Pharmacogenetic factors affecting β-blocker metabolism and response

Abstract

Introduction:

β-blockers are among the most widely prescribed of all drugs, used for treatment of a large number of cardiovascular diseases. Herein we evaluate literature pertaining to pharmacogenetics of β-blocker therapy, provide insight into the robustness of the genetic associations, and determine the appropriateness for translating these genetic associations into clinical practice.

Areas covered:

A literature search was conducted using PubMed to collate evidence on associations between CYP2D6, ADRB1, ADRB2, and GRK5 genetic variation and drug-response outcomes in the presence of β-blocker exposure. Pharmacokinetic, pharmacodynamic, and clinical outcomes studies were included if genotype data and β-blocker exposure were documented.

Expert opinion:

Substantial data suggest that specific ADRB1 and GRK5 genotypes are associated with improved β-blocker efficacy and have potential for use to guide therapy decisions in the clinical setting. While the data do not justify ordering a CYP2D6 pharmacogenetic test, if CYP2D6 genotype is available in the electronic health record, there may be clinical utility for understanding dosing of β-blockers.

1. Introduction

Cardiovascular disease is the leading cause of death in the United States, causing approximately 660,000 deaths annually.[1,2] Clinically, β-adrenergic receptor blockers are commonly used in cardiovascular disease, including ischemic heart disease, hypertension, cardiac arrhythmias, and heart failure.[3–5] β-blockers modulate sympathetic nervous system (SNS) activation to produce their negative chronotropic, inotropic, and dromotropic effects. Despite their prevalent use, variation within genes that affect β-blocker pharmacokinetic and pharmacodynamic properties contribute to wide inter-patient variability in β-blocker response. The efforts toward precision medicine approaches present a unique opportunity to re-evaluate literature pertaining to pharmacogenomics of β-blocker therapy and determine the appropriateness for translating these genetic associations into clinical practice. The genes with the strongest evidence for associations with response to β-blocker therapy include CYP2D6, ADRB1, ADRB2 and GRK5. The purpose of this expert opinion article is to describe pharmacogenetic factors affecting β-blocker metabolism and response, provide insight into the robustness of the genetic associations, and discuss clinical interpretations for potential translation of variants influencing β-blocker response into practice.

2. CYP2D6 and β-blockers

2.1. Pharmacokinetic effects

Cytochrome P450 2D6 (CYP2D6) serves an important role in the biotransformation of approximately 25% of drugs.[6] Several β-blockers are substrates for CYP2D6, including metoprolol, carvedilol, propranolol, labetalol, nebivolol, and timolol.[7–11] During the 1980s, cytochrome P450 2D6 was purified and mapped to chromosome 22q13.[12–14] Currently, CYP2D6 genotype is utilized in clinical practice to predict CYP2D6 metabolism phenotype, which is used to facilitate prescribing recommendations for a variety of medications.[15–20] CYP2D6 genotype is used to assign CYP2D6 phenotype, and in many cases, the CYP2D6 activity score method is used in clinical practice to translate CYP2D6 genotype to predicted phenotype.[15–21] Currently, four phenotypes are used, with increasing metabolism potential in order from poor metabolizer (PM; no CYP2D6 activity), intermediate metabolizer (IM, approximately 25%−50% normal metabolism), normal metabolizer (NM), to ultra-rapid metabolizer (UM; increased CYP2D6 activity secondary to gene duplication).[22] While there are no Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines for β-blockers, associations between genotype and β-blocker pharmacokinetics and pharmacodynamics date back to the 1980s and 1990s.

Metoprolol metabolism is mediated extensively by the cytochrome P450 system, with CYP2D6 contributing to approximately 70–80% of metoprolol’s metabolism to inactive metabolites.[23] CYP2D6 is a highly polymorphic gene, with more than 100 alleles described.[24] CYP2D6 genotype is associated with altered metoprolol metabolism and an allele-dose response effect between CYP2D6 genotype and plasma metoprolol concentrations has been previously described.[25–33] As the number of active CYP2D6 alleles increases (from zero in PMs to > 2 in UMs), so does the metabolism of metoprolol. This is reflected by variability in pharmacokinetic parameters (e.g., oral clearance, area under the plasma concentration-time curve, elimination half-life) based on CYP2D6 genotype.[29–31,33,34] A meta-analysis of metoprolol pharmacokinetics by CYP2D6 phenotype demonstrated an approximate 6-fold decrease in metoprolol oral clearance in CYP2D6 PMs compared with NMs.[35] Metoprolol oral clearance was 1.5-fold lower in IMs and 2.6-fold higher in UMs compared with NMs.[35]

There is the largest body of literature on the impact of CYP2D6 phenotype on metoprolol pharmacokinetics because it is the most commonly prescribed of the β-blockers that are CYP2D6 substrates, and because CYP2D6 contributes more to the metabolism of metoprolol than the other β-blockers. However, the literature clearly documents the role of CYP2D6 on the pharmacokinetics of carvedilol, propranolol, nebivolol, and timolol.[36–49] Carvedilol also exhibits stereoselective metabolism, with CYP2D6 PMs demonstrating decreased clearance of R-carvedilol compared with CYP2D6 NMs.[50] Similar to metoprolol, there is a linear relationship between the CYP2D6 activity score and carvedilol clearance.[34,51–53] Regarding nebivolol, plasma concentrations, area under the plasma concentration-time curve, and clearance were consistently different between NMs and PMs.[43–49] Of the β-blockers discussed, labetalol has the weakest literature base.[36] Other commonly used β-blockers not subject to metabolism via CYP2D6 include: acebutolol, atenolol, betaxolol, bisoprolol, and nadolol.

