CYP2D6 genetic polymorphisms are associated with metoprolol pharmacokinetics. Whether the clinical response to metoprolol is also affected remains uncertain.

Methods

We conducted a systematic review on the effects of CYP2D6 polymorphism on the clinical response to metoprolol. Searches were conducted using MEDLINE. Meta-analyses were performed on the impact of CYP2D6-inferred phenotypes on heart rate (HR) reduction, diastolic (DBP) and systolic (SBP) blood pressure reduction, average daily doses, all-type adverse events and bradycardia.

Results

Our qualitative assessment indicated inconsistent results in individual studies and endpoints, but CYP2D6 poor metabolizers (PM) generally presented a greater reduction in HR. The meta-analysis of 15 studies, including a total of 1146 individuals, found a reduction in HR of 3 beats/min (P = .017), and of SBP and DBP by 3 mmHg (P = .0048) for PM compared to non-PM individuals using similar metoprolol doses. Bradycardia appeared more frequent by 4-fold for PM, although significant heterogeneity was observed regarding bradycardia, which limits the scope of this finding.

Conclusion

Patients without any CYP2D6 metabolic capacities appear to have increased reduction in DBP, HR and SBP during metoprolol treatment and may be at a higher risk of bradycardia compared to patients with active CYP2D6 phenotypes. Further prospective data are required to determine whether CYP2D6 is associated with clinical events in patients treated with metoprolol, as well as to demonstrate the clinical utility of an individualized approach of prescribing metoprolol using CYP2D6-inferred phenotypes.

What is already known about this subject

CYP2D6 highly contributes to drug metabolism. It contains several genetic polymorphisms affecting its enzymatic activity.
Metoprolol pharmacokinetics are significantly affected between genotype-inferred phenotypes of CYP2D6, but the impacts on the clinical response of patients to patients have not been summarized.

What this study adds

This paper reviews the cumulative results of the effects of CYP2D6 genetic polymorphisms for patients on metoprolol treatment.
Patients possessing an inactive CYP2D6 phenotype have increased clinical effects and bradycardia with metoprolol, compared to those with an active CYP2D6 metabolic capacity.
More prospective data should be generated to suggest the implementation of a personalized approach to CYP2D6-mediated β-blocker therapy.

1 INTRODUCTION

The cytochrome P450 (CYP) 2D6 accounts for nearly 20% of all hepatic enzyme expression and is responsible for the metabolism of up to 25% of all registered drugs.¹ The Pharmacogenomics Knowledge Base regroups growing evidence that demonstrates a causal relationship between allelic variants of the CYP2D6 gene and variable response to certain drugs.² More than 100 variations, primarily single nucleotide polymorphisms and copy number variants, have been reported.¹ Such sequence variations can confer a null to increased CYP2D6 metabolic activity, and the effects on drug metabolism and pharmacokinetic profiles have been repeatedly documented across multiple drug classes.³ Guidelines have been elaborated to classify the CYP2D6 metabolic activity of individuals into 4 genetically-inferred phenotypes: ultrarapid metabolizers (UM), extensive, or normal, metabolizers (EM), intermediate metabolizers (IM) and poor metabolizers (PM).^3-6 The predicted prevalence of CYP2D6 genotype-predicted phenotype is estimated to be 1–21%, 67–90%, 0.4–11% and 0.4–5.4%, for UM, EM, IM and PM respectively, depending on the population studied.⁷

The selective β₁-antagonist metoprolol is one of the most commonly used cardiovascular drugs to treat conditions such as heart failure, hypertension and atrial fibrillation.^8-10 Approximately 80% of metoprolol is metabolized by CYP2D6,¹¹ whose metabolites exert negligible pharmacological activity.¹² An earlier meta-analysis has confirmed the significant effect of CYP2D6 genotype on the pharmacokinetics of metoprolol.¹³ However, whether this pharmacogenetic effect translates into differences in clinical response to metoprolol treatment is uncertain. This is largely attributable to the fact that previous studies only had small sample sizes. Hence, the objective of this article was to systematically review and summarize the impact of CYP2D6 polymorphisms on the clinical response to metoprolol.

2 METHODS

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement.¹⁴

2.1 Data sources and search strategy

A series of comprehensive searches in the MEDLINE database were conducted on articles written from 1946 to September 2018 to investigate the association between CYP2D6 and the pharmacodynamics of metoprolol. The searches were restricted to articles written in English or in French, while also including ahead-of-print, in-process and nonindexed records. The following terms were used to guide and refine the iterative search process: metoprolol; cytochrome P-450 2D6; genomics; genes; pharmacogenetics; polymorphism; genetic; single-stranded conformation; amplified fragment length polymorphism analysis; single nucleotide; restriction fragment length; phenotype; genotype (Table S1). Database updates were activated to identify newly published articles. The reference lists of the articles retrieved were manually searched to find studies that did not appear based on our strategy.

2.2 Eligibility criteria and study selection

Studies to be included were considered for this review irrespectively of study design, follow-up duration, metoprolol dosing regimen, and drug formulation. Studies evaluating drug interactions between metoprolol and other CYP2D6 substrates were excluded, unless patients took metoprolol without or before administration of a potentially interacting agent. In such cases, only the relevant baseline information was included. Previous reviews and meta-analyses were also rejected after having searched their reference lists.

In terms of subject inclusion, participants who were healthy, as well as patients with cardiovascular and noncardiovascular diseases, were included. Studies involving special populations such as twins, or pregnant and/or lactating women were excluded. No constraints were applied on the use of polypharmacy regimens. Publications where participants had not been genotyped, or studies that had classified patients primarily according to their pharmacokinetic profiles using probe drugs, or to drug responses without consideration to genetic information, were also excluded.

2.3 Data collection and data items

Eligibility assessment and data collection was performed independently by the first and second authors. Differences were resolved through discussions and consensus, and any remaining issues were settled with the senior author. Results from the database searches were ordered alphabetically according to the names of the first author to easily identify duplicates, which were removed afterwards. A primary screening with the titles and abstracts of all the articles was made. Articles considered eligible for review were entirely assessed and the data were collected in a MS Excel spreadsheet. The following details were included: year of publication; first author; title of study; presence of duplicates; study design; sample size and ethnicity; patients' diseases; investigated medications and concomitant drugs; dosing regimens; follow-up; outcome measures; effects on reviewed outcomes; genotyped single nucleotide polymorphisms; and phenotype assignment.

The reviewed parameters between all phenotypes were limited to changes from baseline at study enrolment of heart rate (ΔHR), diastolic (ΔDBP) and systolic (ΔSBP) blood pressure, mean daily metoprolol doses, and frequency of all-type adverse events (AE) and bradycardia. End-of-study values for each endpoint were collected. Mortality and hospitalizations, or other composite criteria reflecting death and morbidity were also reviewed.

2.4 Genotype-inferred phenotype assignment

We used the inferred phenotype assigned by the authors, according to their genotype classification. In studies having reported genotypes only, we attributed inferred phenotypes in accordance with the current Dutch Pharmacogenetics Working Group (DPWG) recommendations for groups which had not been assigned one. The DPWG annotations on metoprolol and CYP2D6 are referred by the Pharmacogenomics Knowledgebase website. Thus, subjects were categorized as UM if they carried >2 fully functional alleles (*1, *2, *33, *35, *39, *53), EM if they possessed 2 fully functional alleles or 1 fully functional and a semi-active allele (*9, *10, *17, *29, *36, *41, *53, *54, *59, *72, *84), IM if they had only 1 fully functional allele and a null allele (*3–*8, *11, *12, *14–*16, *18–*21, *38, *40- *42, *44, *51, *56, *57, *60, *62, *68, *69, *92, *96, *99–*101, *114) or 2 semi-functional alleles, and PM were determined to have 2 null alleles. All known CYP2D6 polymorphisms are indicated on the Pharmacogene Variation Consortium online database.¹⁵

2.5 Statistical methods and meta-analyses

Statistical analyses were carried using primarily with the meta package on the R software version 3.5.1 (The R Foundation for Statistical Computing). Random-effects models for meta-analyses were chosen. Irrespective of their designs, studies having reported quantitative values for the assessed parameters were included in our meta-analyses. For continuous variables (i.e. average daily doses, ΔDBP, ΔHR, ΔSBP), data were reported as mean difference, alongside confidence intervals (CI) set at 95% of the distribution (Table S2). For such parameters, random-effects models were conducted using the Hartung–Knapp adjustment.¹⁶ For binary outcomes (i.e. all-type AE, bradycardia), odds ratios (OR) were retrieved or calculated (Table S3). Summary effects sizes were generated by means of a restricted maximum likelihood method with an inverse-variance estimator. Other models have recently been brought forward to increase probability coverage for binary outcomes.^{17, 18} Recent simulations have shown that they do not considerably outperform restricted maximum likelihood estimations in random-effects meta-analyses with few studies.¹⁹ Nevertheless, considering the criticism that has been made surrounding the application of the Hartung–Knapp adjustment,^{20, 21} we performed sensitivity analyses to compare the results to the conventional DerSimonian–Laird approach in the case of continuous variables.

