Pharmacogenomics and Cardiovascular Disease

Abstract

Variability in drug responsiveness is a sine qua non of modern therapeutics, and the contribution of genomic variation is increasingly recognized. Investigating the genomic basis for variable responses to cardiovascular therapies has been a model for pharmacogenomics in general and has established critical pathways and specific loci modulating therapeutic responses to commonly used drugs such as clopidogrel, warfarin, and statins. In addition, genomic approaches have defined mechanisms and genetic variants underlying important toxicities with these and other drugs. These findings have not only resulted in changes to the product labels but also have led to development of initial clinical guidelines that consider how to facilitate incorporating genetic information to the bedside. This review summarizes the state of knowledge in cardiovascular pharmacogenomics and considers how variants described to date might be deployed in clinical decision making.

Introduction

Pharmacogenomic research aims to identify how genetic variability affects drug responsiveness and toxicity. Substantial progress has been made over the past decade in improving our understanding of genetic determinants influencing response to cardiovascular drugs such as warfarin, clopidogrel, and simvastatin. These advances have fueled a hope for “personalized medicine” which includes tailoring of diagnostic and treatment strategies to the needs and characteristics of individual patients – including genomic variation as well as other features such as personal preferences or access to care – aiming to improve drug responsiveness and lessen risk of toxicity. This article describes recent findings in the realm of cardiovascular pharmacogenomics and discusses the potential for clinical implication.

Clopidogrel

Clopidogrel is a thienopyridine derivative commonly used to prevent thromboembolic events including myocardial infarction (MI) and stent thrombosis among patients with cardiovascular disease. The antiplatelet activity of clopidogrel is characterized by considerable inter-individual responsiveness with certain genotypes causing attenuated platelet inhibition and ultimately more adverse cardiovascular events [1]. Clopidogrel is a prodrug that requires oxidation by cytochrome P-450 (CYP) iso-enzymes to generate the pharmacologically active thiol-containing metabolite that inhibits the platelet adenosine diphosphate receptor, P2RY₁₂ [2].

Polymorphisms in CYP2C19, the major CYP isoform mediating clopidogrel bioactivation, have been associated with both decreased and increased enzymatic function. CYP2C19*2 (rs4244285) is the most common loss-of-function allele and is associated with impaired clopidogrel activity, i.e., decreased platelet inhibition [3-6]. Homozygous carriers for the loss-of-function allele (*2/*2; “poor metabolizers”) have decreased platelet inhibition compared to individuals who carry the reference alleles (*1/*1; “extensive metabolizers”), and heterozygotes (*1/*2; “intermediate metabolizers”) [4, 7, 8]. The *1/*2 genotype is found in ~25 % of Caucasians, and ~2 % are *2/*2 [9••] Other loss-of-function alleles that mimic the effects of CYP2C19*2 have also been identified: *3 (rs498693), *4 (rs28399504), *5 (rs72552267). These are rare (minor allele frequency [MAF] <1 %) in Caucasian populations, but *3 are common among Asians (MAF: 2-9 %). In contrast to these loss-of-function alleles, *17 (rs3758581) is a gain-of-function allele and has been associated with increased enzymatic activity and improved platelet inhibition [10]. Carriers of the *17 variant have been termed ultra-rapid metabolizers.

