Author + information
- Received June 7, 2011
- Revision received August 7, 2011
- Accepted August 23, 2011
- Published online January 3, 2012.
- Babar Parvez, MD⁎,
- Nagesh Chopra, MD⁎,
- Shane Rowan, MD⁎,
- Joseph C. Vaglio, MD⁎,
- Raafia Muhammad, MD⁎,
- Dan M. Roden, MD⁎,† and
- Dawood Darbar, MD⁎,†,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Dawood Darbar, Division of Cardiovascular Medicine, Vanderbilt University School of Medicine, 2215B Garland Avenue, Room 1285A MRB IV, Nashville, Tennessee 37323-6602
Objectives In this study, we evaluated the impact of 2 common β1-adrenergic receptor (β1-AR) polymorphisms (G389R and S49G) in response to ventricular rate control therapy in patients with atrial fibrillation (AF).
Background Randomized studies have shown that ventricular rate control is an acceptable treatment strategy in patients with AF. However, identification of patients who will adequately respond to rate-control therapy remains a challenge.
Methods We studied 543 subjects (63% men; age 61.8 ± 14 years) prospectively enrolled in the Vanderbilt AF registry and managed with rate-control strategy. A “responder” displayed adequate ventricular rate control based on the AFFIRM (Atrial Fibrillation Follow-Up Investigation of Rhythm Management) criteria: average heart rate (HR) at rest ≤80 beats/min; and maximum HR during a 6-min walk test ≤110 beats/min or average HR during 24-h Holter ≤100 beats/min.
Results A total of 295 (54.3%) patients met the AFFIRM criteria. Baseline clinical characteristics were similar in responders and nonresponders except for mean resting HR (76 ± 20 beats/min vs. 70 ± 15 beats/min; p < 0.01) and smoking (6% vs. 1%; p < 0.01). Multiple clinical variables (age, gender, hypertension) failed to predict response to rate-control therapy. By contrast, carriers of Gly variant at 389 were more likely to respond favorably to rate-control therapy; 60% versus 51% in the Arg389Arg genotype, p = 0.04. This association persisted after correction for multiple clinical factors (odds ratio: 1.42, 95% confidence interval: 1.00 to 2.03, p < 0.05). Among responders, subjects carrying the Gly389 variant required the lowest doses of rate-control medications; atenolol: 92 mg versus 68 mg; carvedilol: 44 mg versus 20 mg; metoprolol: 80 mg versus 72 mg; diltiazem: 212 mg versus 180 mg, and verapamil: 276 mg versus 200 mg, respectively (p < 0.01 for all comparisons).
Conclusions We have identified a common β1-AR polymorphism, G389R, that is associated with adequate response to rate-control therapy in AF patients. Gly389 is a loss-of-function variant; consequently, for the same adrenergic stimulation, it produces reduced levels of adenyl cyclase, and hence, attenuates the β-adrenergic cascade. Mechanistically, the effect of rate-control drugs will be synergistic with that of the Gly389 variant, which could possibly explain our findings. These findings represent a step forward in the development of a long-term strategy of selecting treatment options in AF based on genotype.
Atrial fibrillation (AF) is the most common arrhythmia in clinical practice in the United States today, affecting over 2 million people (1). A common target for pharmacological therapies in cardiovascular diseases is the β1-adrenergic receptor (β1-AR) ADRB1, and β-blocking medications are considered first-line agents for ventricular rate control in patients with AF (2,3). The ADRB1 is a member of the superfamily of cell surface receptors that carry out signaling via coupling to guanine nucleotide binding proteins (G proteins), activated by catecholamine binding, which increases intracellular cyclic-adenosine monophosphate activity, resulting in increased chronotropic and inotropic effects on the myocardium.
There are 2 common nonsynonymous single nucleotide polymorphisms (SNPs) of the ADRB1 found in the humans. At position 49 in the amino-terminus of the receptor, a serine is substituted by a glycine (Ser49Gly) with an allele frequency of 12% to 16% in Caucasians and Asians and 13% to 28% in African Americans (4,5). At position 389 in the proximal portion of the carboxy-terminus in the cytoplasmic tail, an arginine is substituted by a glycine (Arg389Gly) with an allele frequency of 24% to 34% in Caucasians and Asians and 39% to 46% in African Americans (5).
