Issue Navigator

Volume 10 No. 05
Earn CME
Accepted Papers

Scientific Investigations

Severity of Obstructive Sleep Apnea Influences the Effect of Genotype on Response to Anti-Arrhythmic Drug Therapy for Atrial Fibrillation

Sandeep K. Goyal, M.D.1; Li Wang, M.S.2; Raghu Upender, M.D.3; Dawood Darbar, M.D.1,4; Ken Monahan, M.D.1
1Division of Cardiovascular Medicine; 2Department of Biostatistics; 3Division of Sleep Medicine/Department of Neurology; 4Division of Clinical Pharmacology, Vanderbilt Medical Center, Nashville, TN


Study Objectives:

To examine the impact of genotype on the relationship between obstructive sleep apnea (OSA) and anti-arrhythmic drug (AAD) efficacy in atrial fibrillation (AF).


Registry based.




Eighty-four individuals from Vanderbilt AF registry who had polysomnography, genotyping, and serial comprehensive evaluations of AF status.



Measurements and Results:

Response to AADs was defined as a decrease in AF burden score by ≥ 75% or the combination of sinus rhythm on follow-up EKGs, stable AAD therapy for at least 6 months, objective AF burden below an established threshold, and the absence of non-pharmacologic therapies. Participants were genotyped for common AF susceptibility alleles at chromosomes 4q25 (near PITX2), 16q22 (in ZFHX3), and 1q21 (in KCNN3), and common SNPs in the β1-adrenergic receptor (ARDB1). Wild-type status for rs10033464 at 4q25 was associated with increased success of AAD therapy in patients with no or mild OSA (odds ratio: 10.0, 95% confidence interval: 1.03 to 97.5; p < 0.05), but did not influence response to AAD therapy in those with moderate-severe OSA. A similar trend was observed for rs1801252 on ARDB1.


In this hypothesis-generating pilot study of predominantly Caucasian men, the effect on AF response to AAD therapy of rs10033464 at 4q25 varied based on OSA status. The impact of genotype on AAD efficacy may be greatest in mild OSA and attenuated in more severe disease.


Goyal SK, Wang L, Upender R, Darbar D, Monahan K. Severity of obstructive sleep apnea influences the effect of genotype on response to anti-arrhythmic drug therapy for atrial fibrillation. J Clin Sleep Med 2014;10(5):503-507.

Obstructive sleep apnea (OSA) has been increasingly recognized as a risk factor for atrial fibrillation (AF) and has an adverse impact on therapy for AF. Individuals with OSA have a higher prevalence of AF than those without OSA, and OSA is more prevalent in those with AF than those without AF.1,2 Paroxysms of AF are more common during and shortly after OSA-related respiratory events compared to periods of normal breathing.3 Anti-arrhythmic drugs (AADs) are less effective in achieving rhythm control when severe OSA, as compared to milder disease, is present; similarly, electrical cardioversion and catheter ablation are less successful in those with OSA than in unaffected individuals.46

Recent genome-wide association studies have identified loci on chromosomes 4q25 (near PITX2),7 16q22 (in ZFHX3),8,9 and 1q21 (in KCNN3),10 that associate with either typical or lone AF. Wild-type status for rs10033464 at 4q25 is a predictor of successful treatment of AF with AADs,11 and variants in rs10033464 and rs2200733 at 4q25 increase the chance of AF recurrence after catheter ablation.12,13 In addition, a common β-1 adrenergic receptor (ADRB1) polymorphism at rs1801253 has been shown to predict response to rate-control therapy in AF.14 These findings indicate that variable mechanisms contribute to AF susceptibility and suggest that response to therapy is also influenced by genotype.


Current Knowledge/Study Rationale: Individuals with OSA have a higher prevalence of AF and both genetics and OSA have been individually known to affect response to AAD therapy. Given that both OSA and common variants influence the response of AF to AAD therapy, it is plausible that the impact of OSA on AF is determined in part by genotype.

Study Impact: This study points toward the role of genetics in determination of anti-arrhythmic drug efficacy in patients with AF and OSA. This will help direct further focused research on defining exact extent of genetic influence in these patients.

