pgx required

Pharmacogenomics:

Increasing the safety and
effectiveness of drug therapy

Introduction to personalized medicine

The goal of personalized medicine is to individualize health care by using knowledge of
patients’ health history, behaviors, environments, and, most importantly, genetic variation
when making clinical decisions. Research on genetics has provided advances that can be
used to more accurately predict the risk of developing certain diseases, personalize screening
and surveillance protocols and, in some cases, prevent the onset of disease. In certain cases,
genetics can also be used to diagnose diseases and tailor therapies and disease management
strategies. These advancements in personalized medicine rely on knowledge of how a
patient’s genotype (genetic makeup) influences his or her phenotype (observable traits or
characteristics). Using the principles of personalized medicine, health care providers may be
better equipped to move beyond the “one-size-fits-all” approach that defined much of patient
care in the past, to care that is appropriate for unique patient subgroups.1

Pharmacogenomics

One of the most important components of personalized medicine is pharmacogenomics, the
study of genetic variations that influence individual response to drugs. Enzymes responsible
for drug metabolism and proteins that determine the cellular response to drugs (receptors)
are encoded by genes, and can therefore be variable in expression, activity level and
function when genetic variations are present. Knowing whether a patient carries any of
these variations may help health care professionals individualize drug therapy, decrease the
number of adverse drug reactions and increase the effectiveness of drugs. Pharmacogenomics
has been characterized as “getting the right dose of the right drug to the right patient at
the right time.”2 This brochure is intended to introduce the concept of pharmacogenomics
to physicians and other health care providers using a case-based approach. Note that the
terms “pharmacogenomics” and “pharmacogenetics” are often used interchangeably; for this
brochure, “pharmacogenomics” will be used.

Genes commonly involved in pharmacogenomic drug
metabolism and response

There are several genes responsible for differences in drug metabolism and response. Among
the most common are the Cytochrome P450 (CYP) genes, encoding enzymes that control
the metabolism of more than 70 percent of prescription drugs. People who carry variations
in certain CYP genes often do not metabolize drugs at the same rate or extent as in most
people, and this can influence response in many ways. Other genes known to affect drug
response encode the receptors for regulatory molecules such as neurotransmitters, hormones,
cytokines and growth factors, and cellular proteins such as enzymes, transporters, carriers,
ion channels, structural proteins and transcription factors. Variations in these genes can lead
to poor response and adverse drug reactions by disabling, inactivating, interfering with, or
inaccurately inducing the signaling mechanisms or cellular machinery that must function for
the body to respond properly to the drug; or by causing side effects that prevent continued
use of the drug.

2

Selected drugs whose safety and efficacy are affected
by gene variations

The table below is a partial list of drugs that exhibit reduced therapeutic effectiveness
and/or safety concerns in patients carrying certain genetic variations. These variations often
make the drug unsafe or unsuitable for patients who carry the variations. This list contains
examples of drugs that are affected by inherited genetic variations and by variations that are
acquired and present in tumor tissue.

Drug Gene(s) Drug Gene(s)

Clopidogrel (Plavix®) CYP2C19 Azathiopurine (Imuran®) TPMT

Atomoxetine (Strattera®) CYP2D6 Irinotecan (Camptosar®) UGT1A1

Codeine CYP2D6 Cetuximab (Erbitux®)* EGFR

Tamoxifen (Nolvadex®) CYP2D6 Erlotinib (Tarceva®)* EGFR

Warfarin (Coumadin®) CYP2C9, VKORC1 Imatinib mesylate (Gleevec®)* C-KIT

Abacavir (Ziagen®) HLA-B*5701 Panitumumab (Vectibix®)* KRAS

Carbamazepine (Tegretol®) HLA-B*1502 Trastuzumab (Herceptin®)* Her2/neu



* R esponse to these drugs is dependent on genetic variations that are present in tumor tissue.

Adapted from U.S. Food and Drug Administration website (www.fda.gov/Drugs/ScienceResearch/ResearchAreas/
Pharmacogenetics/ucm083378.htm). Accessed February 7, 2011.

Metabolizer phenotype

The metabolizer phenotype describes the patient’s ability to metabolize certain drugs and
is based on the number and type of functional alleles of certain genes that a patient carries.
These genes most commonly encode the CYP enzymes, which are the focus of much of this
brochure. The metabolizer phenotype can range from “poor,” used to describe patients with
little or no functional activity of a selected CYP enzyme, to “ultra-rapid,” used to describe
patients with substantially increased activity of a selected CYP enzyme. Depending on the
type of CYP variation present, the patient’s metabolizer phenotype and the type of drug
(active pharmacologic agent or inactive prodrug precursor), therapeutic drug response is
often suboptimal. The table on the following page summarizes the effects of CYP variation on
therapeutic efficacy.3 For example, poor metabolizers are unable to metabolize certain drugs
efficiently, resulting in a potentially toxic build-up of an active drug or the lack of conversion
of a prodrug into an active metabolite. In contrast, in ultra-rapid metabolizers, an active
drug is inactivated quickly, leading to a subtherapeutic response, while a prodrug is quickly
metabolized, leading to rapid onset of therapeutic effect.