Other statistical analyses were made via Comprehensive Meta-Analysis, version 2.2.064 (Biostat, Englewood, NJ, USA). The various ways of reporting results from the included studies required that certain values be pooled. For example, if individuals carrying different genotypes but having the same inferred phenotype were reported separately in some articles, their respective values were weighed and aggregated together, according to the DPWG guidelines, generating single values of mean and standard deviation (SD) for each phenotype. The same weighing process was applied to calculate means and SDs for pooled phenotypes. Whenever needed, reported values were transformed into common units of measure. In some studies, only medians, interquartile ranges, and/or full ranges with sample sizes were reported. To transform those parameters into values that could be standardized and properly integrated into our models, we used the formulas proposed by Wan et al.²² Briefly, this approach utilizes the sample size, median, maximum and minimum values, or the interquartile range of a given distribution to estimate the mean and SD of a dataset. Values obtained from Wan et al.’s method²² are displayed in Table S4 (Supporting Information), and all values used in our meta-analyses for daily doses are listed in Table S5. If warranted, the estimated parameters were pooled as described earlier. Even though our primary comparison of interest was to compare PM individuals with those presenting any CYP2D6 activity (i.e. UM, EM, IM), we performed pairwise comparisons between single phenotypes as well.

2.6 Quality and risk of bias

The first and second authors (M.M., M.K.) performed an evaluation on the quality and risk of bias in each study. Multiple critical evaluation tools for study biases were available from different consortia and institutes, but their qualitative assessment methods were only partially adapted for pharmacogenetic studies. Thus, the following quality components were inspected without referring to general critical evaluation tools: study design, randomization, appropriateness of group comparisons, baseline characteristics, patients' diseases and medications, dosing regimens and concomitant medications, and inferred phenotype assignment from genotype. However, we did not exclude studies based on these criteria.

2.7 Ethical review

Ethical approval was unnecessary for this systematic review and meta-analysis, because only previously published data were retrieved. No recruitment of human participants was done, and personal data were not needed to perform this review.

3 RESULTS

3.1 Summary of systematic review

Figure 1 presents the article selection process. The comprehensive search generated 75 results from the MEDLINE database. Seven additional record from reference lists were added, along with 5 from manual searches, totalling 87 results. Of these, 2 articles were removed due to language barriers (German, Chinese), along with 1 off-topic systematic review, 2 earlier meta-analyses, and 1 editor correspondence. After removal of duplicates (n = 10), 13 articles were excluded based on their titles and 11 on their abstracts. The remaining 47 articles were completely assessed for eligibility. Twenty-five additional articles were then excluded because they did not meet all the eligibility criteria: 5 studies did not perform any genotyping, 4 studies did not present any reviewed parameters, 1 study was a retrospective assessment of 2 prior studies, and 1 study did not classify subjects according to their CYP2D6 genotype-inferred phenotype. Fifteen articles were rejected for presenting pharmacokinetic parameters only. The complete list of fully assessed and excluded articles is available as Supporting Information (Table S6). In total, 21 studies were included for presenting clinical parameters, 15 of which that were used in our meta-analyses.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Flow chart of the review process

Table 1 contains the main characteristics of the included studies with regards to their clinical parameters. Study designs, types of populations and diseases, medications and dosing regimens, durations and follow-ups, inferred phenotype classification, and the summarized results of every parameter reviewed are indicated. Values in the table were all explicitly reported in the original articles.

Table 1. Summary of included studies evaluating the effects of CYP2D6 phenotypes on the clinical response to metoprolol