Given that both OSA and common variants influence the response of AF to AAD therapy, it is plausible that the impact of OSA on AF is determined in part by genotype. Understanding the effects of common variants and OSA on the response to pharmacologic treatment of AF could influence AAD choice and may also factor into decisions to pursue nonmedication-based treatment strategies, such as pulmonary vein isolation or AV nodal ablation/pacemaker placement. Therefore, we sought to explore the relationship between common AF single nucleotide polymorphisms (SNPs) and symptomatic response to AADs in a cohort with AF and OSA. Specifically, we hypothesized that common genetic variants associated with AF decrease the efficacy of AADs to a greater degree in moderate or severe OSA as compared to no or mild OSA.


Study Cohort and Definitions

The Vanderbilt University Medical Center institutional review board approved the study protocol. The cohort used for this study and the techniques for measuring AF symptom burden have been previously described.4,15 Briefly, adults with documented AF treated with at least 1 conventional AAD were prospectively enrolled in the Vanderbilt AF Registry between November 2002 and October 2005. At enrollment and at 3, 6, and 12 months of follow-up, patients completed the modified University of Toronto AF Severity Scale (range 3-30) to gauge symptomatic AF burden.15 Patients' electronic medical records were also reviewed, and an electrophysiologist blinded to genotype status of the patients ascertained AF burden during enrollment.

Response to AAD Therapy

Response to AADs was defined as a decrease in AF burden score by ≥ 75% on a validated symptom scoring tool16 or the combination of sinus rhythm on follow-up EKGs, stable AAD therapy for ≥ 6 months, AF burden below the threshold for increased stroke risk described in the ASSERT trial17 on device diagnostics or ambulatory monitoring, and the absence of nonpharmacologic therapies (cardioversion, nodal ablation, AF ablation). In all participants, EKGs were obtained at enrollment and at the 1-, 3-, and 6-month follow-up visits. In the first 6 months of enrollment, one-third of participants had an assessment of AF burden by either device interrogation or ambulatory monitoring.

Polysomnography Data

The Vanderbilt AF registry was screened for participants who underwent a diagnostic overnight polysomnogram (PSG) for clinical reasons at the Vanderbilt Sleep Center.18 Sleep parameters were abstracted from clinical reports generated by specialists at our institution certified by the American Board of Sleep Medicine. If the study was a split-night examination, data from the diagnostic portion (without application of continuous or bilevel positive airway pressure) were utilized.

Full-night PSGs were carried out using the Polysmith Sleep system (Nihon Kohden America Incorporated; Foothill Ranch, CA). Airflow was monitored using both a thermistor placed over the philtrum and a nasal pressure transducer via oro-nasal cannula. Respiratory effort was monitored by impedance plethysmography. Electroencephalogram, electro-oculogram, and submental electromyogram were recorded according to American Academy of Sleep Medicine standards. Single lead electrocardiographic signals, as well as oxyhemoglobin saturation via digital pulse oximetry, were continuously recorded throughout the study. Apnea was defined as ≥ 90% decrease in the airflow signal from baseline for ≥ 10 seconds. For all patients, hypopneas were defined as ≥ 50% reduction in airflow from baseline for ≥ 10 sec accompanied by a decrease in oxyhemoglobin saturation of ≥ 3% or an associated EEG arousal.19

This definition of hypopneas was used to calculate the “standard AHI” for the entire cohort. A separate analysis was also performed for Medicare patients (n = 44), in which hypopneas were defined as ≥ 30% reduction in airflow from baseline for ≥ 10 sec associated with ≥ 4% desaturation. This definition of hypopnea was used to calculate for Medicare patients a respiratory disturbance index (RDI) that included apneas, hypopneas, and respiratory effort-related arousals (flow limitation that resolved with an arousal). The standard AHI and RDI were similar numerically; use of the RDI would have resulted in re-classification of 2 Medicare participants from “none” to “mild” and 1 Medicare participant from “mild” to “moderate.” Therefore, the standard AHI was used for all participants. Individuals with an apnea-hypopnea index (AHI) (sum of apneas and hypopneas divided by the total sleep time) ≥ 5 events/h were considered to have OSA. Severity of OSA was further categorized by AHI as follows: mild: 5-15 events/h; moderate: 16-30 events/h; severe: > 30 events/h.