3

Effects of CYP variants on therapeutic efficacy:

Metabolizer Active drug Prodrug
phenotype (inactivated by metabolism) (needs metabolism to produce
active metabolite)
Poor Drug enzyme Inactive Metabolite
Prodrug enzyme Active Metabolite
Ultra-rapid Increased efficacy;
active metabolite may accumulate; Decreased efficacy; prodrug may
usually require lower dose to avoid accumulate; may require lower dose
toxic accumulation to avoid toxic accumulation, or may
require alternate drug
Decreased efficacy;
active metabolite rapidly inactivated; Increased efficacy;
usually require higher dose to offset rapid onset of effect; may require
inactivation lower dose to prevent excessive
accumulation of active metabolite

Some drugs and foods cause altered metabolizer phenotype

Certain drugs mimic the effect of genetic variations, effectively causing changes in
metabolizer phenotype. For example, quinidine is an inhibitor of CYP2D6 activity. A patient
taking quinidine is therefore a CYP2D6 poor metabolizer, similar to someone who carries
a loss-of function variation in CYP2D6. In those patients, drugs that require the activity of
CYP2D6, such as atomoxetine, will not be metabolized at the same rate as in most people.4
Certain foods can also mimic the effects of genetic variations. One of the most common
examples is grapefruit juice, which is an inhibitor of CYP3A4. In people regularly drinking
grapefruit juice, drugs that require the activity of CYP3A4, such as diazepam, will not be
metabolized at the same rate as in most people.4

Pharmacogenomics in the clinical setting

Awareness of the influence of gene variations on patient response to certain drugs can help
physicians decide which type of drug therapy may be appropriate, and identify cases in which
a patient isn’t responding as anticipated to a drug. The examples that follow illustrate three
categories for which pharmacogenomic knowledge can help inform therapeutic decisions:
predicting and preventing adverse reactions, determining the efficacy of a drug for a particular
patient, and predicting the optimal drug dose.

Using pharmacogenomics to predict and prevent
adverse drug reactions

Several drugs can cause severe or life-threatening reactions in patients with variations
in genes that encode proteins that metabolize or are targets of the drugs. Knowing about
patients’ genetic variations can help physicians avoid drugs that may cause adverse reactions.
On the following two pages are examples of drugs in this category.

4

Abacavir

Abacavir (Ziagen®) is a nucleoside reverse transcriptase inhibitor used in combination with
other antiretrovirals to treat HIV infection.5 An immunologically-mediated hypersensitivity
reaction occurs in 5–8 percent of patients taking abacavir, usually during the first six weeks
after initiation of therapy. The hypersensitivity symptoms include a combination of fever,
rash, gastrointestinal tract symptoms and respiratory symptoms that become more severe with
continued dosing.6 Discontinuation of abacavir therapy results in reversal of symptoms.
The abacavir hypersensitivity described above is associated with a variant allele of the major
histocompatibility complex, HLA-B*5701. Patients who carry the HLA-B*5701 allele have
an increased risk for developing a hypersensitivity reaction.6 HLA-B*5701 screening before
abacavir treatment results in a significantly reduced number of hypersensitivity cases.6
The abacavir product labeling recommends genetic testing to detect the presence of
HLA-B*5701.5 In the “Warnings and Precautions” section, the labeling states:

Serious and sometimes fatal hypersensitivity reactions have been associated with ZIAGEN and
other abacavir-containing products. Patients who carry the HLA-B*5701 allele are at high risk
for experiencing a hypersensitivity reaction to abacavir. Prior to initiating therapy with abacavir,
screening for the HLA-B*5701 allele is recommended; this approach has been found to decrease
the risk of a hypersensitivity reaction. Screening is also recommended prior to reinitiation of
abacavir in patients of unknown HLA-B*5701 status who have previously tolerated abacavir.
For HLA-B*5701-positive patients, treatment with an abacavir-containing regimen is not
recommended and should be considered only with close medical supervision and under exceptional
circumstances when the potential benefit outweighs the risk. (Label updated September 2010)

Case study
JS is a 35 year-old man who has recently been diagnosed as HIV positive. Before initiating abacavir
anti-retroviral therapy, his physician orders genetic testing to determine whether he carries the
HLA-B*5701 allele, knowing that JS would develop fever, rash, nausea and fatigue if he carried the
variant allele. The test confirmed that JS carries the HLA-B*5701 allele.
How did genetic testing help JS and his physician?
Knowing that JS carried the HLA-B*5701 variation indicated that he would likely experience a
hypersensitivity reaction. JS’s physician will probably not include abacavir in his antiretroviral
regimen.

5

Codeine

Codeine is a prodrug with analgesic properties due primarily to its conversion into
morphine.7,8 Conversion to morphine is mediated by the cytochrome P450 enzyme CYP2D6.
Variations that decrease the metabolic activity of CYP2D6 result in a poor analgesic
response due to the reduced conversion of codeine into morphine,3 and patients carrying
such a variation are considered poor metabolizers and receive little therapeutic benefit from
codeine. It is estimated that 5–10 percent of Caucasians are CYP2D6 poor metabolizers; the
percentage is approximately 2–3 percent in other racial and ethnic groups.7,9,10

Variations (such as gene duplications) can also result in increased metabolic activity of
CYP2D6; these result in an enhanced analgesic response due to the rapid conversion of
codeine into morphine. Patients who carry such variations are at risk for opioid toxicity,
which includes moderate to severe central nervous system depression. The prevalence of
the CYP2D6 ultra-rapid metabolizer phenotype has been estimated at 1–10 percent in
Caucasians, 3–5 percent in African Americans, 16–28 percent in North Africans, Ethiopians
and Arabs, and up to 21 percent in Asians.3,11

The product labeling (updated in July 2009) of drugs containing codeine include warnings
that CYP2D6 ultra-rapid metabolizers “may experience overdose symptoms such as extreme
sleepiness, confusion or shallow breathing, even at labeled dosage regimens,” and encourages
physicians to “choose the lowest effective dose for the shortest period of time and inform
their patients about the risks and the signs of morphine overdose.”11 Of additional concern is
the use of codeine in nursing mothers. Product labeling includes the statement that women
who are CYP2D6 ultra-rapid metabolizers achieve “higher-than-expected levels of morphine
in breast milk and potentially dangerously high serum morphine levels in their breastfed
infants.” A 2007 U.S. Food and Drug Administration (FDA) Public Health Advisory warns
about the dangers to neonates of codeine use in breast-feeding mothers and states that
the only way to determine prior to drug administration whether a patient is an ultra-rapid
metabolizer is by the use of a genetic test.12 Canadian guidelines also state that genetic testing
is available.13

Case study

DS is a 30 year-old woman who gave birth by caesarian section 10 days ago. Her physician
prescribed codeine for post-caesarian pain. Despite taking no more than the prescribed dose, DS
experienced nausea and dizziness while she was taking codeine. She also noticed that her breastfed
infant was lethargic and feeding poorly. When DS mentioned these symptoms to her physician,
he recommended that she discontinue codeine use. Within a few days, both DS’s and her infant’s
symptoms were no longer present.

