV dysfunctiona (%) | 7 (2) | 20 (2) | 0 (0) | 0.11 |
Moderate or severe TR (%) | 6 (2) | 18 (2) | 0 (0) | 0.11 |
Prior AF ablation (%) | 39 (13) | 42 (5) | 38 (8) | 0.56 |
CHF (%) | 36 (8) | 33 (4) | 19 (4) | 0.30 |
HTN (%) | 61 (20) | 75 (9) | 52 (11) | 0.20 |
DM (%) | 12 (4) | 8 (1) | 14 (3) | |
Vascular disease (%) | 12 (4) | 25 (3) | 5 (1) | 0.13 |
CAD (%) | 30 (10) | 42 (5) | 24 (5) | 0.85 |
CPAP (%) | 67 (8) | N/A | N/A | |
Previously failed >1AAD (%) | 58 (19) | 58 (7) | 57 (12) | 0.95 |
ACE/ARB (%) | 39 (13) | 50 (6) | 33 (7) | 0.28 |
Beta blocker (%) | 42 (14) | 67 (8) | 29 (6) | 0.03 |
Spironolactone (%) | 12 (4) | 25 (3) | 5 (1) | 0.13 |
Non-dihydropyridine calcium channel blocker (%) | 30 (10) | 42 (5) | 24 (5) | 0.25 |
Statin (%) | 36 (12) | 42 (5) | 33 (7) | 0.46 |
Implantable device (%) | 6 (2) | 17 (2) | 0 (0) | 0.13 |
. | All (n = 33) . | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|---|
Age (years) | 62 (11) | 61.3 (11.2) | 63.0 (11.8) | 0.32 |
Female (%) | 27 (9) | 25 (3) | 29 (6) | 0.73 |
AF type | 0.42 | |||
Paroxysmal (%) | 30 (10) | 17 (2) | 38 (8) | |
Persistent (%) | 58 (19) | 67 (8) | 52 (11) | |
Longstanding persistent (%) | 12 (4) | 17 (2) | 10 (2) | |
BMI (kg/m2) | 30.5 (5.8) | 33.6 (4.6) | 28.8 (5.7) | 0.01 |
CHA2DS2VASC | 2.2 (1.9) | 2.5 (2.0) | 2.0 (1.9) | 0.24 |
LVEF | 54 (9) | 52 (10) | 55 (9) | 0.62 |
LA diameter (cm) | 4.4 (0.7) | 4.7 (0.6) | 4.3 (0.8) | 0.08 |
Moderate or severe LAE (%) | 42 (14) | 58 (7) | 33 (7) | 0.16 |
Moderate or severe RAE (%) | 27 (9) | 33 (4) | 24 (5) | 0.42 |
RV dysfunctiona (%) | 7 (2) | 20 (2) | 0 (0) | 0.11 |
Moderate or severe TR (%) | 6 (2) | 18 (2) | 0 (0) | 0.11 |
Prior AF ablation (%) | 39 (13) | 42 (5) | 38 (8) | 0.56 |
CHF (%) | 36 (8) | 33 (4) | 19 (4) | 0.30 |
HTN (%) | 61 (20) | 75 (9) | 52 (11) | 0.20 |
DM (%) | 12 (4) | 8 (1) | 14 (3) | |
Vascular disease (%) | 12 (4) | 25 (3) | 5 (1) | 0.13 |
CAD (%) | 30 (10) | 42 (5) | 24 (5) | 0.85 |
CPAP (%) | 67 (8) | N/A | N/A | |
Previously failed >1AAD (%) | 58 (19) | 58 (7) | 57 (12) | 0.95 |
ACE/ARB (%) | 39 (13) | 50 (6) | 33 (7) | 0.28 |
Beta blocker (%) | 42 (14) | 67 (8) | 29 (6) | 0.03 |
Spironolactone (%) | 12 (4) | 25 (3) | 5 (1) | 0.13 |
Non-dihydropyridine calcium channel blocker (%) | 30 (10) | 42 (5) | 24 (5) | 0.25 |
Statin (%) | 36 (12) | 42 (5) | 33 (7) | 0.46 |
Implantable device (%) | 6 (2) | 17 (2) | 0 (0) | 0.13 |
AAD, antiarrhythmic drug; AF, atrial fibrillation; BMI, body mass index; CHF, congestive heart failure; CPAP, continuous positive airway pressure; DM, diabetes mellitus; HTN, hypertension; ICM, ischaemic cardiomyopathy; LA, left atrium; LVEF, left ventricular ejection fraction; OSA, obstructive sleep apnea.