Genomic DNA was isolated from whole blood by a commercial kit (Purgene; Gentra Systems; Minneapolis, MN). Genotyping of the 4q25 SNPs (rs2200733 and rs10033464) was performed using real-time polymerase chain reaction (PCR), iPlex single base primer extension, and MALDI-TOF mass spectrometry in a 384-well-format (Sequenom; San Diego, CA) as previously described.20 Genotyping at the 16q22 SNP (rs7193343 in ZFHX3), the 1q21 SNP (rs13376333 in KCNN3), and ADRB1 (rs1801252; rs1801253) was performed using TaqMan assays (Applied Biosystems Incorporated; Foster City, CA) as previously described8,10 by laboratory personnel who had no knowledge of the response to rhythm control therapy.

Statistical Analysis

Clinical data are reported as mean ± standard deviation for continuous variables and as percentages for categorical variables. Differences between groups were evaluated by the Wilcoxon rank-sum test for continuous variables and with Pearson χ2 test for categorical variables.

Pearson χ2 test was also used to test for Hardy-Weinberg equilibrium. Logistical regression was used to assess the association between rhythm control and genotype, adjusting for OSA status; an interaction term between genotype and OSA status was also included. Due to expected low numbers of homozygous variants, it was defined a priori to include homozygous variants of all SNPs with their respective heterozygous variants.

Analyses were performed with the R software package (version 2.12.0; Vienna, Austria). Comparisons were considered statistically significant if the p value was less than 0.05 for two-tailed tests.


The analysis consisted of 84 individuals who underwent polysomnography and had genotype data as well as serial evaluations of AF symptoms. This cohort includes the group described in our prior work,4 as well an additional ∼ 25 individuals identified in the interim. As such, this study population is very similar to the previously reported cohort in terms of demographics, cardiac history, echocardiographic characteristics, and key sleep parameters (Table 1). Briefly, the study population was primarily comprised of obese, middle-aged, Caucasian males. More than half had paroxysmal AF. Severe OSA was present in 52% of the cohort, with a mean AHI at the lower end of the severe range. Approximately 15% of those with at least mild OSA (12/79) were on continuous positive airway pressure (CPAP) therapy at the time of enrollment in the AF Registry. Three individuals initiated CPAP during the first 6 months after enrollment, and 2 participants had their CPAP therapy intensified (i.e., an up-titration of mean PAP) over that time frame. Two-thirds were taking either a β-blocker and/or calcium channel blocker; 50% were taking amiodarone; 25% were taking sotalol; and 25% were taking either flecainide or propafenone. Approximately half of the cohort (49%) responded to AADs according to the definition used in this study. As in our prior study,4 when compared to responders to AADs, non-responders had higher AHIs, greater likelihood of severe OSA, and a higher prevalence of non-paroxysmal AF. There was no difference in left atrial size in non-responders as compared to responders.

Cohort characteristics stratified by response to anti-arrhythmic drugs


table icon
Table 1

Cohort characteristics stratified by response to anti-arrhythmic drugs

(more ...)

The availability of genotype status was as follows: 4q25. rs2200733: 99% (83/84); 4q25.rs10033464: 95% (80/84); ZFXH3.rs7193343: 99% (83/84); KCNN3.rs13376333: 97% (81/84); ADRB1.rs1801252: 98% (82/84); ADRB1.rs1801253: 99% (83/84). The wild-type frequency across the SNPs of interest was between 40% and 75% (4q25.rs2200733: 65%; 4q25.rs10033464: 73%; ZFXH3.rs7193343: 67%; KCNN3. rs13376333: 44%; ADRB1.rs1801252: 68%; ADRB1. rs1801253: 53%).

Carriers of the 4q25.rs10033464 wild-type allele were more likely to respond to AAD therapy if they had no or mild OSA (odds ratio [OR]: 10.0, 95% confidence interval [CI]: 1.03 to 97.5; p < 0.05), but this genotype did not influence the success of AAD response in participants with moderate or severe OSA (OR: 0.86, 95% CI: 0.26 to 3.20). Those with the ADRB1. rs1801252 wild-type allele showed a trend towards more successful response to AAD therapy in the presence of no or mild OSA; again, there was no effect in those with moderate or severe OSA. The response to AADs stratified by OSA status for both 4q25.rs10033464 and ADRB1.rs1801252 is displayed in Figure 1. For comparison, when stratified in a similar manner by OSA level and genotype status, there was no difference in the prevalence of a given clinical AF subtype (i.e., paroxysmal, persistent, permanent) between any subgroups.

