ADRB1 variants in Atrial Fibrillation: Small Steps and Giant Leaps toward Personalized Therapy in Cardiovascular Disease

Atrial fibrillation (AF) is the most common cardiac arrhythmia, affecting about 3 million adults in the US and costing over 26 million dollars annually to treat.^{1, 2} As a common cause of symptoms and precipitant of decompensated heart failure, atrial fibrillation contributes greatly to cardiovascular morbidity and disability. Many of the symptoms and adverse outcomes associated with untreated atrial fibrillation may be attributed to the typically rapid ventricular rate, and therefore a cornerstone of management is rate control. Rate control for atrial fibrillation can ameliorate symptoms and improve outcomes, and in prospective, randomized testing, shows no disadvantage compared with rhythm control [AFFIRM].³ Nevertheless, the pharmacologic management of atrial fibrillation remains challenging, and clinicians caring for patients with atrial fibrillation have limited tools to help tailor and/or predict response to medical therapy for the individual patient.

In this issue of the Journal, Parvez et al.⁴ provide evidence that represents a significant step towards realizing the "giant leap" of personalized medicine for individuals with atrial fibrillation by advancing a large body of work with their analysis of the association between two common genetic variants in the gene encoding the β-1-adrenergic receptor (ADRB1) and ventricular rate control in response to treatment in patients with atrial fibrillation. Their study took advantage of the extensive data available in the Vanderbilt AF Registry, a clinical and genetic registry of adult patients with documented atrial fibrillation or atrial flutter. Two strengths of this registry, and of Dr. Parvez et al.’s investigation, are the documentation of number and dose of prescribed medications for control of atrial fibrillation and the availability of echocardiographic measurements at the time of enrollment into the registry for all patients. The investigators pre-specified that they would restrict their analyses to attempted rate control using a β-blocker, calcium channel-blocker, or digoxin. In addition, well defined and specific AFFIRM (Atrial Fibrillation Follow-up Investigation of Rhythm Management) study criteria provided an objective endpoint for adequate rate control (‘responders’) within 6 months from study entry.

543 Caucasian subjects were included in the study (344 men; 199 women), with a mean age of 61.8 +/− 14 years. Fifty four % (n=295) of those enrolled in the study responded to β-blockers, calcium channel-blockers, and/or digoxin therapy within 6 months according to AFFIRM criteria (meeting one of two criteria: (1) average resting heart rate ≤ 80 beats per minute and maximum heart rate during a 6-minute walk ≤110 beats per minute OR (2) average heart rate ≤100 beats per minute on a 24-hour ambulatory holter ECG monitor and no heart rate >110% maximum predicted age-adjusted exercise heart rate). Forty six % (n=248) of the cohort failed to respond. Of the non-responders, 51% had an anti-arrhythmic drug added and 12.5% received AV node ablation and pacemaker implantation.

The authors investigated the association of ADRB1 Arg389Gly and ADRB1 Ser49Gly variants with response and found that the ADRB1 Gly389Arg variant was associated with significant differences in response to therapy while the ADRB1 Ser49Gly variant was not. The authors found that 60% of individuals with at least one ADRB1 389Gly allele were responders compared with 51% of ADRB1 Arg389Arg homozygote individuals (P=0.04). Although the authors found no difference in response according to ADRB1 Ser49Gly genotype alone (52% vs. 55%; P=0.45), when combined with ADRB1 Arg389Gly genotype, prediction of response by genotype became even more robust. Two thirds (67%) of individuals who were homozygous for ADRB1 Ser49Ser and carried at least one ADRB1 389Gly allele were responders (compared with ~50% in all other haplotypes groups; P< 0.001). Importantly, in multivariable regression analysis, while clinical variables (including echocardiographic variables) failed to significantly predict adequate response to rate control therapy, individuals carrying the ADRB1 389Gly allele were 40% more likely to respond (OR: 1.44, 95% CI: 1.01–2.04, P <0.05). This association persisted after adjustment for age and gender (OR: 1.42, 95% CI: 1.00-2.03, P <0.05).