In acute coronary syndrome (ACS) and percutaneous intervention (PCI), the CYP2C19*2 loss-of-function variant has been associated with increased risk of adverse cardiovascular events including stent thrombosis during clopidogrel treatment (Table 1) [7, 10-12, 13••, 14•]. Table 1 lists pharmacogenomic targets where the level of evidence of clinical implementation is strong. Data from a meta-analysis on 9685 clopidogrel treated patients identified a gene dose–response relationship between the number of loss-of-function variants and the composite end-point of cardiovascular death, myocardial infarction, and stroke. Compared to non-carriers, carriers of one loss-of-function allele had increased risk of a cardiovascular event (hazard ratio (HR) = 1.55, 95 % confidence interval (CI): 1.11-2.17) and carriers of two loss-of-function variants had an even greater risk (HR = 1.76, CI: 1.24-2.50) (Table 1) [13••]. Notably, the risk of stent thrombosis was particularly high among carriers of one loss-of-function variant (HR = 2.81, CI: 1.81-4.37), and even higher in homozygous (*2/*2) patients (HR = 3.97, CI: 1.75-9.02) (Table 2) [13••]. This pharmacogenomic interaction between clopidogrel and carriers of a CYP2C19*2 loss-of-function variant was not been replicated in other analyses [15, 16]; one possible explanation is that the interaction is especially important in ACS/PCI compared to other indications for clopidogrel. The Food and Drug Administration (FDA) has issued a black-box warning that loss-of-function carriers may exhibit lower capacity to metabolize clopidogrel with subsequent elevated risk of thromboembolic complications [17]. Conversely, carriers of the gain-of-function variant allele *17 have been reported to display increased risk of bleeding [18] and increased protection from ischemic events [16]; these data have not yet been widely replicated (Table 2). Table 2 lists the pharmacogenomic targets where current level of evidence for clinical application is weak.

Table 1.

Cardiovascular pharmacogenomic targets with strong level of evidence for clinical implementation

Gene	Variant	Chr	Position	Reported allele frequency of the risk-allele (%)			Drug	Drug response associated with risk allele	Possible clinical implications
				Caucasians	Africans	Asians
CYP2C19	*2 (rs4244285)	10	96531606	16	14	27	Clopidogrel	Reduced active metabolite  concentration. Impaired  platelet inhibition.	Carriers of the risk allele  who are treated for ACS  or PCI have a graded  risk of cardiovascular  complications with 0,1,  or 2 alleles including MI,  stroke, and stent  thrombosis
VKORC1	−1639G>A (rs9923231)	16	31107689	60	98	2	Warfarin		Carriers of the A allele  require lower warfarin  dosing. Increased time  to stable dose
CYP2C9	*2 (rs1799853)	10	96702047	10	<1	<1	Warfarin	Reduced clearance of the  potent S-warfarin	Stepwise reduction in  warfarin dosing  required with 0, 1, or 2  alleles. Increased risk of    bleeding
CYP2C9	*3 (rs1057910)	10	96741053	6	<1	4	Warfarin	Reduced clearance of the  potent S-warfarin	Stepwise reduction in  warfarin dosing  required with 0, 1, or 2  alleles. Increased risk of  bleeding
SLC01B1	rs4149056	12	21331549	15	<1	13	Simvastatin	Impaired transporter  function leading to  increased simvastatin  levels	Simvastatin induced  myopathy, particularly  among patients on high  dose simvastatin

ACS acute coronary syndrome; PCI percutaneous coronary intervention; MI myocardial infarction

Table 2.

Cardiovascular pharmacogenomic targets with weaker level of evidence for clinical implementation

Gene	Variant	Chr	Position	Reported allele frequency of the risk-allele (%)			Drug	Drug response associated with risk allele	Possible clinical implications
				Caucasians	Africans	Asians
ABCB1	C3435T (rs1045642)	7	89676581	57	11	46	Clopidogrel	Reduced bioavailability of  substrate drug	Reduced platelet inhibition
CYP2C19	*17 (rs12248560)	10	96511647	5	10	6	Clopidogrel	Increased active metabolite  concentration. Increased  platelet inhibition	Increased risk of bleeding
CYP2C19	*3 (rs4986893)	10	96540410	<1		2-9	Clopidogrel	Reduced active metabolite  concentration. Impaired  platelet inhibition.	Carriers of the risk allele have  increased risk of stent thrombosis  following PCI
ADBR1	Ser49Gly (rs1801252)	10	115804036	20	<1	<1	Metoprolol	Impaired down-regulation  of ADBR1 leading to  higher signal transduction	Carriers of risk allele have greater  reductions in BP and HR
ADBR1	Arg389Gly (rsl801253)	10	115805056	31	41	20	Metoprolol	Impaired down-regulation  of ADBR1 leading to  higher signal transduction	Carriers of risk allele have greater  reductions in BP and HR as  well as improved LVEF
CYP2D6	*4 (rs3892097) and  many others	22	42524947	18		23	Metoprolol	Impaired metabolism	Increased plasma  metoprolol concentrations  with reduction in BP and  HR. Increased risk of ADR  especially among vulnerable HF  patients
KCNE1	D85N (Rsl805128)	21	35821680	5	4	2	Multiple	Modulates important  electronic currents  of the heart	Attenuated risk of TdP
HMGCR	H7 haplotype (defined  by three intronic SNPs:  rs17244841, rs3846662,  and rsl7238540)	5	74642855,  74651084,  74655498	3	6		Simvastatin,  pravastatin	Alternatively spliced  MHGCR transcript that is  less sensitive to therapy	Impaired LDLc lowering
APOE	ε2, ε3, ε4 (defined by  rs7412 and rs429358)	19	45412079,  45411941				Statins	Major binding protein for  VLDL/IDL cholesterol	Impaired LDLc reduction