Response to anti-arrhythmic drugs (AADs) by genotype and OSA status.

Responder rates to AAD therapy are displayed as a function of OSA status for 4q25.rs10033464 (A), and ADRB1.rs1801252 (B). In those with no or mild OSA (n = 22), the prevalence of responders is higher in the wild-type (WT) than in minor allele carriers (MAC) for both SNPs. For moderate or severe OSA (n = 59), the responder rate does not vary based on WT vs. MAC status for either SNP.


Figure 1

Response to anti-arrhythmic drugs (AADs) by genotype and OSA status. Responder rates to AAD therapy are displayed as a function of OSA status for 4q25.rs10033464 (A), and ADRB1.rs1801252 (B). In those with no or mild OSA (n = 22), the prevalence of responders is higher in the wild-type (WT) than in...

(more ...)

For the KCNN3, ZFXH3, 4q25.rs2200733 and ADRB1. rs1801253 variants, there was no difference in AAD response rate when stratified by OSA status.


The key finding of this study suggests that the effect of genotype on the response of AF to AAD therapy depends partially on OSA severity. The high non-response rates in more severe OSA,4 regardless of genotype, suggest that the impact of common variants on response to AADs may be “overwhelmed” by the increased frequency of apneic events and their pro-arrhythmic consequences.3 However, at lower levels of OSA, a favorable genotype may be able to “protect” against arrhythmogenesis given the less frequent proarrhythmic stimuli.

Polymorphisms in ARDB1 are associated with differences in resting hemodynamics and response to positive airway pressure therapy in those with OSA21,22 and also track with a higher prevalence of OSA in hypertensive individuals.23 Our findings provide additional support for a link between OSA and ADRB1 polymorphisms. One such connection could be increased downstream changes in sympathetic tone induced by OSA-related respiratory events, which, in turn, could contribute to a more pro-arrhythmic milieu. Correspondingly, although perhaps even more speculative, it appears plausible that PITX2, the closest gene to rs10033464 at 4q25 and a regulator of development of pulmonary venous myocardium, could influence susceptibility to AF based on OSA-related local fluctuations in intrathoracic pressure, oxygen tension, and sympathetic tone.24 These factors may contribute to electroanatomical remodeling in the atria, which has been observed in those with OSA compared to unaffected individuals.25 Whether a synergistic effect of common genetic variants and OSA on resistance of AF to treatment may contribute to the observed decrease in procedural success of electrical cardioversion5 or catheter ablation6 for AF when severe OSA is present (and increased success when OSA is treated26) remains to be determined.

The current results extend recent work from our group that established rs10033464 status as a predictor of AF response to AAD therapy11 and an ADRB1 receptor polymorphism as a modulator of response to a rate-control strategy.14 In addition to using genotype as a guide when selecting AF treatment, consideration of OSA status when formulating therapeutic plans may be reasonable as well.


Our cohort includes detailed longitudinal follow up of AF burden and symptoms, availability of sleep-lab overnight PSGs, and detailed genotyping data. Changes in OSA therapy (i.e., initiation or titration of CPAP) over the course of the study were infrequent, suggesting that effects of OSA treatment on AAD efficacy did not unduly influence the results.

The study population primarily consisted of a Caucasian male population; therefore, the results may not be generalizable to other groups. A large percentage of our cohort had severe OSA, which may be expected in a population with numerous risk factors for OSA referred for PSG. AAD therapy was recorded at the time of enrollment into the AF Registry, and changes in pharmacotherapy over time were noted; however, detailed information regarding the rationale behind AAD choices was not available. Variation in AAD use could influence the relationship between OSA, AF, and genotype, but disparate effects of AADs would be expected to dilute the main association; for the present cohort, the small sample size does not permit investigation of this hypothesis. The low rate of recorded CPAP use among the cohort may reflect a time lag between enrollment in the AF Registry, recognition of OSA, and initiation of therapy, which could have occurred (or have been recorded) outside the follow-up period examined in this study. However, the relative paucity of OSA therapy does minimize the potentially heterogenous and confounding effects of CPAP treatment on our analyses and permits evaluation of the “natural history” of OSA in this context. Furthermore, non-uniform treatment with CPAP of a larger portion of the cohort may have decreased the association between response to AAD therapy and OSA, thus diminishing the magnitude of our main finding.