2.2. Pharmacodynamic Effects

CYP2D6 genotype is associated with altered HR response to β-blocker therapy,[25,27,30,31,54] and one study showed CYP2D6 phenotype was one of the most significant predictors of variability in HR response to metoprolol.[54] Of note, HR response, in the presence of β-blocker therapy, is a surrogate measure of the degree of β1-adrenergic receptor blockade. Therefore, due to variability in CYP2D6 activity, the β-blockade and HR response can vary by dose. At equal doses in the lower end of the dosing range (e.g. 50 to 100 mg), CYP2D6 PM and IM phenotypes are associated with greater reductions in HR after metoprolol exposure compared to NMs.[25,27,30,31,54] Additionally, among CYP2D6 IM/PMs with New York Heart Association (NYHA) class II-IV heart failure with reduced ejection fraction (HFrEF), the odds of achieving a HR ≤ 60 beats/min are approximately 2-fold higher than CYP2D6 NMs during dose titration.[25] Genotypes associated with reduced CYP2D6 metabolism have also been linked to increased odds of achieving a target HR of less than 70 beats/min in patients undergoing elective PCI.[55] CYP2D6 no functional alleles (e.g., CYP2D6*4) are also associated with lower maintenance doses of metoprolol.[56] While the relationships between CYP2D6 genotype and metoprolol and carvedilol pharmacokinetics are similar, a pharmacodynamic relationship between CYP2D6 and carvedilol does not appear as strong. This observation is expected given that CYP2D6 accounts for less of S-carvedilol’s metabolism; S-carvedilol is primarily responsible for β-blockade.[50] Among healthy volunteers receiving carvedilol, maximum HR reduction and resting HR did not differ by CYP2D6 activity score.[34] Similar to carvedilol, variability in nebivolol pharmacokinetics does not translate to pharmacodynamic differences between PMs and NMs.[44,45,48,57,58] Given the sigmoid dose/concentration-response relationship for β-blockers, at a certain point, increasing plasma drug concentrations does not lead to increasing response.[25,59]

While the associations between CYP2D6 genotype and HR response are consistent, the data are weaker for BP and other responses. There are fewer studies that identify a positive association between CYP2D6 genotype and blood pressure response.[25,60] In studies that identified an association between CYP2D6 genotype and DBP response, a difference of approximately 5 mmHg between CYP2D6 PMs and NMs has been reported.[25,60] Only one study identified an association between CYP2D6 genotype and SBP.[60] More consistently, negative findings for CYP2D6 genotype and blood pressure response have bene reported.[25,26,30,31,54] Overall, the evidence for metoprolol points towards no association between CYP2D6 genotype and clinical efficacy of metoprolol as it relates to SBP and DBP response. In a post hoc pharmacogenomic sub-study of heart failure patients in MERIT-HF[61,62], CYP2D6 genotype was also not associated with the composite clinical outcome of death and hospitalization.[25] In another study of HFrEF participants, CYP2D6 phenotype was not associated with risk of cardiac decompensation, increased need for background HF medications, or final daily metoprolol CR/XL doses.[33] While data are limited for carvedilol, one study suggested CYP2D6 genotype was not associated with BP response.[34]

Data also suggest a lack of association between CYP2D6 phenotype and adverse responses to β-blockers. CYP2D6 genotype was not associated with increased incidence of dizziness, fatigue, or absent-mindedness in healthy volunteer studies of carvedilol.[34] Data regarding adverse drug reactions with metoprolol use are more abundant. In a real-world cohort of primary care participants where metoprolol was prescribed, PM and IMs did not have a higher incidence of common metoprolol adverse drug reactions (headache, dizziness, fatigue or drowsiness, sleep disturbances, dyspnea, or cold extremities) compared with NM and UMs, however sexual dysfunction was more common in NMs and UMs compared with PM and IMs.[26] This latter finding is somewhat paradoxical since this finding is not consistent with a concentration-related adverse effect. Similarly, in a prospective, randomized clinical trial with metoprolol in hypertensives, CYP2D6 phenotype was not associated with the incidence of metoprolol-related side effects.[54] In another clinical study of metoprolol in hypertensives, CYP2D6 phenotype was not associated with general, or dose-limiting metoprolol-related adverse events.[63]

Collectively, the data suggest the impact of CYP2D6 genotype/phenotype on response is more evident at lower doses, when a patient is still on the linear portion of the dose-response curve, but pharmacogenetic effects are less evident at higher doses. Overall, pharmacokinetic variability across CYP2D6 genotypes appear to translate primarily to differences in heart rate response after β-blocker initiation, with most data suggesting that CYP2D6 phenotype does not influence BP or other responses, including adverse responses to β-blockers. Given that β-blockers are typically initiated at a low dose, with upward titration based on response or a HR < 60 bpm, then patients who are PM or IM may achieve the target or HR response at a lower than expected dose. In the precision medicine era, CYP2D6 genotype may be used to predict tolerable maintenance doses of metoprolol and carvedilol in patients with HFrEF.[56]

3. ADRB1 and β-blockers

β1-adrenergic receptors, encoded for by ADRB1, are G-protein coupled receptors that are primarily expressed in cardiac tissue to mediate chronotropic, dromotropic, and inotropic effects from SNS activation.[64] ADRB1 variants are common in the overall population and exhibit differences in frequencies across ancestral populations (Table 1).[64] The two most common and well-studied ADRB1 variants are rs1801252 (ADRB1 145A>G; Ser49Gly) and rs1801253 (ADRB1 1165 C>G; Arg389Gly). The variant at rs1801252 results in increased agonist-promoted down-regulation, whereas the variant at rs1801253 alters G-protein coupling, decreases adenylyl cyclase activity, thus attenuating generation of cAMP (Figure 1).[59,65,66] Thus based on the functional effects for both variants, the major allele is predicted to be associated with greater agonist-mediated response. This functional basis for differences in β-blocker response has been studied extensively in healthy volunteers and patients with a variety of cardiovascular diseases.