In vitro studies, which largely have been conducted in mice or rodent ventricular myocytes, have demonstrated that both the Ser49Gly and Arg389Gly SNPs are functional; earlier studies showed that the Gly49 AR expresses higher basal and agonist-stimulated adenylyl cyclase activity and increased agonist-promoted down-regulation compared with the Ser49 variant, whereas the Arg389 allele has 3 times greater adenylyl cyclase activity in response to agonist than the Gly389 variant (5,6). However, recent studies have reported that the effect on adenyl cyclase activity is primarily due to allelic variants on position 389 and not 49, which may only play a modulating role yet to be fully determined (7). Studies have also assessed the effects of these SNPs on resting hemodynamics and incidence of hypertension (HTN). Bengtsson et al. (8) in a study of genotype-discordant siblings reported that siblings homozygous for the Arg389 allele had significantly higher heart rates (HRs) and diastolic blood pressure and were more likely to be hypertensive compared with siblings heterozygous or homozygous for the Gly389 allele. However, several other studies have met with mixed results (9,10).
Although the role of ADRB1 polymorphisms has been extensively investigated in HTN and heart failure, it is only recently that a study showed that the Arg389Gly polymorphism modulates responsiveness to β-blockade in heart failure patients (11). However, the role of these common ADRB1 polymorphisms in modulating response to atrioventricular (AV) nodal blocking drugs in AF patients has not been determined. Here, we tested the hypothesis that 2 common SNPs in the ADRB1 (Ser49Gly and Arg389Gly) modulate response to rate-control therapy in patients with AF.
The study was performed in patients prospectively enrolled in the Vanderbilt AF Registry, which is a clinical and a genetic registry (12). Inclusion criteria include age >18 years, a documented history of AF or atrial flutter, and attempted rate control using a β-blocker, calcium channel blocker, or digoxin. An echocardiogram was obtained on all patients at the time of enrollment into the registry. The treating physicians were blinded to the study protocol and subsequent genotype data. Study investigators were not involved in patient management. Written informed consent was obtained from all patients under a protocol approved by the Vanderbilt University Institutional Review Board.
Arterial HTN was defined by a history of HTN and/or the presence of antihypertensive therapy. Criteria for coronary artery disease included a history of myocardial infarction or typical angina, previous bypass surgery or angioplasty, and drug treatment. Congestive heart failure was defined by a history and/or drug treatment for heart failure. Left atrial and left ventricular measurements from the M-mode echocardiograms were made by an experienced physician blinded to the genotype status of the patient. The echocardiograms were evaluated according to the recommendations of the American Society of Echocardiography. Drug prescription distribution was defined as the total number of prescriptions written by the treating physicians for both responders and nonresponders, including alteration of dose of an already prescribed medication and/or addition of a new medication to control AF. Family history of AF was defined as electrocardiographically documented AF in 1 or more first-degree family members of the study subject.
Response to rate-control therapy
Responders were prospectively defined as patients who achieved adequate rate control, meeting the AFFIRM (Atrial Fibrillation Follow-up Investigation of Rhythm Management) study criteria: average HR ≤80 beats/min at rest and maximum HR ≤110 beats/min during a 6-min walk test or average HR during 24-h ambulatory Holter monitoring electrocardiogram ≤100 beats/min (at least 18 h of interpretable monitoring) and no HR >110% maximum predicted age-adjusted exercise HR (13). Nonresponders were patients who failed to meet the AFFIRM criteria necessitating addition of an anti-arrhythmic drug (AAD) or nonpharmacological therapies such as AV node ablation and pacemaker implantation in 6 months from study entry.