The overall sample size is small, due to lack of PSG data in the majority of AF Registry patients; this is reflected in the wide confidence intervals associated with the main result. Although the cohort included only AF Registry participants who had a PSG, this “selection bias” should not confound the results. It is unlikely that the decision to obtain a PSG (made independent of the patient's participation in the AF Registry) was related to genotype or efficacy of AAD therapy for AF, although it is possible that some of the referrals for PSG were prompted by unsuccessful attempts at AF treatment. The analytical limitations stemming from the size of the cohort include: (1) pooling of AADs as a single group and (2) inability to perform multivariate regression. However, the emergence of a relationship between efficacy of AAD therapy for AF, OSA, and genotype, despite the small cohort, suggests it is reasonable to pursue larger studies of the impact on this relationship of individual AADs and other potential confounders, such as AF subtype (i.e., paroxysmal, persistent, permanent, lone, familial), choice of AAD, and OSA treatment.


The results of this hypothesis-generating pilot study raise the possibility that the impact of common AF susceptibility alleles on the efficacy of AAD therapy for AF may be influenced by OSA severity. Future research will focus on further characterizing the interaction of genotype and OSA status on AF pathogenesis and treatment as part of the effort to personalize therapy for this common and morbid arrhythmia.


This was not an industry-supported study. Research reported in this publication was supported by the National Center for Advancing Translational Sciences of the National Institute of Health under Award Number UL1 TR000445, HL65962, HL09221, and an American Heart Association Established Investigator Award (0940116N). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors have indicated no financial conflicts of interest.



obstructive sleep apnea


atrial fibrillation


anti-arrhythmic drug


single nucleotide polymorphisms




polymerase chain reaction


The authors wish to thank Ann Gage, M.D., for her assistance with data management, abstraction, and analysis and Arthur Walters, M.D., for his thoughtful review of the manuscript.



Mehra R, Benjamin EJ, Shahar E, et al., authors. Association of nocturnal arrhythmias with sleep-disordered breathing: The Sleep Heart Health Study. Am J Respir Crit Care Med. 2006;173:910–6. [PubMed Central][PubMed]


Gami AS, Pressman G, Caples SM, et al., authors. Association of atrial fibrillation and obstructive sleep apnea. Circulation. 2004;110:364–7. [PubMed]


Monahan K, Storfer-Isser A, Mehra R, et al., authors. Triggering of nocturnal arrhythmias by sleep-disordered breathing events. J Am Coll Cardiol. 2009;54:1797–804. [PubMed Central][PubMed]


Monahan K, Brewster J, Wang L, et al., authors. Relation of the severity of obstructive sleep apnea in response to anti-arrhythmic drugs in patients with atrial fibrillation or atrial flutter. Am J Cardiol. 2012;110:369–72. [PubMed Central][PubMed]


Kanagala R, Murali NS, Friedman PA, et al., authors. Obstructive sleep apnea and the recurrence of atrial fibrillation. Circulation. 2003;107:2589–94. [PubMed]


Ng CY, Liu T, Shehata M, Stevens S, Chugh SS, Wang X, authors. Meta-analysis of obstructive sleep apnea as predictor of atrial fibrillation recurrence after catheter ablation. Am J Cardiol. 2011;108:47–51. [PubMed]


Gudbjartsson DF, Arnar DO, Helgadottir A, et al., authors. Variants conferring risk of atrial fibrillation on chromosome 4q25. Nature. 2007;448:353–7. [PubMed]


Gudbjartsson DF, Holm H, Gretarsdottir S, et al., authors. A sequence variant in ZFHX3 on 16q22 associates with atrial fibrillation and ischemic stroke. Nat Genet. 2009;41:876–8. [PubMed Central][PubMed]