Table 1.

Phenotype and Variant Allele Frequencies

	EUROPEAN	AFRICAN	ASIAN
PHENOTYPE FREQUENCIES
CYP2D6
PM	6	2	2
IM	38	45	29
NM	51	44	66
UM	4	4	2
VARIANT ALLELE FREQUENCIES
ADRB1
rs1801252 (Gly49)	13	16	5
rs1801253 (Gly389)	31	37	38
ADRB2
rs1042713 (Gly16)	63	51	46
rs1042714 (Glu27)	42	29	12
GRK5
rs2230345 (Leu41)	2	16	0

IM: intermediate metabolizer; NM: normal metabolizer; PM: poor metabolizer; UM: ultrarapid metabolizer. CYP2D6 phenotype frequencies from https://cpicpgx.org/. ADRB1, ADRB2, and GRK5 variant allele frequencies from dbGaP, with the resulting encoded variant allele amino acid in parentheses.

Figure 1.

Pharmacodynamic Pathways of β1 and β2 Receptors and Resulting cAMP Generation

3.1. Pharmacodynamic effects

A number of studies have addressed the associations between ADRB1 variants and HR and BP responses, in healthy volunteers and patients with cardiovascular disease. Most studies show no association between ADRB1 and HR response to β-blockers, while several studies show an association between the commonly studied ADRB1 SNPs and BP response, particularly the DBP response. In healthy volunteers, up-titration of dobutamine resulted in greater increases in SBP in Arg389Arg participants compared with Gly389 carriers.[67] Consistent with greater agonist-mediated response for Arg389Arg, another study found this genotype was also associated with greater response to β-blockade. Specifically, a healthy volunteer study found Arg389Arg was associated with greater BP lowering and decreases in resting and exercise HR responses, across various metoprolol doses, compared with Gly389Gly,[68] and the association between Arg389 and improved HR response became more apparent as the metoprolol dose was increased.[68] In another healthy volunteer study, no significant differences in HR were seen between ADRB1 Arg389Gly genotype groups.[69] A pharmacodynamic study of patients with chronic HF, initiated on β-blocker therapy with either carvedilol or bisoprolol, showed ADRB1 genotype was not associated with changes in HR or SBP after the maximal tolerated dose was obtained.[70] Subsequent studies conducted in various patient populations did not identify associations between Arg389Gly and maximum HR reduction.[34,41,70–73] Similarly, several studies found no association between heart rate response to β-blockers and genotype at codon 49.[67,68,70,71,73–77]

Data suggest the ADRB1 gene may be important determinant for DBP, but not SBP response to β-blocker therapy. A study of essential hypertension patients treated with metoprolol found those with Arg389Arg genotype had a 6.5 mmHg greater DBP response than Gly389 carriers.[71] These findings were corroborated in another study of participants with essential hypertension.[72] Ophthalmic timolol administration was also associated with higher systolic and diastolic arterial pressures in Ser49Ser vs. Gly49 carriers, while genotype at position 389 was not associated with BP response.[41] A study in healthy volunteers and another study in participants with hypertension and left ventricular hypertrophy produced inconsistent results with DBP.[68,75] Some data suggest that consideration of the ADRB1 haplotypes that result from the two commonly studied variants may be more informative than individual SNPs, and thus studies of single SNPs may have resulted in inconsistent findings with pharmacodynamic parameters.[78] Specifically, multiple studies have shown the Ser49-Arg389 haplotype is associated with improved DBP lowering compared with Ser49-Gly389 haplotype in participants with essential hypertension and coronary artery disease.[71,72,78] Regarding SBP, previous findings do not support an association with ADRB1 genotype, with one positive and five negative association studies.[62,70–72,75,78]

Among study participants with chronic HF, ADRB1 genotype was not associated with HR or BP responses at the end of β-blocker titration, or with the attained maintenance β-blocker dose.[33,62] Similarly, in a pharmacogenetic sub-study of MERIT-HF[61], HR reduction attained during metoprolol titration, and maintenance dose attained was not different based on codon 389 genotype.[62] In another cohort of NYHA class II to IV HFrEF participants receiving carvedilol, HR and BP responses, both at rest and during exercise, were similar between ADRB1 genotypes at codon 389.[79] These findings were validated in a randomized trial of chronic HF participants scheduled to receive carvedilol or bisoprolol.[59] Additionally, no significant interaction between the effects of ADRB1 and CYP2D6 on treatment outcomes has been observed.[80]

Collectively for hemodynamic responses, the strongest data exist for an association between ADRB1 genotype and DBP response to β-blockers, while HR and SBP responses and target dose attainment seems to be minimally affected by ADRB1 genotype.

3.2. Coronary artery disease outcomes

Numerous lines of evidence suggest certain ADRB1 genotypes are associated with risk for cardiovascular events, and such risk may be ameliorated with β-blocker therapy, implying certain genotypes derive greater benefit from β-blockers (Table 2; Supplementary Table 1). Clinical outcomes among a cohort of patients with coronary artery disease randomized to treatment with verapamil SR or atenolol (INVEST trial)[81] focused on ADRB1 haplotype and found an increased risk of major adverse cardiovascular events (MACE; defined as composite of death, nonfatal myocardial infarction, and stroke) among Ser49-Arg389 carriers compared with non-carriers (HR: 1.51; 95%CI: 1.07–2.12).[78] This association was primarily driven by the risk for death among Ser49-Arg389 carriers (HR: 3.66; 95%CI: 1.68–7.99).[78] The association with cardiovascular events identified in the INVEST trial is consistent with a case-control study that reported ADRB1 genotype at codon 389 was associated with acute myocardial infarction.[82] When the INVEST analysis was stratified by treatment allocation, Ser49-Arg389 carriers randomized to verapamil remained at increased risk for death (HR: 8.58; 95%CI: 2.06–35.8), while the haplotype’s risk was reduced for Ser49-Arg389 carriers randomized to atenolol (HR: 2.31; 95%CI: 0.82–6.55; p=0.11). Therefore, treatment with atenolol may offset the increased risk for MACE observed with the ADRB1 Ser49-Arg389 haplotype.[78] These data suggest Ser49-Arg389 allele carriers are likely to derive the greatest benefit from β-blocker therapy.