PCI percutaneous coronary intervention; BP blood pressure; HR heart rate; LVEF left ventricular ejection fraction; ADR adverse drug reaction; TdP Torsade de Pointes

ABCB1 encodes P-glycoprotein, an ATP-dependent drug efflux transporter. Carriers of a common polymorphism C3435T in ABCB1 (homozygous carriers in particular) have been shown to have reduced bioavailability of substrate drugs such as clopidogrel [19] and have higher rates of adverse cardiovascular outcomes [15, 20, 21••], although this finding has not been consistent (Table 2) [22]. Patients treated with prasugrel or ticagrelor are not seemingly affected by common variants at ABCB1 or CYP2C19 [15, 23].

Common variants in P2YR₁₂ (rs6798347, rs6787801, rs9859552, rs6801273, rs9848789, and rs2046934), a G-protein-coupled ADP receptor involved in regulation of platelet aggregation, have been associated with improved platelet inhibition in a candidate-gene study [24]. However, this finding has not been consistently replicated [20]. One report found that the common Q192R (rs662) variant in Paraoxonase-1 (PON1), thought to be involved in the bioactivation of clopidogrel, had greater risk of stent thrombosis [25], although this finding has not been replicated in multiple other datasets [7, 14•, 26]. A common variant in another CYP (CYP2C9*3; rs1057910) showed an association with reduced platelet inhibition in initial reports [4], but has not been validated [7, 10].

Clinical Implications

An ACC/AHA task force did not recommend routine CYP2C19 testing [27]. The Clinical Pharmacogenetics Implementation Consortium (CPIC) poses a different question: if the genotype were available, what action, if any should be considered [9••]? In this context, CPIC recommends that patients with ACS or PCI who carry one or two copies of the loss-of-function allele should receive alternative antiplatelet agents (i.e., prasugrel or ticagrelor) in order to alleviate the risk of cardiovascular adverse events as these do not undergo extensive bioactivation by the CYP2C19 enzyme. Investigators at Vanderbilt University have already described a program implementing preemptive CYP2C19 genotyping in patients likely to receive clopidogrel [28•].

Warfarin

Vitamin K antagonists (warfarin and others) are effective and widely used for treatment and prevention of venous and arterial thromboembolic events. Warfarin targets blood coagulation by inhibiting the vitamin K epoxide reductase multiprotein complex, subunit 1 (VKORC1)[29] and inhibits the activation of vitamin K dependent clotting factors II, VII, IX, and X [30]. However, wide inter-individual dosing variability and a narrow therapeutic index necessitates frequent monitoring to balance the risks of over anticoagulation and bleeding with those of under anticoagulation and clotting and this monitoring adds to the complexity of warfarin use [31, 32]. Warfarin-related bleeding complications are one of the most common reasons for emergency room visits [33]. Variables contributing to steady state warfarin dose include clinical factors such as age, weight, diet, and interacting drugs (notably amiodarone) and genetic factors. Polymorphisms at three loci have been associated with warfarin dose: CYP2C9, VKORC1, and CYP4F2.