Two patients had mild RV dysfunction and the remaining patients (n = 27) had normal RV function; RV function could not be adequately assessed in four patients.
Obstructive sleep apnea patients had a higher BMI (33.6 vs. 28.8 kg/m2, P = 0.01) and were more commonly on beta blockers (67% vs. 29%, P =0.03) compared with patients without OSA, but were otherwise similar regarding baseline characteristics, medication use, and prior AF treatments, including antiarrhythmic drugs and prior ablation. There was trend towards larger LA diameter among OSA patients compared with those without OSA (4.7 vs. 4.3 cm, P = 0.08). Notably, there was no difference in the frequency of diabetes, hypertension, or heart failure (Table 1).
Intraprocedural findings and outcomes among patients with and without OSA are detailed in Table 2. OSA patients had a higher mean RA pressure compared with patients without OSA (11 ± 5 mmHg vs. 7 ± 5 mmHg, P = 0.03); although LA pressures (mean and peak V wave) were higher among OSA patients, the differences were not statistically significant. The overall procedure, ablation, and fluoroscopy times were 350 ± 101, 58 ± 27, and 33 ± 15 min, respectively. The most commonly used basket sizes were 60 mm (n = 9), 50 mm (n = 9), and 48 mm (n = 4). The remaining patients were mapped with a 42 mm basket (n = 1), a 70 mm basket (n = 1), or both 48 and 60 mm baskets (n = 1).
. | All (n = 33) . | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|---|
Mean RA pressure (mmHg) | 8 (5) | 11 (5) | 7 (5) | 0.03 |
Mean LA pressure (mmHg) | 14 (5) | 16 (6) | 14 (5) | 0.14 |
LA V wave (mmHg) | 19 (7) | 22 (6) | 18 (7) | 0.13 |
Total # of rotors | 2.2 (1.0) | 2.6 (0.9) | 2.0 (1.0) | 0.03 |
#RA rotors | 0.70 (0.73) | 1.0 (0.7) | 0.5 (0.7) | 0.04 |
#LA rotors | 1.48 (0.76) | 1.7 (0.8) | 1.4 (0.7) | 0.15 |
Acute procedural success composite (%) | 39 (12) | 36 (4) | 47 (8) | 0.44 |
AF to SR with FIRM, n | 5 | 2 | 3 | |
AF to AT/AFL, n | 2 | 0 | 2 | |
10% AF CL slowing with FIRMa, n | 5 | 2 | 3 | |
Procedure time (min) | 350 (101) | 298 (92) | 380 (96) | 0.97 |
Radiofrequency time (min) | 58 (27) | 52 (28) | 62 (27) | 0.69 |
Fluoroscopy time (min) | 33 (15) | 29 (13) | 36 (15) | 0.77 |
Radiation dose | ||||
mGy | 542 (656) | 762 (992) | 431 (389) | 0.17 |
mGy-cm2 | 85 078 (79 607) | 99 407 (108 120) | 76 890 (59 300) | 0.26 |
. | All (n = 33) . | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|---|
Mean RA pressure (mmHg) | 8 (5) | 11 (5) | 7 (5) | 0.03 |
Mean LA pressure (mmHg) | 14 (5) | 16 (6) | 14 (5) | 0.14 |
LA V wave (mmHg) | 19 (7) | 22 (6) | 18 (7) | 0.13 |
Total # of rotors | 2.2 (1.0) | 2.6 (0.9) | 2.0 (1.0) | 0.03 |
#RA rotors | 0.70 (0.73) | 1.0 (0.7) | 0.5 (0.7) | 0.04 |
#LA rotors | 1.48 (0.76) | 1.7 (0.8) | 1.4 (0.7) | 0.15 |
Acute procedural success composite (%) | 39 (12) | 36 (4) | 47 (8) | 0.44 |
AF to SR with FIRM, n | 5 | 2 | 3 | |
AF to AT/AFL, n | 2 | 0 | 2 | |
10% AF CL slowing with FIRMa, n | 5 | 2 | 3 | |
Procedure time (min) | 350 (101) | 298 (92) | 380 (96) | 0.97 |
Radiofrequency time (min) | 58 (27) | 52 (28) | 62 (27) | 0.69 |
Fluoroscopy time (min) | 33 (15) | 29 (13) | 36 (15) | 0.77 |
Radiation dose | ||||
mGy | 542 (656) | 762 (992) | 431 (389) | 0.17 |
mGy-cm2 | 85 078 (79 607) | 99 407 (108 120) | 76 890 (59 300) | 0.26 |
AF, atrial fibrillation; AFL, atrial flutter; CL, cycle length; FIRM, focal impulse and rotor modification; LA, left atrium; OSA, obstructive sleep apnea; RA, right atrium; SR, sinus rhythm.