Table 2:

Evidence Table for Pharmacogenetics of Cardiovascular Outcomes with respect to ADRB1, ADRB2, and GRK5

Study Typea	Study Population	β-blocker	No. of Subjects	Duration	Gene	SNPs	Outcomes	Findings
Prospective (Pacanowski, 2008)	CAD secondary prevention	Atenolol	5895	2 years	ADRB1, ADRB2	ADRB1: rs1801252 (Ser49Gly), rs1801253 (Arg389Gly) ADRB2: rs1042713 (Gly16Arg), rs1042714 (Gln27Glu), rs1042718 (Arg175Arg)	MACE (death, MI, stroke)	Ser49-Arg389 carriers – increased MACE and death risk Death risk in Ser49-Arg389 carriers attenuated in atenolol, but not verapamil treated participants ADRB2 not associated with MACE overall or in treatment specific manner
Prospective (Fiuzat, 2013)	HFrEF (LVEF < 35%; NYHA class II-IV)	Multiple	902	2.5 years	ADRB1	ADRB1: rs1801252 (Ser49Gly), rs1801253 (Arg389Gly)	All-cause death or all-cause hospitalization All-cause death	Arg389Arg - increased risk of death with low-dose β-blocker that was offset with high-dose β-blocker. β-blocker dose not associated with outcomes in Gly389 carriers Ser49Gly: no association
Prospective (White, 2003)	HFrEF (LVEF ≤ 40%; NYHA class II-IV)	Metoprolol CR/XL	600	1 year	ADRB1	ADRB1: rs1801253 (Arg389Gly)	All-cause mortality or hospitalization	ADRB1 genotype not associated with outcomes
Prospective (Liggett, 2006)	HFrEF (LVEF ≤ 35%; NYHA class III to IV)	Bucindolol	1040	3 years	ADRB1	rs1801253 (Arg389Gly)	All-cause mortality HF Exacerbations (Hospitalization because of HF)	Arg389Arg – bucindolol decreased risk for death and HF exacerbations vs. placebo Gly389 carriers – no association with mortality benefit or reduced HF exacerbations with bucindolol vs. placebo
Prospective (Magvanjav, 2017)	Small artery ischemic stroke secondary prevention	Multiple	926	4 years	ADRB1	ADRB1: rs1801252 (Ser49Gly), rs1801253 (Arg389Gly)	MACE (death, MI, stroke)	Gly49 carriers + β-blocker - increased risk for MACE vs. Ser49Ser + β-blocker Arg389Arg not associated with MACE in the presence of β-blocker therapy
Prospective (Liggett, 2008)	HFrEF (LVEF < 40%; NYHA class II-IV)	Multiple	375	2.5 years	GRK5	rs2230345 (Gln41Leu)	Death or heart transplantation	Increased risk of primary endpoint in Gln41Gln was offset by β-blocker treatment Leu41 carriers derive minimal benefit from β-blocker treatment
Prospective (Cresci, 2009)	HFrEF (LVEF < 40%)	Multiple	2451	4 years	ADRB1, GRK5	ADRB1: rs1801253 (Arg389Gly) GRK5: rs2230345 (Gln41Leu)	Death or heart transplantation	Gly389 carrier – Death or heart transplantation risk attenuated with β-blocker Among African Americans: Increased risk of primary endpoint in Gln41Gln was offset by β-blocker treatment
Prospective (Lobmeyer, 2011)	CAD	Atenolol	1258	2 years	GRK5	rs2230345 (Gln41Leu)	MACE (death, MI, stroke)	Leu41 carrier – decreased odds of MACE regardless of treatment strategy

^aProspective studies were designed specifically to test pharmacogenetics hypotheses, or the primary outcome studied in the pharmacogenetic study was the primary outcome in the clinical trial.

3.3. Chronic heart failure

Studies suggest a variety of β-blocker response phenotypes in heart failure that may vary by ADRB1 genotype (Table 2; Supplementary Table 1). Among participants with NYHA Class II or III symptoms with HFrEF, ADRB1 genotype was associated with increased requirements of other heart failure medications (e.g., diuretics) for worsening symptoms of heart failure during β-blocker titration.[33] Gly389 carriers were more likely than Arg389Arg participants (48 vs. 14%) to have increased requirements of heart failure medications. Additionally, Ser49 homozygotes were more likely to require increases in heart failure medication during titration compared to Gly49 carriers (41 vs. 11%).[33] When the analysis was expanded to evaluate the risk between ADRB1 haplotypes, approximately 50% of participants with the Ser49Ser/Arg389Gly diplotype required increases in concomitant heart failure medications during metoprolol titration. ADRB1 genotype was not associated with metoprolol discontinuation rates, or metoprolol dose attained at the end of β-blocker titration.[33,62,83] The absence of an association between ADRB1 genotype and attained β-blocker dose has been validated in cohorts of participants with coronary artery disease receiving atenolol, essential hypertension receiving metoprolol, and Marfan syndrome receiving atenolol.[71,73,78] Arg389 carriers with NYHA class II to III chronic HF were also more likely to utilize emergency department resources compared to non-carriers.[84]