How would genetic testing help DS and her physician?
Genotyping of DS’s CYP2D6 gene may have revealed a duplication of CYP2D6 genes, placing her
in the ultra-rapid metabolizer category. Armed with this knowledge, DS’s physician would likely
have prescribed a different analgesic that would have spared DS and her infant the symptoms of
opioid toxicity.

6

Using pharmacogenomics to predict effectiveness

Several drugs are subtherapeutic or ineffective in patients with variations in genes that
encode drug-metabolizing proteins or targets of the drugs. Knowing about patients’ genetic
variations can help physicians select drug therapies that will be most effective for individual
patients. Below is an example of a drug in this category.

Clopidogrel

Clopidogrel (Plavix®) is a platelet inhibitor used in the treatment of a number of
cardiovascular diseases. It is often prescribed for secondary prevention following acute
coronary syndromes and for those undergoing percutaneous coronary intervention. However,
despite clopidogrel treatment, up to one-quarter of patients experience a subtherapeutic
antiplatelet response, resulting in a higher risk for ischemic events.14,15

Clopidogrel is a prodrug; its antiplatelet properties are exerted once it is converted to an
active metabolite.16 A cytochrome P450 enzyme, CYP2C19, mediates the conversion of
clopidogrel into the active metabolite. Patients who carry certain variations in CYP2C19 are
considered poor metabolizers and show reduced ability to convert clopidogrel into its active
metabolite, resulting in a diminished antiplatelet effect.16,17 Further, these patients are more
likely to have an ischemic event following clopidogrel therapy.17 Approximately 2–20 percent
of patients (depending on ethnicity) are likely to carry CYP2C19 variations.3,18

A Boxed Warning is included in the product labeling to alert health care providers about its
reduced effectiveness in patients who are CYP2C19 poor metabolizers, and to inform health
care providers that genetic tests are available to detect genetic variations in CYP2C1918:

WARNING: DIMINISHED EFFECTIVENESS IN POOR METABOLIZERS
The effectiveness of Plavix is dependent on its activation to an active metabolite by the
cytochrome P450 (CYP) system, principally CYP2C19. Plavix at recommended doses forms less
of that metabolite and has a smaller effect on platelet function in patients who are CYP2C19
poor metabolizers. Poor metabolizers with acute coronary syndrome or undergoing percutaneous
coronary intervention treated with Plavix at recommended doses exhibit higher cardiovascular
event rates than do patients with normal CYP2C19 function. Tests are available to identify a
patient’s CYP2C19 genotype; these tests can be used as an aid in determining therapeutic strategy.
Consider alternative treatment or treatment strategies in patients identified as CYP2C19 poor
metabolizers. (Label updated February 2011).

7

Case study

JM is a 58 year-old man who recently had an acute myocardial infarction. To prevent subsequent
ischemic events, JM’s physician recommends antiplatelet therapy and prescribes clopidogrel. Six
months later, JM suffered another acute MI, and his physician suspects that the patient has been
non-adherent, or alternatively, that clopidogrel therapy may have been ineffective.

How would genetic testing help JM and his physician?
Determining JM’s CYP2C19 genotype may reveal that he carries a variant that diminishes the
antiplatelet effect of clopidogrel. If this were the case, alternative anti-platelet therapies may have
been considered, reducing the chance that JM would suffer a second cardiac event.

In patients taking both clopidogrel and the proton pump inhibitors omeprazole or esomeprazole,
diminished antiplatelet activity and adverse cardiac outcomes have been observed.19 It is thought
that omeprazole and esomeprazole inhibit the activity of CYP2C19, making patients who take
these drugs de facto CYP2C19 poor metabolizers, similar to someone who carries a loss-of function
variation in CYP2C19. In these patients, drugs that require the activity of CYPC19, such as
clopidogrel, will not be metabolized at the same rate as in most people.

Using pharmacogenomics to predict optimal dose

Genetic variations can lead to an altered dosage regimen. Knowing whether a patient carries these
genetic variations can assist physicians in determining optimal therapeutic dose. An example of
a drug with an altered dosage regimen in patients with certain genetic variations is warfarin.

Warfarin

Warfarin (Coumadin®) is the most widely-prescribed anticoagulant used to treat and prevent
thromboembolic diseases. It is metabolized by the CYP2C9 enzyme, and its anticoagulant
effect is mediated by the enzyme VKORC1. Variation in the CYP2C9 gene causes some
patients to have slow metabolism of warfarin and a longer half-life of the drug, resulting in
higher than usual blood concentrations of warfarin and greater anticoagulant effect. Certain
variations in the VKORC1 gene result in reduced activity of the enzyme and subsequently
reduced synthesis of coagulation factors. The combination of slow warfarin metabolism caused
by CYP2C9 gene variations and reduced coagulation caused by VKORC1 gene variations
increases the risk of bleeding during warfarin therapy.