CYP2C9 is the predominant enzyme responsible for metabolism (to inactive metabolites) of the potent warfarin S-enantiomer [29]. Common polymorphisms CYP2C9*2 (rs1799853) and CYP2C9*3 (rs1057910) have been identified as the loci most commonly associated with reduced enzyme activity (~30-40 % and ~80-90 %, respectively) [34]. As compared with non-carriers, heterozygous carriers of the CYP2C9*2 or CYP2C9*3 reduced-function alleles require ~19 % and ~33 % reduction in dosing, respectively, and homozygous carriers require even greater reductions (~36 % and ~78 %, respectively) [35]. Carriers of CYP2C9*2 or CYP2C9*3 reduced-function alleles have also been reported to be at increased risk of hemorrhagic complications during warfarin treatment (Table 1) [36, 37].

VKORC1 acts by converting the oxidized, inactive form of vitamin K to the active form and represents the pharmacological target of warfarin. Multiple common promoter polymorphisms influencing VKORC1 gene expression and dosing requirements include −1173T>C (rs9934438), −3730G>A (rs7294), and −1639G>A (rs9923231) [32, 38]. The −1639G>A variant reduces protein expression [39] which results in lower warfarin maintenance dose requirement compared to non-carriers (Table 1) [32]. Multiple rare non-synonymous variants in VKORC1 leading to warfarin resistance and higher dose requirements have also been identified [40].

Genome-wide association studies have identified variants in CYP4F2, a vitamin K(1) oxidase associated with reduced capacity to metabolize vitamin K1 and higher warfarin dose requirements[41], although findings have been inconsistent [42].

Clinical Application

Collectively, ~50-60 % of warfarin variability can be accounted for if patient related factors are evaluated in concert with genetic factors [43, 44]. In order to facilitate the clinical implementation of genetic information guidelines incorporating genetic information have been published [45••]. The pharmacogenetic interaction relating to warfarin has also been recognized by the FDA [46]. Ongoing clinical trials aimed to evaluate the benefit of incorporating genetic information into the warfarin dosing algorithms will provide more insight [47, 48•]. The extent to which genotype-guided therapy will be implemented in practice will also be influenced by uptake of newer oral anticoagulants such as dabigatran or rivaroxaban.

Statins

3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR) inhibitors have in large clinical trials been shown to reduce the risk of cardiovascular events and are considered to be first-line therapy in conjunction with life-style changes for prevention of cardiovascular disease. Competitive HMGCR inhibitors, or statins, lower low-density lipoprotein cholesterol (LDLc) by attenuating endogenous cholesterol production, an early rate limiting step in cholesterol synthesis; they cause up-regulation of the liver LDLc receptor (LDLR) expression and increased LDLc clearance. However, the wide inter-individual variability in the extent of LDLc lowering by statins [49], attributable to environmental and genomic factors, means that some patients are at risk of cardiovascular events despite multiple dose adjustments [50]. In addition, concomitant drug therapy and genetic heterogeneity have been associated with statin-induced myopathy.

Responsiveness and Cardiovascular Events

The H7 haplotype of HMGCR (defined by three intronic SNPs: rs17244841, rs3846662, and rs17238540) has been associated with impaired reduction in LDLc lowering (5-20 %) following therapy with pravastatin [51] or simvastatin [52]. The attenuated LDLc response associated with the minor H7 haplotype has been attributed to an alternatively spliced HMGCR transcript that is less sensitive to simvastatin inhibition [53]. However, this finding has not been confirmed for all types of statins (Table 2) [54]. The H2 haplotype (defined by the intronic SNP rs3846662) and the LDLRL5 haplotype (defined by six SNPs within the LDLR3 un-translated region) have also been associated with attenuated LDLc response following statin treatment, particularly in blacks [52].

Genetic variation in apolipoprotein E (APOE) (two variants: rs429358 and rs7412 define the haplotypes ε2, ε3, and ε4) has previously been associated with variability in cholesterol levels across multiple populations [55] including the first statin response genome-wide association study (GWAS) [54]. Carriers of the ε2 haplotype have the greatest attenuation in lipid response followed by ε3 and ε4, respectively [56]. However, the validity of these findings been questioned by the lack of replication in a recent large meta-analysis (Table 2) [57].