Authors (y)	Study design	Population and diseases	Follow-up	Genotype-inferred phenotypes	Medication and dosing regimens	Results of reviewed markers
Batty et al. (2014)²⁸	Retrospective substudy of the metoprolol CR/XL randomized Intervention trial in congestive heart failure (MERIT-HF)	718 patients from a subpopulation of the MERIT-HF trial with NYHA II-IV, and with LVEF≤40% 313 patients included in the analyses	Mean 339 d	Patients taking metoprolol: 189 EM: 1/1 112 IM: 1/4 12 PM: 4/4	Low initial dose metoprolol CR/XL (12.5 mg/d for patients NYHA III-IV, 25 mg/d for patient NYHA II); progressive, bi-weekly, upward titration for a target dose of 200 mg/d Individual dosing regimens freely adjustable to the judgement of the investigator	• No change in dosing profiles and mean daily doses between all genotypes (P = .29)
						• No differences between all genotypes in achieving a 200 mg/d target dose (P = .25)
						• Risks of discontinuation nonsignificant for the 3 groups compared to placebo:
						- EM: OR = 0.55, 95%CI = [0.25–1.19], P = .12
						- IM: OR = 0.96, 95%CI = [0.49–1.91]; P = .92
						- PM: OR = 0.59, 95%CI = [0.05–7.43]; P = .81
						• Significant decreases in resting HR at wk 4, 6 and 8 for IM and PM compared to EM from their respective baseline values (all P < .05), but not at 2 wk and 3 mo
						• No difference of resting HR between IM and PM during all follow-up period, but greater differences in HR decreases from baseline for IM compared to EM (P = .02), and for PM compared to EM (P = .04) at maximum doses. No differences after 6 mo, which corresponded to the longest time of measurement.
						• No difference for SBP between all genotypes (P = .56)
						• Significant decreases of DBP for IM compared to EM after 4 wk and 3 mo (P = .03; P = .004), but not at wk 2, 6, 8, and after 6 mo, and between PM and EM after wk 2 and 4, and 3 mo (P = .04; P = .04; P = .03), not at wk 6 and 8, and at 6 mo
						• Significant decreases of DBP for IM and PM compared to EM at maximum doses (P = .04; P = .01)
						• Genotype conferred no effect on all-cause mortality and hospitalization; combined endpoints were similar in all 3 groups compared to placebo:
						- EM: OR = 0.62; 95%CI = [0.37–1.06], P = .08
						- IM: OR = 0.71; 95%CI = [0.36–1.45], P = .36
						- PM: OR = 0.80; 95%CI = [0.11–5.77], P = .83
Baudhuin et al. (2010)³⁰	Retrospective, nonrandomized cohort study	93 patients with CHF for ≥6 mo, NYHA I-IV, LVEF ≤45%	Total 14 mo	Patients taking metoprolol: 5 UM: 1/2A, 1/2AxN 1 EM: 1/1 10 IM: 1/4, 1/5, 2/4, 2/9, 4/9 3 PM: 3/10, 4/4, 4/6	25 mg/d metoprolol IR, doubling every 10 d until MTD achieved; target dose 200 mg/d, or 6.25 mg/d carvedilol, doubling every 10 d until MTD achieved; target dose of 25 mg/d if BW ≤ 85 kg, and 50 mg/d if BW > 85 kg	Phenotype comparisons for patients taking metoprolol:
		93 patients with CHF for ≥6 mo, NYHA I-IV, LVEF ≤45%				• Nonsignificant difference in metoprolol dosage and MTD between all genotypes (data not shown)
		19 patients treated with metoprolol
Bijl et al. (2009)³⁷	Prospective, population-based cohort study (Rotterdam study)	6348 β-blocker users at baseline (2878 at final follow-up); 199, 243, 205, and 196 metoprolol users at baseline, first, second and third follow-up	Total 14 y	ΔHR measurements: 451 EM: 1/1, 255 IM: 1/4, 34 PM: 4/4 ΔDBP/ΔSBP measurements (same genotypes): 496 EM 276 IM 37 PM Phenotypes attributed according to DPWG guidelines	Dosage titrated on hypertensive effect and dependant on CYP2D6 genotype (data not shown) Comedications tolerated during the study, including strong and weak CYP2D6 inhibitors Metoprolol formulation not mentioned	• HR, adjusted for age and sex, was significantly decreased in PM compared to EM (P < .0001), and in IM compared to EM (P = .013)
						• Metoprolol conveyed more risks of bradycardia for PM compared to EM (P = .0014)
						• Mean difference in HR significantly decreased for IM and PM compared to EM that were all β-blocker users (P = .012; P < .001)
						• Mean difference in DBP was significant for PM compared to EM that were metoprolol users (P = .045), yet no changes were seen for SBP
						• Within-person change for HR decreased significantly only for PM compared to EM (P = .019), but not between IM and EM (P = .06). Discontinuation did not lead to significant differences between any phenotype.
						• DBP did not change for every phenotype when starting (n = 315) or discontinuing (n = 161) metoprolol therapy
						• Differences in mean dose was observed for CYP2D6-mediated β-blockers, which consisted mainly of metoprolol (P = .03)
Fedorinov et al. (2018)³¹	Prospective, observational cohort study	201 patients of Russian or Yakut ethnicity with CHD on routine β-blocker therapy (atenolol, bisoprolol, metoprolol)	Specific duration not mentioned	63 EM: 1/1, 1/10 16 IM: 1/4	Patients receiving metoprolol: Titrated dose between 12 and 150 mg/d Metoprolol formulation not mentioned	• Titrated doses between the 39 patients genotyped as 1/1 carriers and the 16 other 1/4 patients did not vary significantly, although a trend was observed (P = .071)
						• The differences in titrated doses were even less significant when comparing the 39 1/1 with the 24 1/10 carriers (P = .240)
		79 patients receiving metoprolol
Fux et al. (2005)³²	Multicentre, prospective, open-label, nonrandomized study	121 subjects with various cardiovascular and noncardiovascular diseases (HT, CHD, arrhythmias, migraine, anxiety disorder) not previously treated with metoprolol	6 wk	5 UM: More than 2 fully functional alleles 91 EM: 1/3-10, 1/41 21 IM: 8-10/8-*10, 8-10/3-8, 8-10/*41, 8-10/3-8 4 PM: 3-8/3-8	Any daily dose allowed, with any dosing scheme of metoprolol CR/XL Individual dosage freely adjustable based on clinical grounds	• Mean daily doses did not differ between all genotypes (P = .78)
						• Tendency toward more frequent cases of cold extremities observed in the PM + IM vs EM + UM (P = .56, CI95% = [1.03–14.3]). Sexual dysfunction lower for PM + IM vs EM + UM (P < .05)
						• EM and IM had significant SBP, DBP, and HR reductions compared to baseline (all P < .01), not UM or PM
						• Mean reductions between all phenotypes were nonsignificant for SBP (P = .92), DBP (P = .88), and HR (P = .30)
Goryachkina et al. (2008)²³	Prospective, nonrandomized, open-label study	187 Caucasians hospitalized after AMI occurrence	10 d	Patients kept for clinical parameter analyses: 5 UM: 1/1xN 108 EM: 1/1, 1/3, 1/4, 1/4xN, 1/4, 1/10, 10/4 2 PM: 4/4	75 mg/d ±38 mean metoprolol IR or CR/XL	• Median resting HR at discharge (15–20 d after admission) were different between groups with 0, 1, 2, or 3 functional alleles (P = .01). Changes occurred between d 7–10
Goryachkina et al. (2008)²³	Prospective, nonrandomized, open-label study	115 patients genotyped and included for analysis, including the 17 depressed AMI patients used for the previous study; Goryachkina et al. (2008a)	10 d		Use of concomitant drugs accepted and considered, with multiple drug classes coadministered other than β-blockers
Goryachkina et al. (2008)³⁴	Prospective, nonrandomized, open-label, study	34 patients taking metoprolol at baseline, with 17 depressed AMI patients, and 17 nondepressed control patients	10 d	17 EM: 1/1, 1/3, 1/4	Titrated metoprolol IR or CR/XL on d 1	Baseline measurements:
					Coadministration of 20 mg/d paroxetine on d 2–8, or placebo	• Nonsignificant differences in metoprolol dose allocation between homozygous and heterozygous genotypes before administration of paroxetine (not shown)
					Use of concomitant drugs accepted
Hamadeh et al. (2014)²⁶	Multicentre, prospective, open-label, sequential monotherapy, randomized trial (PEAR2 -Pharmacogenomic evaluation of antihypertensive responses 2)	218 patients with uncomplicated HT starting metoprolol therapy	8 wk	8 UM: 1/1xN, 1xN/2, 1/2N, 2xN/10, 2xN/2 184 EM: 1/1, 1/2, 1/3, 1/4, 1/5, 1/6, 1/10, 1/17, 1/41, 1/4xN, 1/10xN, 4/17xN, 2/2, 2/4, 2/4xN, 2/5, 2/10, 2/17, 2/41, 10/17, 10xN/41, 17/17, 17/41, 17/41 15 IM: 3/17, 4/17, 4/10, 4/41, 5/10, 5/17 11 PM: 3/4, 4/4, 4/5, 4/6	Starting dose 100 mg/d metoprolol IR for 2 wk, titration up to 200 mg/d or MTD for 6 wk if blood pressure > 120/70 mmHg and HR > 55 BPM Treatment interrupted with 50 mg decreases every 2 d until stopped	• Significant differences in changes for HR from baseline values when comparing all 4 phenotypes (P < .0001), but not for SBP (P = .91) and DBP (P = .37)
						• No difference in AE frequencies between all genotypes, including bradycardia (P = .87). No difference in the type of AE observed (P = .41)
						• No difference in metoprolol mean doses between all genotypes (P = .77). Only 5% of all subjects (9 EM, 1 IM) did not reach the maximum recommended dose of 200 mg/d
Hamelin et al. (2000)⁴²	Prospective, randomized, double-blind, placebo-controlled, 2-period crossover study	16 healthy male subjects	Blood and urine samplings collected 48 h after metoprolol dose	10 EM: 1/1, 1/5 6 PM: 4/4, 4/5	100 mg single dose metoprolol IR on d 3 Coadministration of 50 mg/d diphenhydramine, or placebo for 5 d, for 2 periods	Baseline measurements:
						• Exercise HR profiles were more decreased for PM than for EM (P < .0009) and persisted from 3 to 12 h postdose. After 12 h, HR was decreased by 14% ± 2 for PM taking only metoprolol, compared to 5% ± 7 for EM.
						• Decrease in SBP during exercise initially more pronounced for PM than for EM compared to baseline (P < .0001). After 12 h, PM still had a decrease of 14–17%, whereas EM returned to baseline values.
						• Decrease in HR more pronounced between 8 and 12 h for PM than for EM (P < .05)
Kirchheiner et al. (2004)⁴¹	Prospective, nonrandomized, controlled study	29 heathy volunteers	Blood samples collected for 24 h after single dose	12 UM: 1-2/2x1, 1- 2/2x2, 1-2/2x35, 9- 10/2x1, 9-10/2x2, 9 -10/2x35, 35/2x1, 35/2x2, 35/2x35, 41/2x1, 41/2x2, 41/2x35 13 EM: 2 any combination of alleles 1, 2, 9, 10, 35, and 41 4 PM: 3-6/3-6	100 mg single dose metoprolol IR	• Marginal effect on resting HR between all phenotypes, but significant (P = .03)
						• Significant reductions for PM compared to EM (P = .027) and UM (P = .017) when evaluating exercise-induced heart rate differences, but not between EM and UM phenotypes
						• No significant effect on resting DBP and SBP between any phenotypes