Determination of β1-AR genotypes
ADRB1 genotyping for the Arg389Gly and Ser49Gly variants was performed by laboratory personnel who had no knowledge of the response to rate-control therapy. Genomic deoxyribonucleic acid was isolated from whole blood by a commercial kit (Purgene; Gentra Systems, Minneapolis, Minnesota). Genotyping was performed for rs1801252 C>G (Ser49Gly) and rs1801253 A>T (Arg389Gly) using a TaqMan assay (Applied Biosystems, Foster City, California) as previously described (14).
All data are expressed as mean ± SD. Chi-square analysis or Fisher exact test was performed on discrete variables and Mann-Whitney U test on continuous variables. After applying quality control criteria to the genotypic data, Hardy-Weinberg equilibrium was assessed. Due to expected low numbers, homozygous variants (Gly49 and Gly389) were pooled with respective heterozygous variants to formulate a cumulative group of carriers of minor allele, which were defined a priori. Logistic regression with dominant, additive, and recessive modeling was performed in response to therapy and ADRB1 polymorphisms to determine the odds ratio (OR) and adjusted for age and gender. Statistical significance was defined as a 2-sided p < 0.05. Statistical analysis was performed with PLINK and PASW Statistics software (version 18.0.0, IBM, Armonk, New York).
Haplotypes were estimated using the 2 SNPs by applying standard expectation–maximization algorithms and measured for degree of linkage disequilibrium (LD) represented by r2. For drug–group distribution analysis, we considered atenolol, carvedilol, metoprolol, and propranolol from β-blockers (BBs), diltiazem and verapamil from calcium channel blockers (CCBs), and digoxin. Given the various types of BBs that can be used in the clinical setting, for drug–haplotype distribution among responders, we considered the most commonly prescribed BBs and CCBs in our study, defined a priori for the final analysis. All drug doses are expressed as milligrams per day representing the final total mean medication dose required for ventricular rate control of AF.
A total of 587 patients were enrolled over a 36-month period. Forty-four subjects were non-Caucasians (37 African Americans and 7 Asians) and hence excluded from the analysis.
Patient demographics and clinical characteristics
The clinical characteristics of the study population are listed in Table 1. Our study population included 543 Caucasian patients; 344 men and 199 women. The mean age of the cohort was 61.8 ± 14 years. Two hundred ninety-five patients (54%) met the AFFIRM criteria and were classified as responders. The final drug prescription distribution was 567 in responders and 595 (310 AADs and 285 rate-control drugs) in nonresponders. In responders, drug class distribution was 61%, 31%, and 8% for BBs, CCBs, and digoxin, respectively, compared with 43%, 28%, and 29%, respectively, in nonresponders. Thirty-one patients (12.5%) in the nonresponder group underwent AV node ablation and pacemaker implantation for treatment of AF, and ∼ 51% of nonresponders required concurrent administration of an AAD (amiodarone: 35%, sotalol: 33%, propafenone: 15%, flecainide: 15%, and procainamide: 2%). In responders, combination treatment was required with BBs and CCBs in 50% of cases, CCBs and digoxin in 25% of cases, BBs and digoxin in 13% of cases compared with BBs and CCBs in 65% of cases, BBs and digoxin in 67% of cases, respectively, of nonresponders. Only 35 (12%) patients among responders were adequately rate controlled with BBs alone versus 17 (7%) patients from nonresponders who were also on at least 1 concurrent AAD. This precluded us from undertaking a subanalysis looking at the modulatory effect of ADRB1 polymorphism on patients taking BBs only.
In our cohort, we observed a similar trend in HR as previously reported, with homozygotes for the Arg389 allele having the highest HR (75 ± 19 beats/min), followed by Arg389Gly (74 ± 20 beats/min) and homozygotes for Gly389 variant (71 ± 15 beats/min).