The kinesin-like-protein 6 (KIF6) is involved in intracellular transport of key molecules including mRNA. Carriers of a missense SNP (Trp719Arg, rs20455) in KIF6 have been associated with increased risk of coronary events [58]. Interestingly, studies have found carriers of the risk allele to also have a greater benefit from statin therapy compared to non-carriers independent from LDLc, triglycerides and C-reactive protein [59, 60]. However, the association between KIF6 and cardiovascular disease was not replicated in a recent meta-analysis [61] which is in accordance with previous study findings that are unable to couple KIF6 to a differential statin response [62].

The drug pump P-glycoprotein serves to reduce bioavailability by promoting drug efflux from hepatocytes and enterocytes. The tri-allelic G2677T/A variant in ABCB1 (encoding the transporter) has been associated with impaired lowering of LDLc [63], whereas others have reported improved LDLc lowering [64].

The function and location of the ATP-binding cassette, subfamily G, member 2 (ABCG2), is similar to that of ABCB1. The common variant (rs2231142) reduces its transport capacity and rosuvastatin treated carriers of the A allele compared to non-carriers experience 5-7 % greater reduction in LDLc [65] although inconsistently [54].

Statin Induced Adverse Events

Genetic variation in the hepatic drug uptake transporter (polypeptide organic anion transporter P1B (OATP1B1)) encoded by SLCO1B1 is involved in the uptake of statins and has been coupled with altered risk of adverse drug reactions (ADRs). The predominant ADRs associated with statin therapy are myopathy and rhabdomyolysis (11.0 and 3.4 per 100,000 person-years, respectively) with the latter having a mortality rate of 10 % [66]. The pharmacokinetic properties of statin are altered via two common variants in SLCO1B1 (rs4149056 and rs2306283) that interferes with the transporter function [67] and hepatocyte localization of the transporter leading to higher plasma statin levels [68]. Findings from a GWAS saw a strong association between the common variant rs4363657 (rs4149056 and rs4363657 are in near-perfect disequilibrium, R²=0.97) and statin-induced myopathy in patients on high-dose simvastatin [69]. Notably, heterozygous carriers of the risk allele had odds ratio of 4.7 for developing myopathy per allele compared to non-carriers; the odds ratio was 17.4 in homozygotes (Table 1). Fifteen percent of the population were carriers of the risk allele (rs4149056) and 60 % of patients who developed myopathy were carriers of the risk allele [69]. These findings have subsequently been replicated [70].

Clinical Application

Using genetic information to determine the dose needed to acquire an adequate drop in LDLc is currently not implemented for several reasons. First, reasonably forecasting an expected LDLc drop is usually not a problem using only non-genetic patient information. Further, a smaller than expected LDLc reduction is not an acute problem and is often amenable to dose adjustments. Second, the genetic effect of genetic markers identified thus far is small. Third, variability in study designs and endpoints make interpretation of results difficult. Recent approaches modeling expected maximal statin effect on the basis of LDLc response to >1 dose may prove to be clinically useful [71].

Evidence of statin-induced ADRs has prompted CPIC Guidelines to generate guidelines with clinical recommendations for Simvastatin-Induced Myopathy (Table 1) [72••]. The extent to which these or other markers might also be useful for myopathy with other statins remains under study.

Beta-Blockers

Beta-adrenergic antagonists (β-blockers) are used for the management of cardiac arrhythmias, angina pectoris, myocardial infarction and hypertension. β -blockers competitively antagonize endogenous catecholamine at the beta-1-adrenergic receptor, encoded by ADBR1, and improve survival, remodeling and left ventricular ejection fraction after MI. However, inter-individual variability in response to β-blocker treatment has raised the question of genetic components influencing β-blocker response. Genes associated with inter-individual β-blocker response include CYP2D6 (pharmacokinetic; some agents), ADBR1, ADBR2, and GRK5 (pharmacodynamic).