Unable to assess for two patients with missing paired pre- and post-rotor ablation electrograms.
. | All (n = 33) . | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|---|
Mean RA pressure (mmHg) | 8 (5) | 11 (5) | 7 (5) | 0.03 |
Mean LA pressure (mmHg) | 14 (5) | 16 (6) | 14 (5) | 0.14 |
LA V wave (mmHg) | 19 (7) | 22 (6) | 18 (7) | 0.13 |
Total # of rotors | 2.2 (1.0) | 2.6 (0.9) | 2.0 (1.0) | 0.03 |
#RA rotors | 0.70 (0.73) | 1.0 (0.7) | 0.5 (0.7) | 0.04 |
#LA rotors | 1.48 (0.76) | 1.7 (0.8) | 1.4 (0.7) | 0.15 |
Acute procedural success composite (%) | 39 (12) | 36 (4) | 47 (8) | 0.44 |
AF to SR with FIRM, n | 5 | 2 | 3 | |
AF to AT/AFL, n | 2 | 0 | 2 | |
10% AF CL slowing with FIRMa, n | 5 | 2 | 3 | |
Procedure time (min) | 350 (101) | 298 (92) | 380 (96) | 0.97 |
Radiofrequency time (min) | 58 (27) | 52 (28) | 62 (27) | 0.69 |
Fluoroscopy time (min) | 33 (15) | 29 (13) | 36 (15) | 0.77 |
Radiation dose | ||||
mGy | 542 (656) | 762 (992) | 431 (389) | 0.17 |
mGy-cm2 | 85 078 (79 607) | 99 407 (108 120) | 76 890 (59 300) | 0.26 |
. | All (n = 33) . | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|---|
Mean RA pressure (mmHg) | 8 (5) | 11 (5) | 7 (5) | 0.03 |
Mean LA pressure (mmHg) | 14 (5) | 16 (6) | 14 (5) | 0.14 |
LA V wave (mmHg) | 19 (7) | 22 (6) | 18 (7) | 0.13 |
Total # of rotors | 2.2 (1.0) | 2.6 (0.9) | 2.0 (1.0) | 0.03 |
#RA rotors | 0.70 (0.73) | 1.0 (0.7) | 0.5 (0.7) | 0.04 |
#LA rotors | 1.48 (0.76) | 1.7 (0.8) | 1.4 (0.7) | 0.15 |
Acute procedural success composite (%) | 39 (12) | 36 (4) | 47 (8) | 0.44 |
AF to SR with FIRM, n | 5 | 2 | 3 | |
AF to AT/AFL, n | 2 | 0 | 2 | |
10% AF CL slowing with FIRMa, n | 5 | 2 | 3 | |
Procedure time (min) | 350 (101) | 298 (92) | 380 (96) | 0.97 |
Radiofrequency time (min) | 58 (27) | 52 (28) | 62 (27) | 0.69 |
Fluoroscopy time (min) | 33 (15) | 29 (13) | 36 (15) | 0.77 |
Radiation dose | ||||
mGy | 542 (656) | 762 (992) | 431 (389) | 0.17 |
mGy-cm2 | 85 078 (79 607) | 99 407 (108 120) | 76 890 (59 300) | 0.26 |
AF, atrial fibrillation; AFL, atrial flutter; CL, cycle length; FIRM, focal impulse and rotor modification; LA, left atrium; OSA, obstructive sleep apnea; RA, right atrium; SR, sinus rhythm.
Unable to assess for two patients with missing paired pre- and post-rotor ablation electrograms.