Koytchev et al. (1998)³⁹	Prospective, randomized, 3-period cross-over, bioequivalence study	36 healthy Caucasians	Total 15 d	16 EM: 1/1 6 IM: 1/*4	200 mg/d metoprolol CR/XL	• Resting HR decreased more significantly for IM than for EM compared to baseline (P < .05)
Koytchev et al. (1998)³⁹		22 subjects included for pharmacogenetic analyses	Total 15 d	Phenotypes attributed according to DPWG guidelines	200 mg/d metoprolol CR/XL	• No significant differences for resting SBP and DBP between IM and EM, even though a trend was observed.
Luzum et al. (2017)²⁷	Retrospective chart-review from the Ohio State University-Coriell personalized medicine collaborative	102 patients with systolic HF 33 patients treated with metoprolol	Median 5.3 y Median 1.7 y for the duration of treatment for patients taking metoprolol	Comparison of 3 classification schemes for genotype-inferred phenotypes; *4 allele carriers: 23 EM, 9 IM, 1 PM CPIC: 28 EM, 4 IM, 1 PM DPWG: 21 EM, 11 IM, 1 PM	Patients taking metoprolol: Titration of normalized* metoprolol CR/XL (n = 32) or IR (n = 1) *normalization of metoprolol daily maintenance dose defined as [carvedilol dose x 4 = metoprolol dose] Comedications permitted, including CYP2D6-interacting drugs	• CYP2D6 genotype-inferred phenotype significantly associated with decreased maintenance dose of metoprolol for both single and double CYP2D64* allele carriers compared to noncarriers (P = .023, OR = 0.13; 95% CI = [0.02–0.75])
						• Significant trend also observed for the DPWG classification (P = .014, OR = 0.18; 95%CI = [0.03–0.66])
						• Nonsignificant differences in dosing regimens observed for the CIPC classification (P = .12, OR = 0.19; 95% CI = [0.24–1.58])
						• Regardless of the method of phenotype classification (DPWG, CIPC, *4 polymorphism), none of the PM or IM subjects achieved the recommended target metoprolol dose for treating systolic HF
Nozawa et al. (2005)³³	Prospective, nonrandomized, comparative study	72 patients with various cardiovascular diseases (AF, IHD, HT, hypertrophic cardiomyopathy) on routine treatment with metoprolol or bisoprolol for ≥2 wk 38 patients treated with metoprolol	Baseline values taken 3 and 4 h after dosing	30 EM: 1/1, 1/10 8 IM: 10/10 Homozygous and heterozygous EM analyzed separately	Patients taking routine metoprolol IR Initiation of metoprolol or bisoprolol treatment at least 2 wk before the beginning of study Dosing regimens not titrated during study	Baseline measurements for patients taking metoprolol:
Nozawa et al. (2005)³³	Prospective, nonrandomized, comparative study		Baseline values taken 3 and 4 h after dosing			• Dosing regimen differences nonsignificant between all 3 phenotypes
Rau et al. (2002)²⁵	Prospective, cross-sectional study	91 Caucasians with various cardiovascular diseases (HT, CAD, HF, MI), ranging from NYHA I-III	Median 12.6 mo	52 EM: 1/1, 1/2, 1/9, 1/41, 2/2, 2/41 31 IM: 1/0, 2/0, 9/0, 10/0, 41/0, 41/41 8 PM: 0/0 0 includes null alleles 3, 4, and 6 Phenotypes attributed according to DPWG guidelines	47.5 mg/d median metoprolol CR/XL Dose titration based on clinical grounds, individual daily drug doses constant for at least 2 wk before blood sampling 78 patients treated with concomitant cardiovascular drugs	• No difference and wide variability between groups in daily drug doses (P = .79)
Rau et al. (2009)²⁴	Prospective, nonrandomized, longitudinal study	91 naïve β-blocker patients with various cardiovascular diseases (AF, CAD, HF, HT), all treated with metoprolol Different cohort from Rau et al. (2002)²⁵	90 d	63 EM: 2 fully functional, or 1 fully and 1 partly functional alleles 8 IM: 0/10, 0/41 17 PM: *0/0* 0 includes null alleles 3, 4, and 6 Patients regrouped as PM (17) and non-PM (71)	47.5 mg/d mean dose of metoprolol CR/XL for PM and non-PM Titration allocated on medical grounds Other cardiovascular comedications allowed during study	• Dose considered too low to witness any disparities in side effects. However, 14 PM out of 71 non-PM subjects (20%) had doses increases during treatment (P < .004), compared to 1 in 17 (6%) for PM subjects (P = .99)
						• Average daily dose was not significantly different for the 2 groups (P = .73)
						• PM had significantly higher decreases compared to non-PM in resting HR (P = .013), and resting DBP (P = .007), but no significant differences observed for resting SBP (P = .36)
						• No difference of all-type AE between groups (P = .55), but significantly more occurrences of bradycardia among PM (log rank P = .0001)
						• More PM reduced their resting HR to <60 BPM and had bradycardia compared to non-PM (P < .001)
						• Differences in plasma concentrations were associated with significantly and persistently enhanced drug effects between phenotypes
Sharma et al. (2005)⁴³	Prospective, randomized, double-blind, placebo-controlled, 2-period crossover study	20 healthy premenopausal women	Total 2 mo	16 EM: 1/1, 1/3, 1/4, 1/5 4 PM: 4/4, 4/5	100 mg single dose metoprolol IR on d 3 Coadministration of 50 mg/d diphenhydramine, or placebo for 5 d, for 2 periods	Baseline measurements:
						• No significant difference between the profile of PM and EM in terms of stroke volume index (P > 0.05). However, the exercise cardiac index appeared different (P = .009)
						• Exercise HR and exercise rate-pressure products were significantly different between PM and EM (P < .05)
Sharp et al. (2009)²⁹	Multi-centre, prospective, nonrandomized, open-label study	52 patients with systolic HF (NYHA II-IV, LVEF ≤45%) not previously on metoprolol therapy	Blood sampling performed at least 2 wk after maximum tolerated dose achieved Specific duration not mentioned	27 EM: 1/1 22 IM: 1/4 3 PM: 4/4 Phenotypes attributed according to DPWG guidelines	23.75 mg/d median initial metoprolol CR/XL, doubled every 2–4-wk intervals titration occurred until MTD of 190 mg/d was achieved (median MTD 95 mg/d) Other treatments permitted during the study	• Nonsignificant differences between EM and IM for percentage change of HR (P = .71), and between all 3 phenotypes for percentage change of SBP (P = .79), and DBP (P = .40)
						• Nonsignificant AE-related withdrawals between all phenotypes (P = .79)
						• Nonsignificant difference in the number of patients who achieved MTD between all phenotypes (P = .26)
						• Proportions not statistically different between patients with 1 or 2 functional alleles that improved condition with metoprolol therapy (P = .71)
Terra et al. (2005)³⁵	Multi-centre, prospective study	61 naïve β-blocker users, with symptomatic ischaemic or nonischaemic HF (NYHA II-III, LVEF ≤40%) 51 patients genotyped	10 wk	37 EM: 1-2/1-2, 1- 2/3, 1-2/4, 1-2/6, 1-2/9, 1-2/10, 1- 2/17, 1-2/29, 1- 2/41 10 IM: Any combination of 2 reduced function alleles 9, 10, 17, 29, 41, or 1 reduced function allele with another null allele 3, 4, 6 4 PM: 3-4/3-4, 3- 4/6, 6/*6	25 mg/d for patients NYHA II, and 12.5 mg/d metoprolol CR/XL for patients NYHA III as starting dose. Titration biweekly until 200 mg dose achieved, or MTD Other comedications allowed during the study	• No differences between all groups in cardiac decompensation rates, which was the composite criteria of death, HF hospitalization, increase in other HF medications, or need to discontinue metoprolol (P = .61)
Terra et al. (2005)³⁵	Multi-centre, prospective study		10 wk			• Final daily dose differences nonsignificant (P = .45)
Wuttke et al. (2002)³⁶	Retrospective, pharmacovigilance census by means of standardized questionnaires	24 German subjects treated with metoprolol	22 mo for recruitment	8 EM: 1/1, 1/2, 1/41 7 IM: 1/4, 1/5, 2/3, 2/41, 41/4 9 PM: *3–5/*4–6	47.5 mg/d median metoprolol IR dose at time of AE Dosing regimens titrated and based on clinical grounds	• Patients who had AE tended to be 5 times more likely to be PM, yet the till AE occurrence, or the time required for dose reduction were nonsignificant between PM and non-PM (P = .29; P = .26)
				Patients divided into PM (9) and non-PM (15)		• Nonsignificant differences in daily metoprolol doses and between EM and PM (P = .51)
				Patients divided into PM (9) and non-PM (15)		• Trend observed for bradycardia occurrences when comparing PM to non-PM, but nonsignificant (P = .089)
Yuan et al. (2008)⁴⁰	Prospective, observational, randomized study	300 Chinese Han patients with essential HT and homozygous Arg389 carriers for the ADRB1 gene 276 patients included in the analyses	8 wk	Group A: 40 EMA: 1/1, 1/2, 2/2 43 IMA: 1/5, 2/5, 1/10, 2/10 60 PMA: 5/5, 5/10, *10/10 Group B (same genotypes): 38 EMB 27 IMB 68 PMB	Subgroups A (3), fixed dose: 100 mg/d metoprolol (PMA, IMA, EMA) Subgroups B (3), ascending doses: PM: 25 mg/d (PMB) IM: 50 mg/d (IMB) EM: 100 mg/d (EMB)	• Significant difference in resting SBP observed between IMA and EMA (P = .027), and between PMA and EMA (P = .046)
						• Differences in resting DBP decreased more for PMA than for EMA compared to their respective baseline values (P = .022)
						• No difference of resting SBP and DBP between EMB, IMB and PMB (statistical analysis not shown)
						• Resting HR for all 6 groups were significantly decreased compared to before treatment; differences between phenotypes not assessed, but values of HR reduction and final HR were all similar. Ascending doses were given according to CYP2D6 phenotype.
Zineh et al. (2004)³⁸	Prospective, nonrandomized study	50 patients with uncomplicated HT	4 wk	42 EM: Any combination of alleles 1, 2, 9, 10, 17, 29, 41, 45, and 46 4 IM: 3-6/9, 3- 6/10 ¸3-6/17 ¸3- 6/29 ¸3-6/41 ¸3- 6/45, 3-6/46 4 PM: 3-6/3-6 EM patients initially divided into subgroups (high EM, medium EM, low EM), but subsequently regrouped to create a common general phenotype for quantitative analyses.	Start 100 mg/d metoprolol, titration doubled doses weekly until 1 of the following criteria was achieved: Clinical DBP was reduced to ≤90 mmHg or till patient had side effects that precluded upward titration of drug, or 400 mg/d achieved Other comedications allowed during study, including CYP2D6 inhibitors Metoprolol formulation not mentioned	• Based on quartile data of S-metoprolol AUCs and CYP2D6 activity scores, no difference was observed for general and dose-limiting AE
						• No differences for changes in DBP between all S-metoprolol AUC quartiles, but differences in DBP not reported directly for any phenotype
						• Total daily metoprolol doses not significantly different between phenotypes (P = .74)