Warfarin has a narrow therapeutic index; variations in CYP2C9 and VKORC1, in addition
to several other patient characteristics, make it difficult to predict the effective dose. Those
carrying certain CYP2C9 and VKORC1 variations are likely to require altered doses and may
require prolonged time to reach a stable maintenance dose. The prevalence of these CYP2C9
and VKORC1 variations is variable among racial and ethnic groups: up to 17 percent of
Caucasians, 4 percent of African-Americans and less than 2 percent of Asians carry at least
one variant of CYP2C9;20 while 37 percent of Caucasians, 14 percent of African-Americans,
and 89 percent of Asians carry at least one variant in VKORC1.21

8

The warfarin product labeling (updated January 2010) states “The patient’s CYP2C9 and
VKORC1 genotype information, when available, can assist in selection of the starting dose,”
and includes the following table of expected therapeutic doses for patients with different
combinations of CYP2C9 and VKORC1 variations:22

Range of Expected Therapeutic Warfarin Doses (in mg) for CYP2C9 and VKORC1
Genotypes†

VKORC1 CYP2C9 genotype
genotype
*1/*1 *1/*2 *1/*3 *2/*2 *2/*3 *3/*3

GG 5–7 5–7 3–4 3–4 3–4 0.5–2

AG 5–7 3–4 3–4 3–4 0.5–2 0.5–2

AA 3–4 3–4 0.5–2 0.5–2 0.5–2 0.5–2

†R anges are derived from multiple published clinical studies. Other clinical factors (e.g., age, race, body weight,
sex, concomitant medications, and comorbidities) are generally accounted for, along with genotype, in the ranges
expressed in the table.

Adapted from Warfarin drug labeling, 2010.22

The product labeling suggests that this table of expected therapeutic doses can be used to
assist prescribers in choosing initial warfarin dose for patients whose CYP2C9 and VKORC1
genotypes are known. For example, if genetic testing reveals that a patient carries the *1/*3
variation of CYP2C9 and the AG variation of VKORC1, his or her expected therapeutic
warfarin dose is likely to be in the 3–4 mg range.

Dosing algorithms that rely on clinical features such as age, sex and weight, along with
genotype, can also assist in the determination of optimal dose (visit www.warfarindosing.org,
where one such algorithm can be found). Genotype is one factor of many that contribute
to variation in a patient’s response to warfarin. Careful monitoring of INR is still required
during titration to steady state and monitoring of long-term therapy.

Case study

ML is a 65 year-old woman who has recently been diagnosed with atrial fibrillation. To reduce
the risk of stroke and other thrombotic events, ML’s physician recommends warfarin therapy. To
estimate the initial dose, ML’s clinical characteristics (age, sex, weight, diet) were considered.
However, ML will need to return to the clinic every day for INR monitoring until a stable dose is
determined, and then every few weeks thereafter for maintenance monitoring.
How would genetic testing help ML and her physician?
Determination of ML’s CYP2C9 and VKORC1 genotype would reveal whether she carries any
variations that alter her ability to metabolize and respond to warfarin. Knowing about any
gene variations before initiating therapy allows for more accurate initial dosing and faster INR
stabilization, and can reduce the risk for bleeding or clotting events.

9

Pharmacogenomics and genetic testing

Genetic tests that detect variations in the genes that control metabolism and response
to certain drugs are commercially available. Costs and insurance coverage are variable,
depending on the type of test ordered. Your hospital and reference laboratories as well
as clinical geneticists are valuable sources of information on the best test to order and
interpretation of results.

It is important to note that the field of pharmacogenomics is still developing, with new
findings being rapidly reported. Clinical trials and other studies focusing on the benefits, risks
and cost-effectiveness of using genetic information to inform drug therapy are underway. In
the meantime, physicians and health care providers should be familiar with the concept that
genetic variations can cause their patients to respond unexpectedly to drug therapy, and that
in some cases, it may be appropriate to use genetic testing to guide therapeutic decisions. For
more information on pharmacogenomics, visit the following websites:

• UC San Diego Skaggs School of Pharmacy and Pharmaceutical Sciences
Pharmacogenomics Education Program (http://pharmacogenomics.ucsd.edu/home.aspx)

• A merican College of Clinical Pharmacology
The Future of Medicine: Pharmacogenomics (http://user.accp1.org/Sample_Home.htm)

• Indiana University School of Medicine, Division of Clinical Pharmacology
Defining Genetic Influences on Pharmacologic Responses (www.drug-interactions.com)

• U.S. Food and Drug AdministrationGenomics (www.fda.gov/Drugs/ScienceResearch/
ResearchAreas/Pharmacogenetics/default.htm)

• Pharmacogenomics Research Network (www.nigms.nih.gov/initatives/pgrn)

Acknowledgements

The American Medical Association, the Arizona Center for Education and Research on
Therapeutics (AzCERT), and the Critical Path Institute thank the American Academy
of Family Physicians, the American College of Clinical Pharmacy, the American Society
for Clinical Pharmacology and Therapeutics, Medco Research Institute, the Personalized
Medicine Coalition, and several individuals with expertise in pharmacogenomics for their
helpful review comments. This brochure was partially funded by a grant from the Agency
for Healthcare Research and Quality to AzCERT. The authors of this document declare no
conflicts of interest. The information contained in this document is not intended to be a
treatment recommendation or an endorsement of any particular product.

10

References

1. C ollins FS. 2010. The Future of Personalized 13. M adadi P, Moretti M, Djokanovic N, Bozzo P,
Medicine. NIH Medline Plus 5(1):2–3. Nulman I, Ito S, Koren G. (2009). Guidelines for
maternal codeine use during breastfeeding. Can
2. L esko L. Regulatory Perspective on Warfarin Fam Physician 55(11):1077–1078.
Relabeling with Genetic Information. May 2007.
Presentation at the 9th National Conference 14. E llis KJ, Stouffer GA, McLeod HL, Lee CR.
on Anti-Coagulation Therapy. www.fda.gov/ Clopidogrel pharmacogenomics and risk of
downloads/Drugs/ScienceResearch/ResearchAreas/ inadequate platelet inhibition: U.S. FDA
Pharmacogenetics/ucm085563.pdf. Accessed June recommendations. (2009). Pharmacogenomics
23, 2010. 10(11):1799–1817.