Studies support greater benefits from β-blocker therapy among HFrEF patients with the Arg389Arg genotype compared to Gly389 carriers. Furthermore, up-titration of β-blocker dose - a key treatment paradigm for management of chronic HFrEF - appears to benefit Arg389Arg patients, but not Gly389 carriers. A study of participants with NYHA class II-IV HFrEF evaluated the impact of the ADRB1 Arg389Gly variant on changes in left ventricular function during maintenance therapy with carvedilol. There was a trend towards improved LVEF in Arg389 carriers compared with Gly389Gly.[79] In a pharmacogenetic substudy of the HF-ACTION trial[85], a significant interaction between ADRB1 genotype at codon 389 and β-blocker dose for all-cause mortality has been identified among study participants with HFrEF.[86] Among ADRB1 Arg389Arg participants, low-dose β-blocker relative to high-dose β-blocker therapy, was associated with a two-fold increase in the risk for all-cause mortality.[86] This suggests the risk associated with higher adrenergic activation in Arg389Arg individuals is not sufficiently offset by low dose β-blockade. This increased risk was not observed among Gly389 carriers receiving low versus high dose β-blockers.[86] No association was observed between ADRB1 genotype at codon 49 and β-blocker dose for all-cause mortality.[86] This is consistent with findings in numerous other settings, where the codon 49 polymorphism is less commonly associated with clinical phenotypes than the codon 389 polymorphism.

In a genetic substudy of the BEST trial,[87] a randomized, controlled trial of participants with NYHA class III to IV HFrEF, risks for all-cause mortality and hospitalization secondary to a HF exacerbation were significantly reduced in bucindolol treated participants with the Arg389Arg genotype compared with placebo. The risks for death and hospitalization were not significantly different between bucindolol and placebo arms among Gly389 carriers, implying that the benefits of bucindolol in the studied HF population are confined to those with the Arg389Arg genotype.[88] Additionally, genotype at codon 49 in this cohort was not associated with any of the various clinical phenotypes tested.[88] Similar to the associations between ADRB1 genotype and all-cause mortality, a genetic sub-study of the same population identified a decreased risk for ventricular tachycardia or fibrillation among bucindolol treated participants with the Arg389Arg genotype compared with placebo; this risk difference was not observed for Gly389 carriers, again suggesting that phenotype risk associated with Arg389Arg can be offset with β-blocker treatment.[89] In the same trial, among those with atrial fibrillation, bucindolol therapy in Arg389Arg participants was associated with lower risk for the composite of death and heart failure hospitalization compared with placebo.[90] The risk for the composite endpoint was not significantly different between Gly389 carriers.[90] In contrast, in the MERIT-HF genetic sub-study, participants with the Arg389Arg genotype were not at increased risk for the composite of death or hospitalization compared to Gly389 carriers; this observation also held within the placebo and metoprolol-treated arms.[62] This finding has been validated in a separate, prospective study of HFrEF participants where risk for death or heart transplantation was similar between Gly389 carriers and Arg389Arg participants on maintenance β-blocker doses.[91] While β-blocker therapy is essential to management of all patients with chronic HFrEF, ADRB1 Arg389Arg participants appear to derive the greatest clinical benefits from β-blockade.

3.4. Marfan syndrome

ADRB1 variation also explains interindividual variability in response to atenolol among subjects with Marfan syndrome, an autosomal dominant connective tissue disorder. In a pharmacogenomic substudy of a randomized trial conducted by the Pediatric Heart Network[92], clinical response to atenolol – assessed by improvements in the aortic-root z-score – was greater among individuals with the ADRB1 Arg389Arg genotype compared with Gly389 carriers.[73] While the larger study found no difference in the rate of aortic-root dilation between atenolol and losartan, an interaction was observed in the pharmacogenomic substudy between ADRB1 genotype at codon 389 and drug response.[73,93] Among those with the Arg389Arg genotype, greater improvements in aortic-root z-score were noted for atenolol-treated compared with losartan-treated participants. ADRB1 Ser49Gly was not associated with variability in treatment response to atenolol. Similar to other disease phenotypes, ADRB1 genotype was not associated with maintenance atenolol dose at study completion.[73] Among patients with Marfan syndrome, ADRB1 genotype has the potential to identify those likely to benefit from atenolol therapy, and potentially to identify those who should be preferentially treated with atenolol over losartan.

3.5. Ischemic stroke outcomes

While genetic differences in cardiovascular outcomes among participants with CAD and chronic HF were associated with variation at codon 389, genotype at codon 49 has been associated with MACE in participants with ischemic stroke (Supplementary Table 1). The pharmacogenomic associations between ADRB1 genotype and clinical outcomes were assessed in data from SPS3, an international multicenter randomized controlled clinical trial that evaluated the effects of different antihypertensive and antiplatelet therapy regimens on the rate of recurrent stroke.[94] In the pharmacogenomic analysis, the cumulative incidence of MACE and ischemic stroke were higher among Gly49 allele carriers compared with Ser49Ser participants.[95] The relative risk for MACE and recurrent ischemic stroke were 1.6 and 1.8-fold higher, respectively, among Gly49 carriers compared with Ser49Ser participants.[95] Among Gly49 carriers treated with β-blockers, the relative risks for MACE and ischemic stroke further increased to 2.8 and 2.9-fold those of Ser49Ser participants.[95] Their risk for MACE and ischemic stroke was similar between Gly49 carriers and non-carriers among those not treated with β-blockers. Overall, these data suggest the association between the Gly49 allele and risk for MACE may be increased with β-blocker therapy.