AE; all-type adverse events AF, atrial fibrillation; AUC, area under the concentration–time curve; AMI, acute myocardial infarction; BPM, beats/min; CAD, coronary artery diseases; CHF, congestive heart failure; CPIC, Clinical Pharmacogenomics Implementation Consortium; CR/XL, controlled-release/extended-release metoprolol formulation; DBP, diastolic blood pressure; DPWG, Dutch Pharmacogenomics Working Group; EM, extensive metabolizers; HF, heart failure; HR, heart rate; HT, hypertension; IHD, ischaemic heart diseases; IM, intermediate metabolizers; IR, immediate-release metoprolol formulation; LVEF, left ventricular ejection fraction; MI, myocardial infarction; MTD, maximum tolerated dose; NYHA, New York Heart Association; PM, poor metabolizers.

3.2 Meta-analyses of clinical parameters

3.2.1 Average daily doses

Fifteen studies were available to review the qualitative differences between the daily titrated metoprolol dosing across phenotypes.^23-37 Titration schemes varied and starting doses differed based on diagnosis. Forced titration by ascending doses was initiated according to fluctuating time periods and on patients' needs until a maximum tolerated dose or the final target dose had been achieved, which averaged between 47.5 and 200 mg daily. Generally, studies observed that PM participants did not reach target dose as often as non-PM, but those differences were not consistent across studies.

Our meta-analyses of daily metoprolol drug dosing differences between CYP2D6 inferred phenotypes included 8 studies and 789 individuals (Figure 2).^{24, 25, 28, 29, 32, 35, 36, 38} Average daily dose values were calculated using Wan et al.’s formulas for 5 studies.^{24, 25, 29, 32, 36} The difference in means (mg/d) was nonsignificant when comparing PM and non-PM (mean difference = −2.12 [95%CI: −14.26 to 10.03]; P = .202). No significant heterogeneity was observed in our main analysis, with I² reaching 28% (P = .20). For pairwise comparisons, all phenotypes had similar dosages.

3.2.2 Heart rate reduction

A total of 9 studies assessed the HR reduction at rest between the genotype-inferred phenotypes.^{24, 26, 28, 29, 32, 37, 39-41} Almost all studies showed significant differences between at least 2 phenotypes. We also found 3 studies in which exercise-induced HR differences had been measured,^41-43 all of which showed significant differences (Table 1).

A total of 5 studies comprising 750 individuals were included in the meta-analysis (Figure 3).^{24, 26, 28, 32, 41} We observed that PM had a significantly greater reduction compared to non-PM when assessing the mean differences in beats/min (mean difference = 3.16 [95%CI: 0.94 to 5.37]; P = .017). After removal of 1 fixed-dose study, the same conclusions were obtained when comparing PM against non-PM (P = .044). No heterogeneity was observed in our main analysis (P = .89). The only 2-by-2 comparisons that reached significant differences were EM vs PM (P = .029), and UM vs PM (P = .048).

3.2.3 Blood pressure reduction

Eight studies published data on resting ΔSBP and ΔDBP between CYP2D6 genotype-inferred phenotypes.^{24, 26, 28, 29, 32, 37, 39-41} In all the studies reviewed, patients were initially β-blocker naïve. Our systematic review showed inconsistent results across studies for both parameters. Only 1 of the studies that had examined exercise-induced HR variations also reported values for BP reduction after exercise.⁴² In that study, 12 hours after a single 100 mg metoprolol dose, PM still had a decrease of 14 to 17%, whereas EM had returned to baseline.

Five studies including 842 individuals suitable for meta-analysis were analysed to determine the effect of CYP2D6 genotype-derived phenotypes on resting ΔSBP (Figure 4).^{24, 26, 28, 32, 40} Our results indicated a significantly greater reduction (mmHg) for PM against non-PM (mean difference = 2.88 [95%CI: 1.47 to 4.29]; P = .0048), which was also observed for titrated-dose studies only (P = .031). No significant heterogeneity was observed in this meta-analysis (P = .99). EM against PM was the only pairwise comparison where statistical significance was achieved (mean difference = 4.15 [95%CI: 0.18 to 8.12]; P = .030), which was preserved when removing the fixed-dose study (P = .037).

As for ΔDBP, the same 5 studies were included in our meta-analyses (Figure 5). Significant results were observed between non-PM and PM (mean difference = 2.93 [95%CI: 1.53 to 4.32]; P = .0043). Again, no significant heterogeneity in this meta-analysis, with I² = 0% (P = .91). By excluding the 1 fixed-dose study, results remained significant (P = .021), with PM showing greater reduction than other phenotypes. IM and PM showed significantly greater reductions compared to EM when including all studies (EM vs IM: P = .014; EM vs PM: P = .014). Titrated-dose studies also showed that the phenotype with less metabolic capacity had a greater ΔDBP, in both comparisons. For fixed-dose studies, comparisons involving EM and IM showed no statistical differences (P = .150). As was the case for ΔSBP, our quantitative analysis also showed no differences in ΔDBP between UM and other phenotypes.

3.2.4 All-type adverse events and bradycardia

In total, 6 studies reported the frequency or the number of all-cause adverse reactions, either dose-limiting or not.^{24, 26, 29, 32, 36, 38} Five studies specifically monitored the frequency of bradycardia,^{24, 26, 28, 36, 37} which was uniformly defined as a HR < 60 beats/min. Two studies reported symptomatic bradycardia and all bradycardia events, while 2 reported bradycardia without specifying whether these events were symptomatic. All studies reported the use of multiple doses.

Only 3 studies including 356 individuals were pooled to evaluate the frequency of all-type AE.^{24, 26, 38} Considerable variability was observed for each individual study regarding confidence intervals. Neither PM or non-PM were more prone to adverse reactions due to their CYP2D6 phenotype (OR: 0.76 [95%CI: 0.33 to 1.74]; P = .514; Figure 6). No heterogeneity was found in this analysis (I² = 0%, P = .59) Pairwise comparisons also did not yield significant results for any group.

Values from 4 studies regrouping 643 individuals were used for meta-analyses of bradycardia.^{24, 26, 28, 36} Given that only 2 studies reported symptomatic bradycardia, and that the number of events was low in the 2 studies, we limited the meta-analyses to all cases of bradycardia. The frequency of events was significantly reduced for non-PM compared to PM (OR: 0.27 [95%CI: 0.08 to 0.89]; P = .032; Figure 7) The effect was more important in the 2 smaller studies, as opposed to the 2 larger studies, resulting in significant heterogeneity. As opposed to our previous comparisons, heterogeneity was high in this analysis (I² = 68%, P = .03). Two-by-two meta-analyses showed that only EM were significantly less prone to developing bradycardia during treatment compared to IM (OR: 0.62 [95%CI: 0.40 to 0.96]; P = .032), but no other pairwise comparisons indicated a significant reduction of bradycardia for any phenotype. Again, UM were excluded from our meta-analyses for lack of values.

3.2.5 Mortality and hospitalizations

Hospitalizations, mortality and other composite criteria reflecting outcome events were only evaluated once in the retrospective substudy of the MERIT-HF trial.²⁸ Thus, no meta-analysis was done for this parameter. As part of this study, no effect on all-cause mortality and hospitalizations was observed, and early discontinuation was similar between EM, IM and PM.