3. B elle DJ, Singh H. (2008). Genetic Factors in Drug 15. S imon T, Verstuyft C, Mary-Krause M, Quteineh
Metabolism. Am Fam Physician 77(11):1553–1560. L, Drouet E, Meneveau N et al. (2009). Genetic
determinants of response to clopidogrel and
4. Indiana University Division of Clinical cardiovascular events. N Engl J Med 360(4):
Pharmacology. P450 Drug interaction Table. 2010. 363–375.
http://medicine.iupui.edu/clinpharm/ddis/table.asp.
Accessed June 23, 2010. 16. S huldiner AR, O’Connell JR, Bliden KP,
Gandhi A, Ryan K, Horenstein RB et al. (2009)
5. G laxoSmithKline. Ziagen® (abacavir sulfate) Association of cytochrome P450 2C19 genotype
product labeling. 2010. http://us.gsk.com/products/ with the antiplatelet effect and clinical efficacy of
assets/us_ziagen.pdf. Accessed February 8, 2011. clopidogrel therapy. JAMA 302(8):849–857.

6. M allal S, Phillips E, Carosi G, Molina JM, 17. M ega JL, Close SL, Wiviott SD, Shen L,
Workman C, Tomazic J et al. (2008). HLA-B*5701 Hockett RD, Brandt JT et al. Cytochrome p-450
screening for hypersensitivity to abacavir. N Engl J polymorphisms and response to clopidogrel.
Med 358(6):568–579. N Engl J Med 2009; 360(4):354–362.

7. Madadi P, Ross CJD, Hayden MR, Carleton 18. B ristol Meyers Squibb/Sanofi Pharmaceuticals
BC, Gaedigk A, Leeder JS, Koren G. (2009). Partnership. Plavix® (clopidogrel bisulfate) product
Pharmacogenomics of Neonatal Opioid Toxicity labeling. 2011. www.accessdata.fda.gov/drugsatfda_
Following Maternal Use of Codeine During docs/label/2011/020839s052lbl.pdf. Accessed
Breastfeeding: A Case-Control Study. Clin February 8, 2011.
Pharmacol Ther 85(1):31–35.
19. H o PM, Maddox TM, Wang L, Fihn SD, Jesse RL,
8. W illmann S, Edginton AN, Coboeken K, Ahr A, Peterson ED, Rumsfeld JS. (2009). Risk of Adverse
Lippert J. (2009). Risk to the Breast-Fed Neonate Outcomes Associated With Concomitant Use of
From Codeine Treatment to the Mother: A Clopidogrel and Proton Pump Inhibitors Following
Quantitative Mechanistic Modeling Study. Clin Acute Coronary Syndrome. JAMA 301(9):937–944.
Pharmacol Ther 86(6):634–643.
20. G age, BF. (2006). Pharmacogenetics-Based
9. R ollason V, Samer C, Piguet V, Dayer P, Desmeules Coumarin Therapy. Hematology ASH Educ Prog
J. (2008). Pharmacogenomics of analgesics: Book. 467–473.
Toward the individualization or prescription.
Pharmacogenomics 9(7):905–933. 21. R eider MJ, Reiner AP, Gage BF, Nickerson
DA, Eby CS, McLeod HL et al. (2005). Effect
10. Ingelman-Sundberg M, Sim SC, Gomez A, of VKORC1 Haplotypes on Transcriptional
Rodriguez-Antona C. (2007). Influence of Regulation and Warfarin Dose. N Engl J Med
cytochrome P450 polymorphisms on drug therapies: 352(22):2285–2293.
Pharmacogeneic, pharmacoepigenetic, and clinical
aspects. Clin Pharmacol Ther 116:496–526. 22. B ristol Meyers Squibb Company. Coumadin®
(warfarin sodium) product labeling. 2010.
11. Roxane Laboratories Inc. Codeine sulfate product www.accessdata.fda.gov/drugsatfda_docs/
labeling. 2009. www.accessdata.fda.gov/drugsatfda_ label/2010/009218s108lbl.pdf. Accessed
docs/label/2009/022402s000lbl.pdf. Accessed February 8, 2011.
March 3, 2010.

12. U S Food and Drug Administration. (2007).
Public Health Advisory: Use of Codeine By
Some Breastfeeding Mothers May Lead To
Life-Threatening Side Effects In Nursing
Babies. http://www.fda.gov/Drugs/DrugSafety/
PostmarketDrugSafetyInformationfor Patientsand
Providers/default.htm. Accessed February 8, 2011.

11

For more information contact:

American Medical Association
www.ama-assn.org/go/genetics
The American Medical Association, with its mission to promote the art and science
of medicine and the betterment of public health, is a non-profit organization that
unites physicians nationwide to work on the most important professional and public
health issues.

Arizona Center for Education and Research on Therapeutics
www.azcert.org
The Arizona Center for Education and Research on Therapeutics is an independent,
collaborative program of the Critical Path Institute and the University of Arizona
College of Pharmacy. Its mission is to improve therapeutic outcomes and reduce
adverse events caused by drug interactions.

Critical Path Institute
www.c-path.org
The Critical Path Institute is an independent, non-profit organization dedicated to
bringing scientists from the FDA, industry and academia together to improve the
path for innovative new drugs, diagnostic tests and devices to reach patients in need.

© 2011 American Medical Association. All rights reserved.