3.6. Summary

In summary, a number of studies identified associations between ADRB1 genotype and hemodynamic and clinical outcomes across various cardiovascular disease phenotypes. Regarding hemodynamic responses, the strongest data exist for an association between ADRB1 genotype and DBP response to β-blockers. Among hypertensive patients with coronary artery disease, β-blocker therapy may offset the risk for MACE among carriers of the Ser49-Arg389 haplotype. The use of β-blockers in patients with chronic HF appears to be associated with greater improvements among those with the Arg389Arg genotype. Furthermore, a genetic substudy of the BEST trial found no difference in mortality among Gly389 carriers between β-blocker therapy and placebo. This suggests the mortality benefit in this population may be heavily driven by the benefits of neurohormonal blockade in Arg389Arg patients. Benefits from β-blocker therapy among participants with the Arg389Arg genotype also extended to populations with Marfan syndrome as greater improvements in aortic-root z-score were noted for atenolol-treated compared with losartan-treated participants. Genotype at codon 49 appears to have the greatest impact among participants with ischemic stroke. Available data suggest cardiovascular risk among Gly49 allele carriers may be increased among β-blocker treated participants. Further research is needed to validate this finding given the clinical implications for treatment of patients with small artery stroke. Collectively, studies have documented associations between ADRB1 and metoprolol, atenolol, bucindolol, and carvedilol, implying that this is a class effect that would be evident for any drug that antagonizes the β1-adrenergic receptor.

4. ADRB2 and β-blockers

ADRB2 encodes the beta-2-adrenergic receptor, a G-protein-coupled adrenergic receptor. Beta-2 adrenergic receptor is expressed in cardiac myocytes, bronchial smooth muscle, and vascular smooth muscle cells. Two common ADRB2 SNPs rs1042713 (ADRB2 46A>G; Arg16Gly) and rs1042714 (ADRB2 79C>G; Gln27Glu) lead to a change in the encoded amino acid and have been commonly studied (Table 1). The variant at rs1042713 enhances agonist-promoted down-regulation, ultimately resulting in decreased cAMP formation.[96,97] In contrast, the Glu27 variant at rs1042714 is resistant to agonist-promoted downregulation, ultimately resulting in increased cAMP formation (Figure 1).[96–98] Due to linkage disequilibrium, these two variants lead to three haplotypes. The haplotype containing the variant allele at both positions (rs1042713 and rs1042714) displays enhanced receptor downregulation.[96,97] The ADRB2 Gly16-Glu27 haplotype and ADRB2 genotype at codon 27 have been associated with on-treatment BP and HR.[34,78,99] Among participants from the INVEST trial, differences in DBP were < 2 mmHg between ADRB2 haplotypes, leading study investigators to challenge the clinical significance of these statistically significant findings.[78] Additionally, ADRB2 genotype has not been associated with maximal tolerated β-blocker dose.[33,78]

Among study participants with chronic heart failure, the associations between ADRB2 genotype and outcomes are mixed. Among carvedilol-treated participants with NYHA class II to IV symptoms and a LVEF ≤ 35%, the ADRB2 haplotype, Gly16Gly/Glu27Glu, was associated with greater improvements in LVEF, mean arterial blood pressure, and pulmonary capillary wedge pressure when compared with Gly16Gly/Gln27Gln and Arg16Arg/Gln27Gln haplotypes.[79] Another study evaluated ADRB2 genotype and found greater improvements in LVEF, resting HR, and the 6-min walk test among carvedilol-treated Glu27 carriers compared with non-carriers.[99] While surrogate markers appear to be associated with ADRB2 genotype, clinical outcomes have not been consistently associated with ADRB2 genotype.[33,78,100] In a study of participants with NYHA class II-IV chronic HF on maintenance β-blocker therapy, Gly16 carriers were not at increased risk of cardiovascular death or heart transplant when compared with Arg16Arg participants.[100] A prospective study of chronic HF participants on metoprolol CR/XL found no differences in decompensated heart failure outcomes, utilization of heart failure medications, or quality-of-life metrics between genotype groups at codon 16.[33] Among participants in INVEST with coronary artery disease and hypertension receiving atenolol, risk for MACE (death, MI, stroke) was not associated with ADRB2 haplotype.[78] Collectively, data among participants with chronic HF and coronary artery disease suggest ADRB2 genotype is not associated with cardiovascular events while on a background of maintenance β-blocker therapy.

5. GRK5 and β-blockers

G protein-coupled receptor kinases (GRK) function to desensitize ligand-occupied G protein-coupled receptors, like β-adrenergic receptors, ultimately leading to receptor down-regulation.[64,101] G protein-coupled receptor kinase 5 is widely expressed in cardiac tissue.[101,102] The GRK5 variant rs2230345 (GRK5 122A>T; Gln41Leu) exhibits allele frequency differences between ancestral populations, with approximately 40% of black and 2% of white patients carrying the variant allele.[101] The GRK5 Leu41 variant encodes an increased function kinase that enhances desensitization of the β−1 adrenergic receptor in vitro, thus reducing receptor stimulation and cAMP generation (Figure 1).[101,103] A prospective study in patients of African ancestry with heart failure showed longer survival times were observed in GRK5 Leu41 carriers compared with GRK5 Gln41Gln participants, in the absence of β-blocker therapy (Supplementary Table 1).[101] Among GRK5 Leu41 carriers, there was no benefit associated with β-blocker therapy, whereas GRK5 Gln41Gln had significant improvements in transplant-free survival when treated with a β-blocker. Overall, GRK5 genotype that encodes the Leu41 allele enhances β-adrenergic receptor desensitization and decreases the risk for death or heart transplant in a gene-dose dependent fashion. In the presence of β-blocker therapy, GRK5 Gln41Gln participants experienced survival advantages similar to that of GRK5 Leu41 carriers with no β-blocker therapy.[101] These findings were corroborated in a prospective study of participants with HFrEF, where GRK5 Leu41 allele carriers, compared with GRK5 Gln41Gln participants, had improved survival among blacks not treated with β-blocker therapy.[91] In contrast, there was no association between GRK5 genotype at codon 41 and HF outcomes among β-blocker treated black participants, suggesting that β-blocker therapy nullifies the risk associated with the Gln41Gln genotype.[91] Additional studies identified the expanded protective role of the GRK5 Leu41 allele to extend beyond patients with HF. Data from a hypertensive cohort with coronary artery disease support decreased odds of MACE (death, myocardial infraction, or stroke) among GRK5 Leu41 carriers compared with GRK5 Gln41Gln participants.[104] Importantly, the effects of the GRK5 Leu41 allele in hypertensive patients were independent of treatment, suggesting the degree of SNS activation in the underlying pathophysiology (e.g., hypertension, HFrEF) may modulate treatment response.[104] While GRK5 Leu41 protects against adverse cardiovascular outcomes, there are no significant differences between GRK5 Gln41Leu genotype and blood pressure response to antihypertensive medication or exercise-induced change in heart rate after atenolol administration.[103,104] Associations between GRK5 genotype and cardiovascular outcomes have been observed with carvedilol, metoprolol, and atenolol, thus implying the effects are consistent for drugs causing β1-adrenergic receptor blockade. Further studies are needed to elucidate the pharmacogenomic associations between GRK5 genotype and cardiovascular outcomes.