3.2.6 Sensitivity analyses

The comparisons of the results for the meta-analyses performed with both the Hartung–Knapp and the DerSimonian–Laird methods are presented in Table S7. For the PM vs non-PM meta-analyses, the effects were consistent across all analyses, although the ΔSBP meta-analysis was not significant using the DerSimonian–Laird method (P = .093). We initially sought to assess publication bias by means of funnel plot and asymmetry,⁴⁴ yet no parameter reached the minimum requirement of 10 included studies for such analyses to be conducted.

4 DISCUSSION

We performed a qualitative systematic review and meta-analyses to assess the effects of CYP2D6 polymorphism on metoprolol pharmacodynamics. Our meta-analyses suggest that CYP2D6 PM have a greater reduction in HR compared to non-PM by >3 beats/min when treated with metoprolol. This enhanced β-blocking effect may also lead to PM being approximately 4 times more at risk of bradycardia, although this result should be interpreted with caution because of significant heterogeneity observed for this endpoint, suggesting an overestimated effect size in smaller studies. Also, given that the doses used in the reviewed studies were similar between phenotypes, these results are therefore consistent with a previous meta-analysis,¹³ indicating that PM present higher metoprolol plasma concentrations than non-PM when treated with a fixed metoprolol dose, resulting in greater reductions in HR and blood pressure. Both systolic and diastolic blood pressures were also significantly more reduced for PM, with average decreases both greater than non-PM by ~3 mmHg. Larger populations than those included in our meta-analyses will be required to investigate the impact of CYP2D6 phenotypes on other AE and reductions in cardiovascular endpoints across all 4 inferred phenotypes.

The clinical significance of our findings requires further investigation. Indeed, although our results suggest that PM could be at a higher risk of bradycardia, one could also argue that the greater HR reduction in PM (3 beats/min) could result in a greater benefit from metoprolol. Indeed, a previous meta-analysis of 17 randomized, placebo-controlled trials of β-blocker combining 17 831 patients with heart failure, indicated a decrease in relative risk of death by 18% per 5 beats/min (95%CI: 6 to 29%) for patients undergoing β-blocker treatment,⁴⁵ which is slightly above the differences in reduction between non-PM and PM. The clinical benefits in resting HR reduction have also been demonstrated during post-myocardial infarction intervention, notably through β-blocker use.⁴⁶ Hence, PM patients would benefit more than other phenotypes from receiving metoprolol treatment at similar doses, as long as symptomatic bradycardia is avoided. One could also speculate that using higher metoprolol doses in non-PM, or using other non-CYP2D6 metabolized β-blockers, could lead to comparable efficacy. For example, in the pharmacogenetic sub-study of the MERIT-HF trial, the authors concluded that to achieve a 1-beat reduction of HR, 7.7, 5.9 and 5.3 mg of metoprolol CR/XL were needed for EM, IM and PM, respectively, which was significantly different between phenotypes (P = .04). Thus, our results are consistent with the DPWG recommending initiation of metoprolol at lower doses for PM, given the apparent greater response and potentially higher risk of bradycardia. Nevertheless, since a greater HR response could not only be associated with a greater risk of bradycardia, but also with a greater clinical benefit,⁴⁵ the optimal metoprolol dosing algorithm based on CYP2D6 genotyping and its clinical utility remains to be established.

We found a statistically significant increase of bradycardia amongst PM compared to non-PM, although significant heterogeneity was observed in the meta-analysis. However, these results are consistent with those from a report by Bijl et al³⁷ in which PM were more likely to suffer from bradycardia than EM (OR: 3.86 [95%CI: 1.68 to 8.86]; P = .0014). Unfortunately, we could not include their data in the current meta-analysis because specific sample sizes and values in each group were not reported. Still, the 4-fold difference in the risk of bradycardia is consistent with the risk observed in the current meta-analysis. Given that only 2 studies specifically reported symptomatic bradycardia, we could not determine whether the differences in frequencies of bradycardia would warrant changes in medication or dosing for PM. Thus, particularly given the low incidence of symptomatic bradycardia in the 2 studies reporting it, future pharmacogenetic investigations on this gene–drug pair should precisely document the occurrence of symptomatic bradycardia.

Beta-blockers, including metoprolol, are known to have modest blood pressure lowering effect.⁸ Nonetheless, the results of this meta-analysis indicate that CYP2D6 genotype-inferred phenotypes could be useful to guide the selection of β-blocker in the treatment of hypertension. Indeed, even if individual studies did not consistently show significant differences between inferred phenotypes, our meta-analysis showed that both reductions in SBP and DBP are greater in PM compared to other phenotypes by approximately 3 mmHg. Although such a difference might appear modest, evidence suggests that from a populational standpoint, a DBP reduction of only 2 mmHg would lead to a decrease of stroke by 15%.⁴⁷ According to recent estimates, an equally small SBP reduction would also significantly reduce the annual incidence of coronary heart disease, stroke and heart failure events.⁴⁸

5 STUDY LIMITATIONS

Our meta-analysis contains some limitations. For this systematic review, we used the DPWG guidelines to assign a phenotype to patients in studies for which only their CYP2D6 genotypes had been mentioned. However, others could have favoured the classification proposed by the Clinical Pharmacogenomics Implementation Consortium. Since we focused our attention on comparisons between PM and non-PM, this would have minimal impact on our conclusions because discrepancies between the 2 groups are mainly related to EM and IM.⁴⁹

In addition, many of the included studies only investigated a limited number of CYP2D6 polymorphisms. As different polymorphisms may be present within specific populations,⁷ limited genotyping assays could prove problematic with this highly polymorphic gene. These technical issues have been addressed prior to this review.^{4, 50} However, such shortcomings would have added heterogeneity to our analyses. Therefore, this limitation should not restrain the scope of our findings.

From a statistical perspective, an important limitation in our review is the small number of studies, and the small and unbalanced number of patients within the compared groups. Thus, whether CYP2D6 inferred phenotypes has an impact of the overall risk of AE or cardiovascular events remains to be established. A greater number of patients would also be required to perform adequately powered comparisons between all 4 phenotypes.

A strength of our study is that our sensitivity analyses for continuous variables show that results generally remain consistent between the 2 methods for calculating confidence intervals. The Hartung–Knapp adjustment has been shown to be a useful tool in making better inferences and refining measurements when performing meta-analyses.²⁰ Despite those improvements, confidence intervals for random-effects models can be overly conservative in some scenarios, while becoming too restrained in others.^{20, 21} This could explain the fact that reductions in SBP were not statistically significant using the DerSimonian–Laird method, as confidence intervals calculated were wider. Although we were able to perform meta-analyses on almost all parameters reviewed, comparisons where ≤3 studies were included may contain considerable uncertainty, regardless of low heterogeneity. Indeed, some small studies with neutral results have not reported all relevant data, thus potentially overestimating the effects observed.³⁸ While our systematic review shows certain trends that seem likely to be confirmed, careful assessment of our results is required.

To conclude, this systematic review provides a comprehensive assessment of the literature regarding the genetic influence of CYP2D6 polymorphisms on metoprolol response. Our meta-analyses suggest that patients with a PM phenotype have greater HR, SBP and DBP reductions compared to non-PM, and may be more subject to bradycardia. This latter finding should be interpreted with caution, given the heterogeneity observed in the analysis of this specific end point. Further prospective data are required to determine whether CYP2D6 is associated with clinical events in patients treated with metoprolol. Ultimately, the clinical utility of CYP2D6-guided β-blocker use should be investigated in a prospective clinical trial.

ACKNOWLEDGEMENTS

Maxime Meloche has received scholarships from the Université de Montréal Faculty of Pharmacy and Faculty of Higher Education.

Simon de Denus holds the Université de Montréal Beaulieu-Saucier Chair in Pharmacogenomics. This study was also supported by the Montreal Heart Institute Foundation.

COMPETING INTERESTS

M.-P.D. is a coauthor on a patent pertaining to pharmacogenomics-guided CETP inhibition and has minor equity interest in DalCor. M.-P.D. has received honoraria from Dalcor and research support (access to samples and data) from AstraZeneca, Pfizer, Servier, Sanofi and GlaxoSmithKline. S.D. has received research support through grants from Pfizer, AstraZeneca, Roche Molecular Science, and DalCor.

CONTRIBUTORS

M.M. performed the systematic review, generated the statistical meta-analyses, and wrote parts of the manuscript. M.K. was the second reviewer and wrote parts of the manuscript. I.K. developed the computer code used in the R software to create the meta-analyses. A.B. contributed to elaborate the statistical models and revised the manuscript. M.-P.D. Dubé contributed to the design of the analyses and revised the manuscript. S.D. designed the project, wrote parts of and revised the manuscript.