10-0290:5/11:jt

REVIEW

CME EDUCATIONAL OBJECTIVE: Readers will discern the implications of pharmacogenomic testing as it emerges

CREDIT

JOSEPH P. KITZMILLER, MD, PhD DAVID K. GROEN, MD MITCH A. PHELPS, PhD WOLFGANG SADEE, Dr rer nat

Department of Pharmacology, Department of Pharmacology, Division of Pharmaceutics Resources, Chairman, Department of Pharmacology;
Division of Clinical Trials, College of Medicine, Division of Clinical Trials, College of College of Pharmacy, The Ohio State Director, Program in Pharmacogenomics,
The Ohio State University, Columbus, OH Medicine, The Ohio State University, University, Columbus, OH College of Medicine, The Ohio State University,
Columbus, OH Columbus, OH

Pharmacogenomic testing:
Relevance in medical practice

Why drugs work in some patients but not in others

■■ABSTRACT I n many patients, certain drugs do not
work as well as expected, whereas in other
Genetics may account for much of the variability in our patients they cause toxic effects, even at lower
patients’ responses to drug therapies. This article offers doses. For some patients, the reason may be
the clinician an up-to-date overview of pharmacoge- genetic.
nomic testing, discussing implications and limitations of Sizeable minorities of the population carry
emerging validated tests relevant to the use of warfa- genetic variants—polymorphisms—­ that affect
rin (Coumadin), clopidogrel (Plavix), statins, tamoxifen their response to various drugs. Thanks to ge-
(Nolvadex), codeine, and psychotropic drugs. It also netic research, our understanding of the vari-
discusses the future role of pharmacogenomic testing in ability of drug response has advanced markedly
medicine. in the last decade. Many relevant polymor-
phisms have been identified, and tests for some
■■KEY POINTS of them are available.
See related editorial, page 241
Polymorphisms that affect the pharmacokinetics and Armed with the knowledge of their pa-
pharmacodynamics of specific drugs are common. tients’ genetic status, physicians could predict
Testing for certain polymorphisms before prescribing their response to certain drugs, leading to bet-
certain drugs could help avoid adverse drug effects and ter efficacy, fewer adverse drug reactions, and a
improve efficacy. better cost-benefit ratio.
Pharmacogenomic testing has only recently begun to The possible impact is substantial, since
enter clinical practice, and routine testing is currently more than half of the drugs most commonly
limited to a few clinical scenarios. However, additional involved in adverse drug reactions are metabo-
applications may be just around the corner. lized by polymorphic enzymes.1 Adverse drug
Many pharmacogenomic tests are available, but testing reactions remain a significant detriment to
has not yet been recommended for most drugs. Needed public health, having a substantial impact on
are large-scale trials to show that routine testing im- rates of morbidity and death and on health-
proves patient outcomes. care costs.2–8 In the United States, adverse
drug reactions are a leading cause of death
doi:10.3949/ccjm.78a.10145 in hospitalized patients4 and are annually re-
sponsible for hundreds of thousands of deaths
and hundreds of billions of dollars in added
costs.2,4,6–8
But the era of truly individualized medicine
is not here yet. For most drugs, pharmacoge-
nomic testing has not been endorsed by expert
committees (and insurance companies will not

  243CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

PHARMACOGENOMIC TESTING

pay for it), since we still lack evidence that Most sequence variations are single nu-
clinical outcomes improve. This, we hope, cleotide polymorphisms (SNPs, pronounced
will change as ongoing clinical trials are com- “snips”), a single DNA base pair substitution
pleted. FIGURE 1 describes the various stages that may result in a different gene product.
involved in translational pharmacogenomic SNPs can be classified as structural RNA
research.11 polymorphisms (srSNPs), regulatory polymor-
In the meantime, physicians can educate phisms (rSNPs), or polymorphisms in coding
their patients and promote efforts to incorpo- regions (cSNPs)10: srSNPs alter mRNA pro-
rate genomic information into standard clini- cessing and translation, rSNPs alter transcrip-
cal decision-making. tion, and cSNPs alter protein sequence and
This article offers an overview of pharma- function.
cogenomic testing, discussing implications Recently, genetic associations with a phe-
and limitations of a few validated tests. Spe- notype have been done on a large scale, with
cifically, we will discuss testing that is relevant millions of SNPs measured in each of many
when using warfarin (Coumadin), clopidogrel subjects. This approach, called a genome-
(Plavix), statins, tamoxifen (Nolvadex), co- wide association study or GWAS, has revealed
deine, and psychotropic medications, as well countless candidate genes for clinical traits,
as the future role of pharmacogenomic testing but only a few have resulted in a practical
in medicine. clinical application. SNPs may by themselves
exert a pharmacokinetic effect (ie, how the
■■ WHAT IS PHARMACOGENOMICS? body processes the drug), a pharmacodynamic
effect (ie, how the drug affects the body), or
Pharmacogenomics is the study of how genet- both, or they may act in concert with other
ic factors relate to interindividual variability genetic factors. Pharmacodynamic effects can
Pharmaco- of drug response. result from a pharmacokinetic effect or can re-
Many clinicians may not be familiar with sult from variations in a pharmacologic target.
the background and terminology used in the Establishing a genotype-phenotype associa-
pharmacogenomic literature. Below, a brief re- tion can involve clinical studies, animal trans-
genomic testing view of the terminology is followed by a sche- genic studies, or molecular and cellular func-
is beginning matic describing the various stages of research tional assays.
involved in pharmacogenomics and the ad- Clinical applications are emerging
to affect vancement of a test into standard practice. Although pharmacogenomic testing is begin-
The review and schematic may be help- ning to affect the way medicine is practiced,
the way ful for evaluating the clinical significance of it is recommended, or at least strongly sug-
medicine pharmacogenomics-related articles. gested, by labeling mandated by the US Food
and Drug Administration (FDA) for only
is practiced From genotype to phenotype a few clinical scenarios, mostly in the treat-
Genotype refers to the coding sequence of ment of cancer and human immunodeficiency
DNA base pairs for a particular gene, and phe- virus (TABLE 1). However, applications are also
notype (eg, disease or drug response) refers to a being developed for a few widely prescribed
trait resulting from the protein product encod- drugs and drug classes in primary care. We
ed by the gene. The name of a gene often re- will therefore focus our discussion on the ad-
fers to its protein product and is italicized (eg, vantages and limitations of a few of these ex-
the CYP3A4 gene encodes for the CYP3A4 amples for which clinical applications may be
enzyme). emerging.
Two alleles per autosomal gene (one pa- ■■ WARFARIN:
ternal and one maternal) form the genotype.
Heterozygotes possess two different alleles, and IMPORTANCE OF CYP2C9, VKORC1
homozygotes possess two of the same alleles. Warfarin is used for the long-term treatment
The most common allele in a population is and prevention of thromboembolic events.
referred to as the wild type, and allele frequen-
cies can vary greatly in different populations.9

244  CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

KITZMILLER AND COLLEAGUES

How a genetic test becomes a standard of care:
Typical advancement pathway

Identify candidate genetic variant
The US Food and Drug Administration (FDA) lists
genetic tests validated for accuracy and reproduc-
ibility at www.fda.gov/Drugs/ScienceResearch/
ResearchAreas/Pharmacogenetics.