Our genotype call rate was 498 of 543 (92%) at position 49 and 513 of 543 (95%) at position 389. The genotype frequencies of ADRB1 polymorphisms in our cohort did not deviate significantly from Hardy-Weinberg equilibrium. Arg389Gly minor allele frequency: 0.28 (responders) versus 0.26 (nonresponders), p > 0.5, Ser49Gly minor allele frequency: 0.14 (responders) versus 0.15 (nonresponders), p > 0.5. The haplotype frequency in our cohort was Arg389Arg-Ser49Ser (37%), Arg389Arg-Ser49Gly (15%), Arg389Arg-Gly49Gly (2%), Arg389Gly-Ser49Ser (30%), Arg389Gly-Ser49Gly (10%), Arg389Gly-Gly49Gly (0%), Gly389Gly-Ser49Ser (6%), Gly389Gly-Ser49Gly (0%), and Gly389Gly-Gly49Gly (0%). We observed weak LD between the 2 SNP, with r2 = 0.045. The clinical characteristics of responders and nonresponders by genotype Ser49Ser versus Gly49 carriers and Arg389Arg versus Gly389 carriers are listed in Tables 2 and 3.⇓⇓
Overall, 13% of our study population had a positive family history of AF. Analysis did not reveal any correlation between family history and polymorphism at either position 49 (Ser49Ser [13%] vs. Gly49 carriers [16%], chi-square p = 0.315) or position 389 (Arg389Arg [15%] vs. Gly389 carriers [11%], chi-square p = 0.214).
Response to rate-control therapy
Two hundred ninety-five (54%) patients responded adequately to rate-control therapy. Carriers of the Gly variant at 389 were more likely to respond favorably to rate-control therapy; 60% responders in carriers of minor allele group versus 51% in Arg389Arg genotype, chi-square p = 0.04. Polymorphism at position 49 did not influence response to rate-control therapy; 52% responders in carriers of minor allele group versus 55% in Ser49Ser genotype; chi-square p = 0.45 (Fig. 1). In regression analysis, multiple clinical variables (age, HTN, gender) failed to significantly predict adequate response to rate-control therapy. By contrast, dominant model (identical effect is expected in heterozygous and homozygous variant carriers), showed a significant association of adequate rate control with the Gly389 variant allele (OR: 1.44, 95% confidence interval [CI]: 1.01 to 2.04, p < 0.05). This association persisted after correction for age and gender (OR: 1.42, 95% CI: 1.00 to 2.03, p < 0.05). Polymorphism at position 49 was not significantly associated with adequate rate control under all 3 genetic models (Table 4). Haplotype-response association analysis also confirmed that the Gly variant at position 389 significantly favors adequate rate-control therapy. Since Ser49Ser did not influence outcome of therapy as a haplotype with Arg389Arg, we think the effect observed with Arg389Gly-Ser49Ser haplotype is primarily driven by the Gly389 variant (Fig. 2).
Among responders, patients with Arg389Arg-Ser49Ser haplotype required the highest doses of BBs and CCBs to achieve adequate rate control than patients with Arg389Gly-Ser49Ser haplotype who required the lowest doses; atenolol: 92 mg versus 68 mg; carvedilol: 44 mg versus 20 mg; metoprolol: 80 mg versus 72 mg; diltiazem: 212 mg versus 180 mg, and verapamil: 276 mg versus 200 mg, respectively (p < 0.01 for all comparisons). It is noteworthy that there was no statistical difference in the number of medications prescribed among these haplotypes (as shown in Table 5), suggesting the difference in doses to be truly related to genotypes. Similarly, in categorizing drug prescription distribution by genotype at position 389 (137 [Arg389Arg] vs. 129 [Arg389Gly]), there was no statistical difference in the number of drugs prescribed for each genotype group: atenolol (22 [Arg389Arg] vs. 26 [Arg389Gly]); carvedilol (13 [Arg389Arg] vs. 11 [Arg389Gly]); metoprolol (47 [Arg389Arg] vs. 46 [Arg389Gly]); digoxin (13 [Arg389Arg] vs. 8 [Arg389Gly]); diltiazem (41 [Arg389Arg] vs. 45 [Arg389Gly]); and verapamil (14 [Arg389Arg] vs. 13 [Arg389Gly]) (p > 0.05 for all comparisons).