Focal impulse and rotor modulation mapping demonstrated an average of 2.2 ± 1.0 rotors per patient. In the overall cohort, rotors were more common in the LA (1.5 ± 0.8) compared with the RA (0.7 ± 0.7). Figure 1 is a schematic depiction of the rotor locations across both atria for OSA and non-OSA patients. Every patient with OSA was observed to have a rotor during mapping. Ninety-two per cent (n = 11) of OSA patients had at least one LA rotor and 75% (n = 9) had at least one RA rotor. Rotors were observed in 90% (n = 19) of patients without OSA. Similar to patients with OSA, patients without OSA frequently had LA rotors (86%, n = 18). In contrast to patients with OSA, patients without OSA had RA rotors less than half of the time (43%, n = 9).

Schematic diagram of the atria depicting rotor locations among OSA (stars) and non-OSA patients (circles). LAA, left atrial appendage; LPV, left pulmonary veins; MV, mitral valve; RA, right atrium; RPV, right pulmonary veins; SVC, superior vena cava; TV, tricuspid valve.
We observed increased rotor burden in OSA patients compared with those without OSA (2.6 ± 0.9 vs. 2.0 ± 1.0, P =0.03). The increased rotor burden was more evident in the RA (1.0 ± 0.7 vs. 0.5 ± 0.7, P =0.04) compared with LA (1.7 ± 0.8 vs. 1.4 ± 0.7, P = 0.15). The increase in rotor burden among OSA patients was particularly evidence along the lateral RA and crista terminalis where 50% (n = 6) of OSA patients had at least one rotor, compared with only 10% (n = 2) of patients without OSA (P = 0.02). Among the population of patients with OSA, CPAP therapy was associated with a lower number of RA rotors (0.8 ± 0.7 vs. 1.5 ± 0.6, P = 0.05) but no significant difference in overall rotors (P = 0.33).
To address the possibility that a higher burden of non-paroxysmal AF in the OSA group may have affected the results, we performed a sensitivity analysis of patients with non-paroxysmal AF (n = 23). Within this subgroup, OSA patients had a higher burden of RA rotors (1.1 ± 0.6 vs. 0.5 ± 0.7, P = 0.02) compared with patients without OSA. Although OSA patients had numerically more total rotors compared with patients without OSA (2.7 ± 0.9 vs. 2.2 ± 0.8), the difference was not statistically significant (P = 0.11) in this smaller subset. We additionally assessed whether a history of prior RA ablation could account for the differences in RA rotor burden by OSA status. In sensitivity analyses excluding patients with definite or possible prior RA ablation (n = 5), OSA patients were more likely to have RA rotors (0.9 ± 0.6 vs. 0.5 ± 0.6, P = 0.05) and more rotors overall (2.6 ± 0.9 vs. 1.9 ± 0.9, P = 0.045).
Of note, there was no correlation between BMI and total number of rotors (r = 0.0961, P = 0.59), left atrial rotors (r = −0.0858, P = 0.63), or right atrial rotors (r = 0.1533, P = 0.39). Patients without OSA were more likely to meet an acute procedural endpoint with FIRM ablation compared with patients with OSA (47% vs. 36%) although this did not meet statistical significance (P = 0.44).
Focal impulse and rotor modulation mapping and ablation were generally safe among those with and without OSA; complications were only observed among those without OSA (Table 3). One patient had an access site haematoma and one patient with baseline pulmonary vein stenosis developed radiographic evidence of worsened pulmonary vein stenosis that was asymptomatic and did not require intervention. A patient with an infiltrative cardiomyopathy had an unremarkable FIRM procedure but she died in follow-up due to cardiac insufficiency and PEA arrest (no evidence of an effusion on echocardiogram).
. | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|
Haematomaa | 0 | 4.8 (1) | 1.00 |
Pulmonary vein stenosisb | 0 | 4.8 (1) | 1.00 |
Deathc | 0 | 4.8 (1) | 1.00 |
Hypotensionc | 0 | 4.8 (1) | 1.00 |
. | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|
Haematomaa | 0 | 4.8 (1) | 1.00 |
Pulmonary vein stenosisb | 0 | 4.8 (1) | 1.00 |
Deathc | 0 | 4.8 (1) | 1.00 |
Hypotensionc | 0 | 4.8 (1) | 1.00 |
No haemopericardium, deep vein thrombosis, tamponade, transient ischaemic attack, myocardial infarction, stroke, blood transfusion, retroperitoneal bleed, atrioventricular fistula, pseudoaneurysm, phrenic nerve injury, air embolism, or atrio-oesophageal fistula were observed.