6. Conclusion

CYP2D6, ADRB1, ADRB2, and GRK5 have the strongest evidence for pharmacogenomic associations with β-blocker pharmacokinetic and pharmacodynamic response and clinical outcomes. Data available to-date have evaluated these pharmacogenomic associations across a range of clinical phenotypes. The effects of genetic polymorphisms in CYP2D6 on CYP2D6-dependent β-blocker (i.e., metoprolol, carvedilol, propranolol, labetalol, nebivolol, and timolol) pharmacokinetics is apparent across many studies. Metoprolol is the principal β-blocker where pharmacokinetic variability translates to differences in pharmacodynamic response, likely due to its extensive reliance on CYP2D6 for its inactivation. Despite greater AUC and decreased apparent oral clearance in CYP2D6 PM and IMs, these pharmacokinetic differences primarily translate into differences in HR, but not other β-blocker response phenotypes. Clinically, this results in CYP2D6 PMs and IMs experiencing greater decreases in HR with certain β-blocker doses in the lower end of the dosing range, though this may be less apparent at higher doses. Though there has been much speculation that PMs and IMs would have increased prevalence of adverse effects to β-blockers, nearly all data indicate this is not the case. Thus, CYP2D6 pharmacogenetics have little to no impact on the efficacy or adverse effects associated with β-blockers.

ADRB1, ADRB2, and GRK5 variants have been studied in patients with many cardiovascular diseases, including essential hypertension, ischemic stroke, coronary artery disease, chronic heart failure, and Marfan syndrome. ADRB1 genotype has been associated with major cardiovascular events and other clinical phenotypes in a variety of clinical settings, including heart failure, ischemic heart disease, and stroke prevention. Among patients with coronary artery disease, β-blocker therapy ameliorates the risk for MACE among Ser49-Arg389 carriers. In patients with chronic HF and the ADRB1 Arg389Arg genotype, the use of β-blockers also appears to reduce the risk for cardiovascular events; a finding less commonly observed for Gly389 carriers. Generally, it appears that differences in MACE and mortality by ADRB1 genotype are mitigated in the presence of β-blocker therapy, thus highlighting the potential for preferential use of β-blocking agents in ADRB1 Arg389Arg patients.

Regarding ADRB2, data among coronary artery disease and chronic HF populations receiving β-blockade are less robust compared to associations with ADRB1. Overall, similar clinical outcomes have been reported between ADRB2 genotype groups among patients receiving maintenance β-blocker therapy. In the case of GRK5 variants, the major allele homozygous genotype (Gln41Gln) was associated with greater treatment response with β-blockers, whereas Leu41 allele carriers experienced similar cardiovascular outcomes between β-blocker treated and untreated participants. While not all β-blockers encountered in clinical practice are represented in the literature, available data for ADRB1 and GRK5 suggest the observed associations apply across the β-blocker medication class. This is evidenced in part by clinical findings across multiple β-blockers. Overall, use of pharmacogenomic data to guide treatment of β-blocker therapy in patients with hypertension, coronary artery disease, and heart failure has the best potential for translation into clinical practice.

7. Expert opinion

In the current precision medicine era, genes involved in β-blocker disposition and response have not yet been translated to clinical practice. Of the genes discussed in this review, only CYP2D6 is used in clinical practice, but not in the context of β-blocker therapy. It is possible for sites that have already adopted CYP2D6 pharmacogenetics to implement strategies for β-blocker prescribing. Of all the β-blockers, metoprolol has the greatest pharmacogenetic evidence with CYP2D6. Regarding clinical utility, the data do not support ordering a CYP2D6 pharmacogenetic test for the purposes of guiding β-blocker therapy. However, if CYP2D6 genotype data are available in the patient’s electronic health record, these data may have clinical utility. Specifically, knowledge that a patient is a PM or IM may provide the prescriber with some assurance if the patient appears to be responding well to a lower than expected dose. CYP2D6 PMs may achieve goal HR reductions at lower β-blocker doses than CYP2D6 NMs, which is the response often used to indicate complete β-blockade. Similarly, CYP2D6 UMs may require increased β-blocker doses to achieve responses comparable to NMs. Additionally, patients with HFrEF and CYP2D6 IM or PM phenotypes may never reach guideline-directed doses of metoprolol yet have a sufficient degree of β-blockade to derive the expected benefits of a full dose. Clinically, HR is an excellent biomarker for the degree of β-adrenergic receptor blockade, and in most settings β-blocker therapy is started at a low dose and titrated upwards based on HR and other response.