Our study is the first to our knowledge to demonstrate that a common ADRB1 polymorphism Arg389Gly predicts a favorable response to rate-control therapy in AF patients. Gly389 variant is present in 24% to 46% of humans across different races, making our finding especially significant with respect to its clinical implications.
AV node physiology and distribution of ion channels, especially L-type calcium channels, and ADRB1 alters greatly from ventricular myocytes, which have been the primary source of in vitro functional studies of ADRB1. The activity of L-type calcium channels in the AV nodal N cells determines the velocity of depolarization during the upstroke of an action potential. These calcium channels are further augmented by catecholamine activation of ADRB1-mediated channel phosphorylation through enhanced cAMP production (15,16). This mechanism explains why the majority of patients with AF required concurrent administration of BBs and CCBs for a synergistic effect for rate control.
Gly389 is a loss-of-function variant; consequently, for the same adrenergic stimulation, it produces reduced levels of adenyl cyclase and hence attenuates the β-adrenergic cascade and its downstream effect in cardiomyocytes, reducing gain in excitation-contraction coupling with catecholaminergic surge (17). Furthermore, the predicted effect of the Arg389Gly variant would be to slow conduction and increase refractoriness in the AV node, thus reducing ventricular rates in AF. Mechanistically, the effect of BBs and CCBs on the conduction and refractoriness of the AV node will be synergistic with that of the Arg389Gly variant, which could possibly explain our findings (Fig. 3).
In our study, only 13% of patients required concurrent use of digoxin with BBs. Digoxin slows AV node conduction and prolongs the effective refractory period. These actions on the AV node are secondary to its effect on increasing vagal tone and release of acetylcholine from the parasympathetic nerve terminals in the AV node, However, there are no studies to our knowledge that have looked at β1-AR polymorphisms and digoxin in patients with AF or heart failure.
Studies have reported weak to strong LD between these 2 ADRB1 SNPs, making it difficult to relate ex vivo with in vivo findings (18,19). In our cohort, we observed weak LD between the 2 SNPs. Furthermore, our results show that AF patients who harbored the functional variant Arg389Gly had a significantly favorable response to rate-control therapy. Taken together, this suggests that of the 2 SNPs, Arg389Gly drives outcome to rate-control therapy in patients with AF.
Clinical studies examining the influence of ADRB1 polymorphism on BB response in patients with CHF have met with variable results. Mialet Perez et al. (20) found substantial improvement in left ventricular ejection fraction with a standardized-dose regimen of carvedilol in Arg389 homozygous patients. However, in a study by Chen et al. (21), similar improvement was observed in Gly389 homozygous patients. Likewise, in CHF patients, Arg389 homozygotes, but not Gly389 carriers, treated with bucindolol had a significant reduction in mortality and morbidity when compared with placebo (22). In a substudy of MERIT-HF (Metoprolol CR/XL Randomized Intervention Trial in Congestive Heart Failure) in which patients were treated with metoprolol, no significant difference in all-cause mortality or hospitalization was observed in patients homozygous for either the Arg389 or Gly389 variants (23). The allelic distribution of ADRB1 polymorphism at codon 49 was found to be associated with long-term survival of patients with chronic HF: 46% mortality rate in Ser49 carriers versus 23% in carriers of the Gly49 variant (24). However, other studies have not been able to reproduce these findings (25). In our study, Ser49Gly polymorphism did not influence the outcome of rate-control therapy in patients with AF.
Recent randomized trials of rhythm versus rate-control strategies of treatment in patients with AF suggest that rate control is a viable first-line strategy in many patients (2,26). It can improve symptoms, exercise capacity, and cardiac function, however, identification of patients likely to respond to this approach is challenging (26–28). Currently, there is no widely accepted method of predicting who will respond adequately to rate-control therapy and who will require escalation and/or change in therapy. Certainly, there are no studies that have assessed genetic predictors of rate-control therapy in AF patients. Rate control of AF can be challenging, and often drugs have to be changed and combination therapy is needed to achieve adequate rate control (13).