Did not require transfusion.
Asymptomatic, did not require intervention.
Same patient. No other periprocedural complications (e.g. tamponade, haemorrhage) were found and the decompensation was attributed to complications of infiltrative cardiomyopathy.
. | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|
Haematomaa | 0 | 4.8 (1) | 1.00 |
Pulmonary vein stenosisb | 0 | 4.8 (1) | 1.00 |
Deathc | 0 | 4.8 (1) | 1.00 |
Hypotensionc | 0 | 4.8 (1) | 1.00 |
. | OSA (n = 12) . | No OSA (n = 21) . | P-value . |
---|---|---|---|
Haematomaa | 0 | 4.8 (1) | 1.00 |
Pulmonary vein stenosisb | 0 | 4.8 (1) | 1.00 |
Deathc | 0 | 4.8 (1) | 1.00 |
Hypotensionc | 0 | 4.8 (1) | 1.00 |
No haemopericardium, deep vein thrombosis, tamponade, transient ischaemic attack, myocardial infarction, stroke, blood transfusion, retroperitoneal bleed, atrioventricular fistula, pseudoaneurysm, phrenic nerve injury, air embolism, or atrio-oesophageal fistula were observed.
Did not require transfusion.
Asymptomatic, did not require intervention.
Same patient. No other periprocedural complications (e.g. tamponade, haemorrhage) were found and the decompensation was attributed to complications of infiltrative cardiomyopathy.
Obstructive sleep apnea patients (with or without prior ablation) demonstrated numerically lower rates of 1-year arrhythmia free survival compared with those without OSA (50% vs. 61%) after FIRM ablation with pulmonary vein isolation, although this difference was not statistically significant (P = 0.72). For comparison, the 1-year arrhythmia free survival rate for patients without OSA who underwent conventional AF ablation without FIRM during the same time period (1:1 matched based on age, sex, AF subtype, history of HF, and previous ablations) was 80%.
Discussion
This study demonstrates a number of key findings regarding the association between OSA, AF drivers, and rotor location and density in symptomatic AF. First, we observed that patients with OSA had a higher overall rotor burden compared with patients without OSA, suggesting a novel mechanism underling the known association between OSA and AF. Second, we observed that the increase in rotor burden among OSA patients is driven primarily by an increase in RA rotors. Third, we found that among OSA patients, CPAP use was associated with a lower number of RA rotors. Finally, OSA patients were less likely to experience an acute intraprocedural endpoint with FIRM (AF cycle length slowing, AF organization, or AF termination) and were more likely to experience an atrial arrhythmia recurrence.
The association between OSA and AF has been well described, although the exact mechanisms underpinning this association have not been fully elucidated. The relationship between OSA and AF is both complex and multifactorial, likely mediated by shared comorbidities (diabetes, obesity, and hypertension) and the impact of OSA on the atrial myocardium and its electrical properties. Obstructive sleep apnea may contribute to AF via hypoxaemia, hypercapnia, fibrosis, and surges in sympathetic and vagal tone, increased inflammation.1 These insults are associated with changes in the anatomic AF substrate evidenced by increased atrial fibrosis (as assessed by bipolar voltage), shortening of the effective refractory period, and slower atrial conduction velocities.14 Obstructive sleep apnea is additionally associated with an increase in arrhythmias associated with a predilection towards AF, including premature atrial contractions, sinus bradycardia, and sinus pauses.15 Importantly, this study provides mechanistic insights into the link between OSA and AF by, for the first time, identifying increased rotor prevalence among OSA patients. This increased rotor burden may be a strong driver of the established association between OSA and risk of AF, AF progression, and AF recurrence.