Currently, ADRB1 genotype is not used to guide β-blocker prescribing decisions despite its associations with major adverse cardiovascular events and mortality. Despite the absence of clinical use of these genetic data, the total number of clinical outcomes studies and consistency of the data are of similar or greater quality than many of the examples for which Clinical Pharmacogenetics Implementation Consortium Guidelines (https://cpicpgx.org/) exist, and thus guidelines for ADRB1 genotype informed β-blocker therapy should be considered. Specifically, numerous studies, including in ischemic heart disease, hypertension, heart failure, stroke and Marfan Syndrome suggest better β-blocker treatment outcomes in those with the Arg389Arg genotype, often because this genotype is associated with increased risk for adverse cardiovascular outcomes, which is offset with β-blocker therapy. Most of these studies point to the Arg389Arg genotype, or the Ser49-Arg389 haplotype being associated with risk for which β-blockers largely moderate that risk. Current evidence suggests this increased risk is ameliorated in the presence of β-blocker therapy. This highlights a gap in current clinical pharmacogenomics implementation, whereby ADRB1 genotype could be utilized to preferentially prescribe β-blocker pharmacotherapy in patients with the ADRB1 Arg389Arg genotype. A genotype-guided strategy for β-blockers may be viewed as less impactful in clinical practice for HFrEF and patients who are in the first few years post-myocardial infarction, where guidelines recommend that all patients receive a β-blocker. However, even in these settings, such data would highlight those for whom achievement of target HR/target doses would be most beneficial and may improve adherence if patients know they are particularly likely to obtain benefit from the β-blocker. In other settings, such as complicated hypertension, ischemic heart disease, stroke prevention, or Marfan syndrome where there might be several options, these genetic data may highlight a β-blocker as the preferred agent for individual patients.

In contrast the current data for ADRB2 do not suggest such data are useful in the clinical setting as it relates to guiding the use of β-blocker therapy, even if the data are already available. While there are a few interesting studies, there is not sufficient consistency in the data to warrant consideration of its use in the clinical environment. While the current state of ADRB2 pharmacogenomics does not warrant clinical implementation, the most promising data are in the setting of heart failure and future research should be conducted in cohorts of participants with heart failure with preserved ejection fraction (HFpEF).

While there are fewer studies on the impact of the GRK5 Gln41Leu polymorphism, the existing literature are highly consistent and thus may have potential for clinical application. This polymorphism is most common in those of African ancestry, with the greatest evidence in patients with heart failure. The Leu41 allele appears to emulate the effects of β-blockade, since GRK5 Leu41 allele carriers have better outcomes than Gln41Gln carriers, and appear to have similar cardiovascular outcomes, regardless of β-blocker utilization. In contrast those with Gln41Gln have increased risk of poor outcomes and show significant benefit with β-blocker therapy. Similar to ADRB1, the clinical use of these data may not seem obvious, since β-blocker therapy is recommended for all patients with HFrEF. However, in patients with Gln41Gln it may also highlight those for whom it is essential to reach the target dose and/or may improve adherence among patients if they understand their genotype is one that derives significant benefit for β-blockers.

In summary, we suggest that CYP2D6 data may have some clinical utility for understanding dosing of commonly used β-blockers if the data are available, though the data are not sufficiently robust to warrant ordering a CYP2D6 pharmacogenetic test. We contend there may be clinical potential to add ADRB1 and GRK5 to the list of genes used clinically in pharmacogenetics, while ADRB2 does not have evidence, at least in the context of β-blockers, that would warrant its clinical use.

Article highlights.

• ADRB1, ADRB2, GRK5 and CYP2D6 have been extensively studied for their associations with efficacy, adverse effects and or clinical outcomes resulting from treatment with β-blockers. We summarize this literature and discuss the potential clinical utility of genotype-guided approaches to β-blocker treatment.

• CYP2D6 intermediate and poor metabolizers have significantly higher plasma drug concentrations than others at an equal dose. This may result in greater heart rate response, or effective β-blockade at lower than usual doses, but otherwise CYP2D6 genotype/phenotype has minimal clinical impact, including that it appears to have minimal to no impact on β-blocker–related adverse effects. If CYP2D6 genotype data are already available to the prescribing clinician, they may provide clinical utility given patients who are poor or intermediate metabolizers may achieve the target HR response at a lower than expected dose. Ordering a CYP2D6 test to guide β-blocker therapy is not recommended.

• Numerous studies suggest greater β-blocker efficacy in those with the ADRB1 Arg389Arg genotype, including for cardiovascular outcomes, the latter of which appear to be due to higher risk for cardiovascular events in this genotype group, which is offset with β-blocker therapy.

• Similarly, GRK5 Gln41Gln is associated with increased risk of adverse cardiovascular outcomes, which is offset by treatment with a β-blocker therapy, resulting in this genotype group being more responsive to β-blockers.

• Data generally do not suggest an important role of ADRB2 variability influences responses to β-blockers.

• Current evidence is consistent with the potential clinical use of ADRB1 and GRK5 genotype to guide β-blocker therapy and development of Clinical Pharmacogenetics Implementation Consortium guidelines should be considered.

Abbreviations

BP

Blood pressure

CAD

coronary artery disease

cAMP

Cyclic adenosine monophosphate

CPIC

Clinical Pharmacogenetics Implementation Consortium

DBP

Diastolic BP

HF

heart failure

HFrEF

heart failure with reduced ejection fraction

HR

heart rate

IM

CYP2D6 intermediate metabolizer

NM

CYP2D6 normal metabolizer

NYHA

New York Heart Association

PM

CYP2D6 poor metabolizer

SBP

Systolic BP

SNS

sympathetic nervous system

UM

CYP2D6 ultra-rapid metabolizer