Two common variants in ADBR1, resulting in Ser49Gly (rs1801252) and Arg389Gly (rs1801253), have previously been associated with impaired down-regulation [73], higher signal transduction [74], and altered biological function invitro [75]. Studies have shown that homozygous carriers of the Arg389 haplotype respond better to β-blocker treatment translating into improved left ventricular ejection fraction compared to carriers of the risk allele Gly389 [76-78], although such findings have not been consistent (Table 2) [79]. Similarly, reports of improved β-blocker response on blood pressure [80] and heart rate have also been inconsistent [81].

Two common variants in the beta-2-adrenergic receptor (ADBR2), resulting in Arg16Gly (rs1042713) and Gln27Glu (rs1042714), have been coupled with stimulation of adenylyl cyclase activity translating into increased agonistpromoted down-regulation of ADBR2 [82]. The majority of studies performed did not find the variants to be associated with improved clinical outcomes such as enhanced left ventricular ejection fraction [76, 83] although positive associations for Gln27Glu have been reported in a small study [84].

Many β-blockers, including propranolol, timolol, and metoprolol, are metabolized by CYP2D6, and loss of function variants (of which there are dozens) are very common. Poor metabolizers, 5-10 % of Caucasian and African subjects, carry two loss-of-function alleles, and generate unusually high plasma metoprolol concentrations with pronounced effects on blood pressure and heart rate lowering [85, 86] with the potential for ADRs (Table 2) [86]. In spite of only few studies providing evidence of a pharmacogenomic interaction, dose adjustments should be considered among patients prone to complications [87] which has also been acknowledged by the FDA [88]. Carvedilol is also a CYP2D6 substrate, but the poor metabolizer trait does not translate into variable clinical effects. Other β-blockers such as atenolol and nadolol do not require CYP2D6 to be metabolized.

A gain-of-function polymorphism (Glu41Leu, rs17098707) in the G protein-coupled receptor kinase 5 (GRK5) has been coupled with a beta-blocker-like phenotype, and appears to protect against catecholamine induced cardiomyopathy in mice [89]. The variant is more common among individuals of African descent and studies of metoprolol treated carriers of the Glu41Leu polymorphism experienced improved survival similar to untreated non-carriers (Leu41).

Variants in the α2c-adrenergic receptor (ADRA2C) have been associated with variable β-blocker responsiveness. A common four amino-acid deletion (Del322-325) has shown reduced ADRA2C activity in transfected cells [90] and has been associated with adverse outcomes in heart failure patients [91] although no effect was seen in the Beta-Blocker Evaluation of Survival Trial among patients on bucindolol (BEST) [78].

Clinical Implication

The findings of loss-of-function polymorphisms in CYP2D6 and the associated β-blocker response is equivocal and highlighted by the added FDA label. Thus, caution should be exerted among vulnerable heart failure patients who carry the loss-of-function in order to avoid ADRs. Moreover, there is some evidence of a pharmacogenomic interaction between β-blockers and the Arg389Gly polymorphism in ADBR1 although variable study findings make clinical interpretation difficult. Combining multiple risk alleles may provide more information [92].

Angiotensin-Converting Enzyme Inhibitors

Angiotensin-converting-enzyme inhibitors (ACE-I) are predominantly used to treat hypertension and have been associated with improved cardiovascular outcomes. A common insertion/deletion (I/D) polymorphism in the ACE gene (rs4646994) has been shown to influence ACE plasma concentrations [93]. An observational study reported that carriers of the DD genotype had a greater 10-year mortality risk compared to II carriers [94••]. However, these findings have not been replicated in prospective trials [95] or a large meta-analysis examining the influence of the variant on outcomes during ACE-I [96]. Thus, the proposed pharmacogenomic interaction between the I/D polymorphism in the ACE gene and ACE-I does not seem to be real and publication bias could have distorted any association [96]. Other candidates in the renin-angiotensin-aldosterone pathway, including the angiotensinogen (AGT) and angiotensin-II receptor types I and II (AGTR1 and AGTR2), have been interrogated and no definite pharmacogenomic associations have been identified.