Determine penetrance of genetic variant Determine genotype-phenotype association
Allele frequency is an important determinant Measured biologic or physiologic effects (associa-
of clinical significance; even genetic variants tions) are often viewed as surrogates for clinical
associated with large pharmacokinetic or response; however, clinical responses do not
pharmacodynamic effects may not have sufficient always correspond as predicted.
clinical utility if they occur only rarely.

Determine clinical significance
Clinical trials that compare implementation
of genetic testing to the current standard of care
should be used for determining clinical significance.

Determine cost-effectiveness About 20%
The cost-effectiveness of implementing a pharmacoge- of patients
nomic testing application should compare favorably starting
with current standard practice. Cost-effectiveness warfarin
assessments measure a variety of outcomes, including are hospitalized
clinical events and quality-adjusted life-years. in the first
Standard of care 6 months
The FDA can require genetic information to be added because of
to a product label (eg, codeine), recommend the bleeding
use of genetic testing data if available (eg, warfarin
[Coumadin]), or even recommend genetic testing
prior to prescribing a medication to specific patient
populations—eg, HLA-B genotyping prior to prescrib-
ing carbamazepine (Tegretol, Equetro) to genetically
at-risk patients.11

FIGURE 1

This drug has a narrow therapeutic win- The findings from a recent study suggest
dow and shows substantial interpatient dose that pharmacogenomic testing may eventu-
variability. The start of warfarin therapy is ally allow more patients to safely benefit from
associated with one of the highest rates of warfarin therapy. In this large, nationwide,
adverse events and emergency room visits prospective study, hospitalization rates were
of any single drug.12 More than 2 million pa- 30% lower when pharmacogenomic testing
tients start warfarin each year in the United was used.14 However, no reduction was seen in
States alone,13 and about 20% of them are the time needed to reach the target interna-
hospitalized within the first 6 months because tional normalized ratio (INR) or in the need
of bleeding due to overanticoagulation.14 for INR checks at 6 months. Furthermore,

  245CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

PHARMACOGENOMIC TESTING

TABLE 1 this study used historical control data, and
FDA-mandated product labeling: some or all of the reduction in hospitalization
Genetic biomarkers and their clinical context rates may be attributed to more frequent INR
checks in the patients who underwent testing
C-KIT than in the historical control group.
(cytokine receptor) A relationship between warfarin dose re-
Imatinib mesylate (Gleevec) is indicated for the treatment of pa- quirements and the genetic status of CYP2C9,
tients with Kit (CD117)-positive unresectable tumors or metastatic which encodes a major drug-metabolizing en-
malignant gastrointestinal stromal tumors (GIST). zyme, has been demonstrated in retrospective
CCR5 and prospective studies.15–17
(C-C chemokine receptor type 5) S-warfarin is metabolized by CYP2C9,
Maraviroc (Selzentry) is indicated for treatment-experienced adult which is polymorphic
patients infected with CCR5-tropic HIV-1 infection. Warfarin contains equal amounts of two iso-
Chromosome 5q mers, designated S and R. S-warfarin, which
Lenalidomide (Revlimid) is indicated for the treatment of patients is more potent, is metabolized principally by
with transfusion-dependent anemia due to low- or intermediate-risk CYP2C9, while R-warfarin is metabolized by
myelodysplastic syndromes associated with a deletion 5q cytogenic CYP1A2, CYP2C19, and CYP3A4.
abnormality with or without additional cytogenic abnormalities. People who possess two copies of the wild
Familial hypercholesterolemia type CYP2C9 gene CYP2C9*1 metabolize
Atorvastatin (Lipitor) dosage adjustment is necessary for children warfarin very well and so are called “extensive
who have homozygous or heterozygous familial hypercholesterol- warfarin metabolizers.” Carriers of the allelic
emia. variants CYP2C9*2 and CYP2C9*3 (which
G6PD have point mutations in exons 3 and 7 of CY-
(glucose-6-phosphate dehydrogenase) P2C9, respectively), have less capacity. Com-
Testing for G6PD deficiency is recommended in high-risk popula- pared with those who are homozygous for the
tions before starting treatment with rasburicase (Elitek). wild-type gene, homozygous carriers of CY-
HER2/neu P2C9*3 clear S-warfarin at a rate that is 90%
(human epidermal growth factor receptor 2) lower, and those with the CYP2C9*1/*3, CY-
HER2 testing is recommended before starting treatment with P2C9*1/*2, CYP2C9*2/*2, or CYP2C9*2/*3
trastuzumab (Herceptin). genotypes clear it at a rate 50% to 75% lower.
HLA-B*1502 A meta-analysis of 12 studies found that the
(major histocompatibility complex, class I, B) CYP2C9 genotype accounted for 12% of the
HLA-B testing is recommended in high-risk populations before interindividual variability of warfarin dose re-
starting treatment with carbamazepine (Tegretol, Equetro). quirements.18
HLA-B*5701 About 8% of whites carry at least one copy
(major histocompatibility complex, class I, B) of CYP2C9*2, as do 1% of African Americans;
HLA-B testing is recommended before starting treatment with the allele is rare in Asian populations. The
abacavir (Ziagen). frequency of CYP2C9*3 is 6% in whites, 1%
Philadelphia (Ph1) chromosome in African Americans, and 3% in Asians.19,20
Dasatinib (Sprycel) is indicated for treatment of adults with People with CYP2C9*4 or CYP2C9*5 have
Philadelphia-chromosome-positive acute lymphoblastic leukemia a diminished capacity to clear warfarin; how-
with resistance or intolerance to prior therapy. ever, these variants occur so infrequently that
TPMT their clinical relevance may be minimal.
(thiopurine methyltransferase) Warfarin’s target, VKOR,
TPMT testing is recommended before starting treatment with is also polymorphic
azathioprine (Imuran). Genetic variation in warfarin’s pharmaco-
logic target, vitamin K 2,3-epoxide reductase
Adapted from US Food and Drug Administration (FDA). Table of pharmacogenomic (VKOR), also influences dose requirements.
biomarkers in drug labels. www.fda.gov/Drugs/ScienceResearch/ResearchAreas/ Warfarin decreases the synthesis of vitamin-
Pharmacogenetics/ucm083378.htm. Accessed 2/3/2011. K-dependent clotting factors by inhibiting