The authors also determined the final total number and dose of β-blockers, calcium channel-blockers, or digoxin therapy needed to achieve ventricular rate control in the responder group according to haplotype. Although the number of medication prescriptions did not differ between haplotype groups, the dose did. Among responders, individuals who were homozygous for both alleles (ADRB1 Arg389Arg-Ser49Ser) required the highest doses and individuals who were ADRB1 Ser49Ser homozygotes and carried at least one ADRB1 389Gly allele required the lowest doses (atenolol 92 mg vs. 68 mg; carvedilol 44 mg vs. 20 mg; metoprolol 80 mg vs. 72mg; diltiazem 212 mg vs. 180 mg and verapamil 276 mg vs. 200 mg respectively; P < 0.01 for all comparisons).

ADRB1 and ADRB2 variants are perhaps the most widely investigated genetic variants in cardiovascular disease to date. The β-1-adrenergic receptor is a G-protein-coupled receptor and is the predominant β- adrenergic receptor subtype expressed in cardiac tissue, mediating the response to sympathetic stimulation. ADRB1 was cloned in 1987 and subsequently mapped to chromosome10q25.3. The gene has no introns, consists of 1,714 base pairs, and codes for a 51.3 kDa protein consisting of 477 amino acids. The ADRB1 Arg389Gly and ADRB1 Ser49Gly variants are the most common and well studied ADRB1 variants; both are coding variants that cause a functional change in the encoded protein. The ADRB1 Arg389Gly variant is located in the G-protein binding domain and changes G-protein coupling and adenylyl cyclase activity. The ADRB1 Ser49Gly polymorphism alters agonist-promoted desensitization.

These ADRB1 variants have been previously been associated with risk and outcomes in cardiovascular diseases such as hypertension⁵ and heart failure^{6, 7} and response to β-blocker ^8–13 and β-agonist therapy.^{14, 15} With respect to atrial fibrillation, ADRB1 49Gly allele carriers¹⁶ and individuals with the ADRB1 Arg389Gly-Ser49Gly haplotype¹⁷ have been reported to have an increased risk of atrial fibrillation; individuals with the ADRB1 Arg389Arg genotype have been reported to have higher heart rates during atrial fibrillation and to have the highest cardioversion response rate to flecanide treatment.¹⁷ Interestingly, developing auto-antibodies to the β-1-adrenergic receptor has also been reported to increase an individual’s risk of developing atrial fibrillation.¹⁸

The study by Parvez et al. significantly extends these previous findings. Although prospective validation in atrial fibrillation cohorts is needed, and some caution should be applied to the findings as drug prescription is not synonymous with drug compliance,^19–21 if validated, the study by Parvez et al. represents a major step forward for cardiovascular electrophysiology pharmacogenomics and has the promise of translation to clinical practice. The authors have elegantly demonstrated that if a (Caucasian) individual with atrial fibrillation is ADRB1 Ser49Ser homozygous and carries at least one ADRB1 389Gly allele, not only will he/she be more likely to respond to rate control with β-blocker, calcium channel-blocker, and/or digoxin, but he/she is likely to require the lowest doses of these medications to achieve this response. Thus, once identified, these individuals, representing more than one third (36%) of the population studied, are predicted to respond favorably. The pressing question is how to guide therapy in those individuals that do not have this haplotype. In the study these individuals more frequently required treatment with either higher doses of β-blockers, calcium channel-blockers, and/or digoxin, or, anti-arrhythmic drugs and/or AV node ablation and pacemaker implantation. Nevertheless, given how commonly atrial fibrillation appears in practice and the high costs associated with its care, predictive instruments incorporating genotype, as suggested by the work of Parvez et al., may have great value in tailoring the therapeutic approach to an individual patient and deserve prospective testing. If confirmed, translating the observations of this report into practice could represent a small, but significant, step toward using genotype to realize the ‘giant leap’ promise of personalized therapy for all patients with cardiovascular disease.