Benjamin EJ, Chen PS, Bild DE, et al., authors. Prevention of atrial fibrillation: report from a National Heart, Lung, and Blood Institute workshop. Circulation. 2009;119:606–18. [PubMed Central][PubMed]


Ellinor PT, Lunetta KL, Glazer NL, et al., authors. Common variants in KCNN3 are associated with lone atrial fibrillation. Nat Genet. 2010;42:240–4. [PubMed Central][PubMed]


Parvez B, Vaglio J, Rowan S, et al., authors. Symptomatic response to antiarrhythmic drug therapy is modulated by a common single nucleotide polymorphism in atrial fibrillation. J Am Coll Cardiol. 2012;60:539–45. [PubMed Central][PubMed]


Benjamin Shoemaker M, Muhammad R, Parvez B, et al., authors. Common atrial fibrillation risk alleles at 4q25 predict recurrence after catheter-based atrial fibrillation ablation. Heart Rhythm. 2013;10:394–400. [PubMed Central][PubMed]


Husser D, Adams V, Piorkowski C, Hindricks G, Bollmann A, authors. Chromosome 4q25 variants and atrial fibrillation recurrence after catheter ablation. J Am Coll Cardiol. 2010;55:747–53. [PubMed]


Parvez B, Chopra N, Rowan S, et al., authors. A common beta1-adrenergic receptor polymorphism predicts favorable response to rate-control therapy in atrial fibrillation. J Am Coll Cardiol. 2012;59:49–56. [PubMed Central][PubMed]


Dorian P, Paquette M, Newman D, et al., authors. Quality of life improves with treatment in the Canadian Trial of Atrial Fibrillation. Am Heart J. 2002;143:984–90. [PubMed]


Darbar D, Roden DM, authors. Symptomatic burden as an endpoint to evaluate interventions in patients with atrial fibrillation. Heart Rhythm. 2005;2:544–9. [PubMed]


Healey JS, Connolly SJ, Gold MR, et al., authors. Subclinical atrial fibrillation and the risk of stroke. N Engl J Med. 2012;366:120–9. [PubMed]


Chesson AL, Ferber RA, Fry JM, et al., authors. The indications for polysomnography and related procedures. Sleep. 1997;20:423–87. [PubMed]


Berry RB, Budhiraja R, Gottlieb DJ, et al., authors. Rules for scoring respiratory events in sleep: update of the 2007 AASM Manual for the Scoring of Sleep and Associated Events. Deliberations of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine. J Clin Sleep Med. 2012;8:597–619. [PubMed Central][PubMed]


Kääb S, Darbar D, van Noord C, et al., authors. Large scale replication and meta-analysis of variants on chromosome 4q25 associated with atrial fibrillation. Eur Heart J. 2009;30:813–9. [PubMed Central][PubMed]


Börgel J, Schulz T, Bartels NK, et al., authors. Modifying effects of the R389G beta1-adrenoceptor polymorphism on resting heart rate and blood pressure in patients with obstructive sleep apnoea. Clin Sci (Lond). 2006;110:117–23.


Goyal SK, Sharma A, authors. Atrial fibrillation in obstructive sleep apnea. World J Cardiol. 2013;5:157–63. [PubMed Central][PubMed]


Bengtsson Bostrom K, Hedner J, Grote L, et al., authors. Polymorphisms in alpha- and beta-adrenergic receptor genes, hypertension, and obstructive sleep apnea: The Skaraborg Sleep Study. Int J Hypertens. 2010;2010.


Somers VK, Dyken ME, Clary MP, Abboud FM, authors. Sympathetic neural mechanisms in obstructive sleep apnea. J Clin Invest. 1995;96:1897–904. [PubMed Central][PubMed]


Dimitri H, Ng M, Brooks AG, et al., authors. Atrial remodeling in obstructive sleep apnea: implications for atrial fibrillation. Heart Rhythm. 2012;9:321–7. [PubMed]


Fein AS, Shvilkin A, Shah D, et al., authors. Treatment of obstructive sleep apnea reduces the risk of atrial fibrillation recurrence after catheter ablation. J Am Coll Cardiol. 2013;62:300–5. [PubMed]