The Perindopril Genetic Association study (PERGENE) sought to predict the treatment benefit of ACE-I and optimize ACE-I therapy by developing a genetic profile in 8907 CAD patients using 52 haplotype tagging polymorphisms across 12 genes [94••]. Two polymorphisms in the AGTR1 gene and one in the bradykinin type 1 receptor gene were significantly associated with improved treatment benefit during 4.2 years of follow-up (The primary composite endpoint was reduction in cardiovascular mortality, non-fatal myocardial infarction, and resuscitated cardiac arrest). Moreover, combining the three previously mentioned polymorphisms into a pharmacogenetic score produced a stepwise decrease in treatment benefit of perindopril with increasing scores [94••] which was replicated using a subset of patients from the Perindopril Protection Against Recurrent Stroke Study (PROGRESS) [95].

Clinical Implications

As of yet no definite pharmacogenomic targets have been identified for ACE-I responsiveness that merits routine genetic testing. However, the findings made in the PERGENE study could represent a true target if further validated.

Anti-Arrhythmic Drugs

Antiarrhythmic drugs that block the repolarizing potassium current I_Kr are among the commonest causes of a prolonged QT interval and are associated with increased risk of developing the malignant arrhythmia Torsade de Pointes (TdP). Importantly, the ability to produce a prolonged QT interval is not limited to cardiovascular drugs only [97]. Multiple polymorphisms in known long-QT genes have been reported among affected patients [98•]. Thus far, the evidence for causation is strongest for the D85N non-synonymous polymorphism (rs1805128) in KCNE1, encoding a potassium channel subunit; in a study of 176 cases of TdP, this variant conferred an odds ratio of 9.0 for developing TdP (Table 2) [99••].

Prioritized Pharmacogenomic Testing

While Clinical Laboratory Improvement Amendments (CLIA) approved laboratories do provide pharmacogenomic testing it is not always clear to clinicians which tests should be ordered, how to interpret the results, or if the results are even clinically actionable, particularly if test are ordered at the time of drug prescription. Given the rapid drop in genotyping costs, the CPIC approaches these issues from a different perspective, assuming that the data are already available from previous “preemptive” genotyping. CPIC provides a three-tier system (A, B, and C) to evaluate the level of evidence associated with a drug-gene interaction. “A” holds the strongest level of evidence and is assigned to pharmacogenetic signals where the “evidence includes consistent results from well-designed, well-conducted studies”. Pharmacogenomic signals that are assigned to “A” evidence level include: CYP2C19*2, CYP2C19*17, VKORC1: −1639G>A, CYP2C9*2, CYP2C9*3, and SLCO1B1: rs4149056. “B” is assigned where the evidence is sufficient to determine the effects, but the strength of the “evidence is limited by the number, quality, or consistency of the individual studies, by the inability to generalize to routine practice, or by the indirect nature of the evidence”. “C” is assigned where the evidence is insufficient to evaluate the effects on health and is not clinically actionable [100].

Conclusion

Fueled by the success of pharmacogenomic research the promise of tailoring diagnostic and treatment strategies to the needs of the individual patient may be within reach. Highlighting this exciting stage of rapid discovery is the development of clinical guidelines that incorporate genetic information to guide warfarin, clopidogrel, and statin therapy to reduce the risk of toxicity. Implementing these on a case by case and patient by patient basis is cumbersome and costly. With the plummeting cost of multiplexed genotyping, the idea of preemptively embedding DNA variant data in patient records, to be acted on when a target drug is prescribed, is beginning to be explored. Evaluation of the mechanics and healthcare outcomes in such systems, as well as ongoing clinical trials in pharmacogenomics and continuing genotype-phenotype association discovery, will determine how this personalized approach can be adopted.

Clinical Trial Acronyms

BEST

Beta-Blocker Evaluation of Survival Trial Among Patients on Bucindolol

PERGENE

Perindopril Genetic Association Study

PROGRESS

Perindopril Protection Against Recurrent Stroke Study