Our data suggest a potential role of screening for a common ADRB1 polymorphism Arg389Gly in AF patients who will favorably respond to rate-control therapy. Furthermore, requirement of higher doses of BBs and CCBs for rate control in AF rate leading to clinical adverse effects is also commonly encountered. In our study population, metoprolol among BBs and diltiazem among CCBs were the most commonly prescribed rate-control agents for the treatment of AF, and subjects with the Arg389Arg genotype consistently required higher doses, which may increase the possibility of potential adverse effects. This also suggests a role for genotyping patients for common variants in ADRB1 to predict requirement for higher doses of rate-control medications. However, further studies are needed to validate our finding before clinical use is recommended.
To our knowledge, this is the first and the largest study to date looking at the effects of ADRB1 polymorphism on rate-control therapy in patients with AF. However, there are certain limitations to consider. First, we cannot correct for daily alterations in HR (conditions such as exercise, stress, sleep) that are driven by catecholamine-mediated activation of AR, but considering that this variability exists in all our patients, the effect would be spread evenly and would not be considered confounding. Second, by selecting the AFFIRM criteria for adequate rate control, we were able to monitor alterations in HRs of our patients with either a 6-min walk test or 24-h Holter monitoring. Third, the study population is derived from patients presenting to a large, tertiary referral center and may not represent the population with AF in other settings; nonetheless, the frequency of ADRB1 alleles identified in this population corresponds to previously published frequencies. Finally, it is uncertain whether the current study's findings can be generalized to non-Caucasian populations due to considerable differences in minor allele frequencies and exclusion from the analysis due to small sample size (n = 44) (4).
There was no standard protocol for titration of rate-control medications or changes to a rhythm-control strategy. The management of study patients was left at the discretion of the treating physicians who were blinded to the study protocol and genotype data. Although this lack of standardization inevitably introduces heterogeneity to the endpoints considered, it reflects a “real world” practice environment and increases the applicability of these results to clinical daily practice. Currently, there are several types of BBs that can be used for rate-control therapy in patients with AF. In our study, we had defined a priori to consider the most commonly prescribed BBs and CCBs in the final drug–haplotype analysis, thus providing a detailed and well-powered analysis.
We have shown that a common ADRB1 polymorphism Arg389Gly predicts favorable response to rate-control therapy in AF patients. Being a loss-of-function variant that decreases conduction and increases refractoriness in the AV node, it is physiologically sound to consider that it works synergistically with BBs and CCBs. Confirmatory studies are needed before clinical use of our data is recommended. These findings represent a step forward in the development of a long-term strategy of selecting treatment options in AF based on genotypes.
This work was supported in part by National Institutes of Health grants HL065962, HL092217, and an American Heart Association Established Investigator Award (0940116N) to Dr. Darbar. Dr. Rowan has received fellowship support from Boston Scientific, Medtronic, and St. Jude; and speaker fees from Medtronic. Dr. Roden is a consultant for Merck, Astellas, Watner Chilcott, and Sanofi. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- antiarrhythmic drug
- atrial fibrillation
- adrenergic receptor
- calcium channel blocker
- congestive heart failure
- confidence interval
- heart rate
- linkage disequilibrium
- odds ratio
- single nucleotide polymorphism
- Received June 7, 2011.
- Revision received August 7, 2011.
- Accepted August 23, 2011.
- American College of Cardiology Foundation
- Levin M.C.,
- Marullo S.,
- Muntaner O.,
- Andersson B.,
- Magnusson Y.
- Bengtsson K.,
- Melander O.,
- Orho-Melander M.,
- et al.
- Liggett S.B.,
- Mialet-Perez J.,
- Thaneemit-Chen S.,
- et al.
- Olshansky B.,
- Rosenfeld L.E.,
- Warner A.L.,
- et al.
- Mason D.A.,
- Moore J.D.,
- Green S.A.,
- Liggett S.B.
- Liggett S.B.
- Borjesson M.,
- Magnusson Y.,
- Hjalmarson A.,
- Andersson B.
- Levy T.,
- Walker S.,
- Mason M.,
- et al.