Supporting Information

Filename	Description
bcp14247-sup-0001-Supporting Information, Table S1.docxWord 2007 document , 13.7 KB	TABLE S1 Online search strategy
bcp14247-sup-0002-Supporting Information, Table S2.docxWord 2007 document , 18.9 KB	TABLE S2 Raw data of reduction in heart rate (ΔHR), systolic blood pressure (ΔSBP), and diastolic blood pressure (ΔDBP) for the meta-analyses
bcp14247-sup-0003-Supporting Information, Table S3.docxWord 2007 document , 18.7 KB	TABLE S3 Extracted raw data of the frequency of all-type adverse events (AE) and bradycardia between phenotypes contributing to meta-analyses
bcp14247-sup-0004-Supporting Information, Table S4.docxWord 2007 document , 19.5 KB	TABLE S4 Estimated means and standard deviations for average daily doses using Wan et al.’s formulas
bcp14247-sup-0005-Supporting Information, Table S5.docxWord 2007 document , 16.1 KB	TABLE S5 Summary of all extracted and estimated data contributing to the meta-analyses of the differences in titrated doses between phenotypes
bcp14247-sup-0006-Supporting Information, Table S6.docxWord 2007 document , 15.5 KB	TABLE S6 Rejected studies from the systematic review for clinical parameters and reasons of exclusion after full-text assessment
bcp14247-sup-0007-Supporting Information, Table S7.docxWord 2007 document , 17.7 KB	TABLE S7 Sensitivity analyses of random-effects model adjustments

Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

REFERENCES

1Zanger UM, Schwab M. Cytochrome P450 enzymes in drug metabolism: regulation of gene expression, enzyme activities, and impact of genetic variation. Pharmacol Ther. 2013; 138(1): 103-141.
10.1016/j.pharmthera.2012.12.007
CASPubMedWeb of Science®Google Scholar
2Whirl-Carrillo M, McDonagh EM, Hebert JM, et al. Pharmacogenomics knowledge for personalized medicine. Clinical Pharmacology & Therapeutics. 2012; 92(4): 414-417.
10.1038/clpt.2012.96
CASPubMedWeb of Science®Google Scholar
3Swen JJ, Nijenhuis M, de Boer A, et al. Pharmacogenetics: from bench to byte--an update of guidelines. Clin Pharmacol Ther. 2011; 89(5): 662-673.
10.1038/clpt.2011.34
CASPubMedWeb of Science®Google Scholar
4Crews KR, Gaedigk A, Dunnenberger HM, et al. Clinical Pharmacogenetics implementation consortium guidelines for cytochrome P450 2D6 genotype and codeine therapy: 2014 update. Clin Pharmacol Ther. 2014; 95(4): 376-382.
10.1038/clpt.2013.254
CASPubMedWeb of Science®Google Scholar
5Hicks JK, Bishop JR, Sangkuhl K, et al. Clinical Pharmacogenetics implementation consortium (CPIC) guideline for CYP2D6 and CYP2C19 genotypes and dosing of selective serotonin reuptake inhibitors. Clin Pharmacol Ther. 2015; 98(2): 127-134.
10.1002/cpt.147
CASPubMedWeb of Science®Google Scholar
6Relling MV, Gardner EE, Sandborn WJ, et al. Clinical Pharmacogenetics implementation consortium guidelines for Thiopurine Methyltransferase genotype and Thiopurine dosing: 2013 update. Clin Pharmacol Ther. 2013; 93(4): 324-325.
10.1038/clpt.2013.4
CASPubMedWeb of Science®Google Scholar
7Gaedigk A, Sangkuhl K, Whirl-Carrillo M, Klein T, Leeder JS. Prediction of CYP2D6 phenotype from genotype across world populations. Genet Med. 2017; 19(1): 69-76.
10.1038/gim.2016.80
PubMedWeb of Science®Google Scholar
8Effect of metoprolol CR/XL in chronic heart failure: metoprolol CR/XL randomised intervention trial in congestive heart failure (MERIT-HF). Lancet (London, England). 1999; 353(9169): 2001-2007.
10.1016/S0140-6736(99)04440-2
CASPubMedWeb of Science®Google Scholar
9Wikstrand J, Warnold I, Tuomilehto J, et al. Metoprolol versus thiazide diuretics in hypertension. Morbidity results from the MAPHY study. Hypertension. 1991; 17(4): 579-588.
10.1161/01.HYP.17.4.579
CASPubMedWeb of Science®Google Scholar
10Kuhlkamp V, Schirdewan A, Stangl K, Homberg M, Ploch M, Beck OA. Use of metoprolol CR/XL to maintain sinus rhythm after conversion from persistent atrial fibrillation: a randomized, double-blind, placebo-controlled study. J am Coll Cardiol. 2000; 36(1): 139-146.
10.1016/S0735-1097(00)00693-8
CASPubMedWeb of Science®Google Scholar
11Lennard MS, Silas JH, Freestone S, Ramsay LE, Tucker GT, Woods HF. Oxidation phenotype--a major determinant of metoprolol metabolism and response. N Engl J Med. 1982; 307(25): 1558-1560.
10.1056/NEJM198212163072505
CASPubMedWeb of Science®Google Scholar
12Borg KO, Carlsson E, Hoffmann K-J, Jönsson T-E, Thorin H, Wallin B. Metabolism of metoprolol-(3H) in man, the dog and the rat. Acta Pharmacol Toxicol. 1975; 36(s5): 125-135.
10.1111/j.1600-0773.1975.tb03329.x
CASPubMedGoogle Scholar
13Blake CM, Kharasch ED, Schwab M, Nagele P. A meta-analysis of CYP2D6 metabolizer phenotype and metoprolol pharmacokinetics. Clin Pharmacol Ther. 2013; 94(3): 394-399.
10.1038/clpt.2013.96
CASPubMedWeb of Science®Google Scholar
14Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 2009; 339(jul21 1): b2700.
10.1136/bmj.b2700
PubMedGoogle Scholar
15Gaedigk A, Sangkuhl K, Whirl-Carrillo M, et al. The evolution of PharmVar. Clin Pharmacol Ther. 2019; 105(1): 29-32.
10.1002/cpt.1275
PubMedWeb of Science®Google Scholar
16IntHout J, Ioannidis JP, Borm GFJBMRM. The Hartung-Knapp-Sidik-Jonkman method for random effects meta-analysis is straightforward and considerably outperforms the standard DerSimonian-Laird method. BMC Med Res Methodol. 2014; 14(1): 25-41.
10.1186/1471-2288-14-25
PubMedWeb of Science®Google Scholar
17Bakbergenuly I, Kulinskaya E. Beta-binomial model for meta-analysis of odds ratios. Stat Med. 2017; 36(11): 1715-1734.
10.1002/sim.7233
PubMedWeb of Science®Google Scholar
18Mathes T, Kuss O. A comparison of methods for meta-analysis of a small number of studies with binary outcomes. Res Synth Methods. 2018; 9(3): 366-381.
10.1002/jrsm.1296
PubMedWeb of Science®Google Scholar
19Bakbergenuly I, Kulinskaya EJBMRM. Meta-analysis of binary outcomes via generalized linear mixed models: a simulation study. Res Synth Methods. 2018; 18(1): 70-87.
Google Scholar
20Jackson D, Law M, Rucker G, Schwarzer G. The Hartung-Knapp modification for random-effects meta-analysis: a useful refinement but are there any residual concerns? Stat Med. 2017; 36(25): 3923-3934.
10.1002/sim.7411
PubMedWeb of Science®Google Scholar
21Wiksten A, Rücker G, Schwarzer G. Hartung–Knapp method is not always conservative compared with fixed-effect meta-analysis. Stat Med. 2016; 35(15): 2503-2515.
10.1002/sim.6879
PubMedWeb of Science®Google Scholar
22Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol. 2014; 14(1): 135-147.
10.1186/1471-2288-14-135
PubMedWeb of Science®Google Scholar
23Goryachkina K, Burbello A, Boldueva S, Babak S, Bergman U, Bertilsson L. CYP2D6 is a major determinant of metoprolol disposition and effects in hospitalized Russian patients treated for acute myocardial infarction. Eur J Clin Pharmacol. 2008; 64(12): 1163-1173.
10.1007/s00228-008-0525-3
CASPubMedWeb of Science®Google Scholar
24Rau T, Wuttke H, Michels LM, et al. Impact of the CYP2D6 genotype on the clinical effects of metoprolol: a prospective longitudinal study. Clin Pharmacol Ther. 2009; 85(3): 269-272.
10.1038/clpt.2008.218
CASPubMedWeb of Science®Google Scholar
25Rau T, Heide R, Bergmann K, et al. Effect of the CYP2D6 genotype on metoprolol metabolism persists during long-term treatment. Pharmacogenetics. 2002; 12(6): 465-472.
10.1097/00008571-200208000-00007
CASPubMedWeb of Science®Google Scholar
26Hamadeh IS, Langaee TY, Dwivedi R, et al. Impact of CYP2D6 polymorphisms on clinical efficacy and tolerability of metoprolol tartrate. Clin Pharmacol Ther. 2014; 96(2): 175-181.
10.1038/clpt.2014.62
CASPubMedWeb of Science®Google Scholar
27Luzum JA, Sweet KM, Binkley PF, et al. CYP2D6 genetic variation and Beta-blocker maintenance dose in patients with heart failure. Pharm Res. 2017; 34(8): 1615-1625.
10.1007/s11095-017-2104-8
CASPubMedWeb of Science®Google Scholar
28Batty JA, Hall AS, White HL, et al. An investigation of CYP2D6 genotype and response to metoprolol CR/XL during dose titration in patients with heart failure: a MERIT-HF substudy. Clin Pharmacol Ther. 2014; 95(3): 321-330.
10.1038/clpt.2013.193
CASPubMedWeb of Science®Google Scholar
29Sharp CF, Gardiner SJ, Jensen BP, et al. CYP2D6 genotype and its relationship with metoprolol dose, concentrations and effect in patients with systolic heart failure. Pharmacogenomics J. 2009; 9(3): 175-184.
10.1038/tpj.2009.9
CASPubMedWeb of Science®Google Scholar
30Baudhuin LM, Miller WL, Train L, et al. Relation of ADRB1, CYP2D6, and UGT1A1 polymorphisms with dose of, and response to, carvedilol or metoprolol therapy in patients with chronic heart failure. Am J Cardiol. 2010; 106(3): 402-408.
10.1016/j.amjcard.2010.03.041
CASPubMedWeb of Science®Google Scholar
31Fedorinov DS, Mirzaev KB, Mustafina VR, et al. Pharmacogenetic testing by polymorphic markers G1846A (CYP2D6*4) and C100T (CYP2D6*10) of the CYP2D6 gene in coronary heart disease patients taking beta beta -blockers in the republic of Sakha (YAKUTIA). Drug Metabolism and Personalized Therapy. 2018; 33(4): 195-200.
10.1515/dmpt-2018-0015
CASPubMedGoogle Scholar
32Fux R, Morike K, Prohmer AM, et al. Impact of CYP2D6 genotype on adverse effects during treatment with metoprolol: a prospective clinical study. Clin Pharmacol Ther. 2005; 78(4): 378-387.
10.1016/j.clpt.2005.07.004
CASPubMedWeb of Science®Google Scholar
33Nozawa T, Taguchi M, Tahara K, et al. Influence of CYP2D6 genotype on metoprolol plasma concentration and beta-adrenergic inhibition during long-term treatment: a comparison with bisoprolol. J Cardiovasc Pharmacol. 2005; 46(5): 713-720.
10.1097/01.fjc.0000184117.76188.68
CASPubMedWeb of Science®Google Scholar
34Goryachkina K, Burbello A, Boldueva S, Babak S, Bergman U, Bertilsson L. Inhibition of metoprolol metabolism and potentiation of its effects by paroxetine in routinely treated patients with acute myocardial infarction (AMI). Eur J Clin Pharmacol. 2008; 64(3): 275-282.
10.1007/s00228-007-0404-3
CASPubMedWeb of Science®Google Scholar
35Terra SG, Pauly DF, Lee CR, et al. Beta-adrenergic receptor polymorphisms and responses during titration of metoprolol controlled release/extended release in heart failure. Clin Pharmacol Ther. 2005; 77(3): 127-137.
10.1016/j.clpt.2004.10.006
CASPubMedWeb of Science®Google Scholar
36Wuttke H, Rau T, Heide R, et al. Increased frequency of cytochrome P450 2D6 poor metabolizers among patients with metoprolol-associated adverse effects. Clin Pharmacol Ther. 2002; 72(4): 429-437.
10.1067/mcp.2002.127111
CASPubMedWeb of Science®Google Scholar
37Bijl MJ, Visser LE, van Schaik RH, et al. Genetic variation in the CYP2D6 gene is associated with a lower heart rate and blood pressure in beta-blocker users. Clin Pharmacol Ther. 2009; 85(1): 45-50.
10.1038/clpt.2008.172
CASPubMedWeb of Science®Google Scholar
38Zineh I, Beitelshees AL, Gaedigk A, et al. Pharmacokinetics and CYP2D6 genotypes do not predict metoprolol adverse events or efficacy in hypertension. Clin Pharmacol Ther. 2004; 76(6): 536-544.
10.1016/j.clpt.2004.08.020
CASPubMedWeb of Science®Google Scholar
39Koytchev R, Alken RG, Vlahov V, et al. Influence of the cytochrome P4502D6*4 allele on the pharmacokinetics of controlled-release metoprolol. Eur J Clin Pharmacol. 1998; 54(6): 469-474.
10.1007/s002280050495
CASPubMedWeb of Science®Google Scholar
40Yuan H, Huang Z, Yang G, Lv H, Sang H, Yao Y. Effects of polymorphism of the beta(1) adrenoreceptor and CYP2D6 on the therapeutic effects of metoprolol. J Int Med Res. 2008; 36(6): 1354-1362.
10.1177/147323000803600624
CASPubMedWeb of Science®Google Scholar
41Kirchheiner J, Heesch C, Bauer S, et al. Impact of the ultrarapid metabolizer genotype of cytochrome P450 2D6 on metoprolol pharmacokinetics and pharmacodynamics. Clin Pharmacol Ther. 2004; 76(4): 302-312.
10.1016/j.clpt.2004.07.002
CASPubMedWeb of Science®Google Scholar
42Hamelin BA, Bouayad A, Methot J, et al. Significant interaction between the nonprescription antihistamine diphenhydramine and the CYP2D6 substrate metoprolol in healthy men with high or low CYP2D6 activity. Clin Pharmacol Ther. 2000; 67(5): 466-477.
10.1067/mcp.2000.106464
CASPubMedWeb of Science®Google Scholar
43Sharma A, Pibarot P, Pilote S, et al. Modulation of metoprolol pharmacokinetics and hemodynamics by diphenhydramine coadministration during exercise testing in healthy premenopausal women. J Pharmacol Exp Ther. 2005; 313(3): 1172-1181.
10.1124/jpet.104.081109
CASPubMedWeb of Science®Google Scholar
44Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997; 315(7109): 629-634.
10.1136/bmj.315.7109.629
CASPubMedWeb of Science®Google Scholar
45McAlister FA, Wiebe N, Ezekowitz JA, Leung AA, Armstrong PW. Meta-analysis: Β-blocker dose, heart rate reduction, and death in patients with heart failure. Ann Intern Med. 2009; 150(11): 784-794.
10.7326/0003-4819-150-11-200906020-00006
PubMedWeb of Science®Google Scholar
46Cucherat M. Quantitative relationship between resting heart rate reduction and magnitude of clinical benefits in post-myocardial infarction: a meta-regression of randomized clinical trials. Eur Heart J. 2007; 28(24): 3012-3019.
10.1093/eurheartj/ehm489
PubMedWeb of Science®Google Scholar
47Cook NR, Cohen J, Hebert PR, Taylor JO, Hennekens CH. Implications of small reductions in diastolic blood pressure for primary prevention. JAMA Intern Med. 1995; 155(7): 701-709.
10.1001/archinte.1995.00430070053006
CASGoogle Scholar
48Hardy ST, Loehr LR, Butler KR, et al. Reducing the blood pressure-related burden of cardiovascular disease: impact of achievable improvements in blood pressure prevention and control. J am Heart Assoc. 2015; 4(10): e002276-e002276.
10.1161/JAHA.115.002276
PubMedWeb of Science®Google Scholar
49 Bank PCD, Caudle KE, Swen JJ, et al. Comparison of the guidelines of the clinical Pharmacogenetics implementation consortium and the Dutch Pharmacogenetics working group. Clin Pharmacol Ther. 2018; 103(4): 599-618.
10.1002/cpt.762
PubMedWeb of Science®Google Scholar
50Hicks JK, Swen JJ, Gaedigk A. Challenges in CYP2D6 phenotype assignment from genotype data: a critical assessment and call for standardization. Curr Drug Metab. 2014; 15(2): 218-232.
10.2174/1389200215666140202215316
CASPubMedWeb of Science®Google Scholar

Citing Literature

Volume86, Issue6

June 2020

Pages 1015-1033

CYP2D6 polymorphism and its impact on the clinical response to metoprolol: A systematic review and meta-analysis

Abstract

Aims