In our study, the increased rotor burden among OSA patients was driven predominantly by an increase in RA rotors. Although AF is often thought to be a disease of the LA, it is well established that non-PV triggers16 and rotors9–11 may be present in the RA, although both are more common in the LA. The preferential increase in RA rotors among OSA patients is physiologically plausible based on the known association between OSA and pulmonary hypertension17 and the fact that each apneic event results in the generation of a very negative intrathoracic pressure which augments venous return and predisposes to nocturnal right sided volume overload.17 Our study demonstrated that OSA patients had increased RA pressures compared with patients without OSA, suggesting that RA hypertension may, at least in part, mediate the association between OSA and rotor burden. It appears that the RA may be more actively involved in the mechanism of AF maintenance among OSA patients; this may be an important factor in the higher rates of AF recurrence after catheter ablation for AF,4 particularly among those not treated with CPAP,3 because AF ablation often includes LA ablation only. It is possible that FIRM guided rotor ablation may be an important consideration among OSA patients as it would allow for a targeted, ‘as needed’ ablation strategy in the RA. Alternatively, outside of FIRM, these findings also raise the hypothesis that other ablation in the RA, such as cavotricuspid isthmus or crista terminals ablation or superior vena cava isolation, might lead to improved outcomes in patients with OSA.
We observed lower rates of acute procedural success and higher rates of AF recurrence among OSA patients compared with non-OSA patients, although these findings did not meet statistical significance due to small numbers and limited statistical power. These observations are consistent with prior reports.3,4 Although it does appear that increased rotor burden may represent a particularly important mechanism for AF maintenance in OSA, other complex mechanisms are likely at play, as evidenced by the fact that the ablation of these ‘excess’ rotors does not appear to lead to similar outcomes among OSA and non-OSA patients.
There is continued debate over the existence of rotors as a mechanism for AF maintenance and their value as a target for AF treatment.18 Several studies of AF mapping have not identified stable rotational activity during AF,19–21 suggesting that the maintenance of AF may rely on other mechanisms, including innumerable transient instances of re-entry that are both meandering and unstable (‘multiple wavelet hypothesis’).18 Although several clinical studies have caused the electrophysiology community to question the overall efficacy of the FIRM system itself,20,22 several studies using this10,11,23 and other mapping techniques7,8,24 strongly support the importance of rotors for the maintenance of AF. These studies suggest that with improved mapping technologies, rotor ablation may become an important tool for the treatment of AF, particularly the non-paroxysmal subtypes.
Limitations
This exploratory study has a number of key limitations worth mentioning. Our study is small, potentially limiting power to detect small differences in patient characteristics, procedure characteristics, and long-term study outcomes, and precluding the use of multivariable analyses. The retrospective nature of this study may make it prone to certain types of biases and errors in data collection that are less common among prospective studies. Sinus rhythm voltage maps are not routinely obtained prior to AF ablation, precluding analyses assessing the association between OSA, atrial scar, and rotor burden. We did not have access to most sleep study results and thus were unable to correlate OSA severity and rotor characteristics. It is well known that the 64-pole basket catheters used for FIRM mapping may be unable to detect all rotors due to poor tissue contact, particularly in the setting of extreme atrial dilation and abnormal atrial morphology.20 This study includes patients from a single centre, potentially limiting generalizability. Atrial fibrillation recurrence was defined as a clinical recurrence and was not determined based on routine ambulatory ECG monitoring. Finally, although AF recurrence was more common among OSA patients, we are unable to determine whether these patients have superior outcomes compared with similar patients who underwent AF ablation without FIRM guided rotor ablation.
Conclusion
Among consecutive patients undergoing FIRM guided AF ablation, OSA patients demonstrated increased rotor prevalence, driven predominantly by an increase in RA rotors, suggesting a novel mechanism underpinning the relationship between OSA and AF. Obstructive sleep apnea patients were less likely to meet an acute intraprocedural endpoint and more likely to have AF recurrence compared with patients without OSA.
Funding
D.J.F. is funded by the National Institutes of Health T 32 training grant HL069749.
Conflict of interest: J.P.D: consulting fees/honoraria from Medtronic, St. Jude Medical, Boston Scientific, Sorin Group, and CardioFocus’ research grants from Boston Scientific, Biosense Webster, Medtronic, and Gilead Sciences. T.D.B.: research grants: Medtronic; St Jude Medical; Consultant to Boehringer Ingelheim, ChanRX, Sequel Pharma, and Sanofi-Aventis. J.P.P.: research grants for clinical research from ARCA biopharma, Boston Scientific, Gilead, Janssen Pharmaceuticals, ResMed, and St Jude Medical and serves as a consultant to Amgen, Allergan, Johnson & Johnson, Medtronic, Parexel, and Spectranetics.