246  CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

KITZMILLER AND COLLEAGUES

TABLE 2
Range of expected warfarin doses
based on CYP2C9 and VKORC1 genotypes

VKORC1 *1/*1 *1/*2 *1/*3 CYP2C9 genotype *2/*3 *3/*3
genotype *2/*2

GG 5–7 mg 5–7 mg 3–4 mg 3–4 mg 3–4 mg 0.5–2 mg

AG 5–7 mg 3–4 mg 3–4 mg 3–4 mg 0.5–2 mg 0.5–2 mg

AA 3–4 mg 3–4 mg 0.5–2 mg 0.5–2 mg 0.5–2 mg 0.5–2 mg

This table and the following comments were taken directly from the warfarin prescribing information. Ranges are derived from
multiple published clinical studies. Other factors (eg, age, race, body weight, sex, concomitant medications, and comorbidities) are
generally accounted for along with genotype in the ranges expressed in the table. The VKORC1 –1639G>A (rs9923231) variant is
used in this table. Other coinherited VKORC1 variants may also be important determinants of warfarin dose. Patients with CYP2C9
*1/*3, *2/*2, *2/*3, and *3/*3 may require a prolonged time (< 2 to 4 weeks) to achieve the maximum international normalized
ratio effect for a given dosage regimen.24

VKOR. This inhibition depends on the pa- CYP2C9 and VKORC1 testing is available
tient’s C1 subunit gene, VKORC1. Patients Currently, the clinical pharmacogenetic tests
with a guanine-to-adenine SNP 1,639 bases relevant for warfarin use are for CYP2C9 and
upstream of VKORC1 (–1639G>A) need VKORC1.10
lower warfarin doses—an average of 25% low- The FDA has approved four warfarin phar-
er in those with the GA genotype (ie, one al- macogenetic test kits, but most third-party pay-
lele has guanine in the –1639 position and the ers are reluctant to reimburse for testing because
other allele has adenine in that position) and it is not currently considered a standard of care. Patients with
50% lower in those with the AA genotype
compared with the wild-type genotype GG.21 Testing typically costs a few hundred dollars, but CYP2C9
This promoter SNP, positioned upstream (ie, it should become less expensive as it becomes variants
before the gene-coding region), greatly influ- more commonplace. The current FDA-ap-
ences VKORC1 expression. proved product label for warfarin does not rec- or VKORC1
A meta-analysis of 10 studies found that ommend routine pharmacogenomic testing for
the VKORC1 polymorphism accounts for determining initial or maintenance doses, but variants
25% of the interindividual variation in war- it does acknowledge that dose requirements are may need
farin dose.18 In one study, the frequency of the influenced by CYP2C9 and VKORC1 and states
AA genotype in a white population was 14%, that genotype information, when available, can lower doses
AG 47%, and GG 39%; in a Chinese popula- assist in selecting the starting dose.24 of warfarin
tion the frequency of AA was 82%, AG 18%,
and GG 0.35%.22 The product label includes a table (TABLE 2)
CYP4F2 and GGCX of expected therapeutic warfarin doses based
also affect warfarin’s dose requirements on CYP2C9 and VKORC1 genotypes, which
Genetic variations in the enzymes CYP4F2 can be used when choosing the initial dose
and gamma-glutamyl carboxylase (GGCX) for patients whose genetic status is known.
also influence warfarin dose requirements. Al- A well-developed warfarin-dosing model in-
though the data are limited and the effects are corporating traditional clinical factors and
smaller than those of CYP2C9 and VKORC1, patient genetic status is available on the non-
people with a SNP in CYP4F2 need 8% high- profit Web site www.warfarindosing.org.25
er doses of warfarin, while those with a SNP in
GGCX need 6% lower doses.23 Clinical trials of warfarin pharmacogenomic
testing are under way
Although genetic status can greatly influence
an individual patient’s warfarin dosing re-

  247CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

PHARMACOGENOMIC TESTING

FDA: quirement, routine prospective pharmacoge- conducted in a cohort of 429 healthy Amish
Poor CYP2C19 nomic testing is not endorsed by the FDA or persons revealed a SNP in CYP2C19 to be as-
metabolizers by other expert panels26 because there is cur- sociated with a diminished response to clopi-
may not rently insufficient evidence to recommend for dogrel and to account for 12% of the variation
benefit or against it. in drug response.33 Traditional factors (the pa-
from Several large prospective trials are under tient’s age, body-mass index, and cholesterol
clopidogrel way. For example, the National Heart, Lung, level) combined accounted for less than 10%
and Blood Institute began a prospective trial of the variation.
in about 1,200 patients to evaluate the use Findings were similar in a subsequent in-
of clinical plus genetic information to guide vestigation in 227 cardiac patients receiving
the initiation of warfarin therapy and to im- clopidogrel: 21% of those with the variant had
prove anticoagulation control for patients.27 a cardiovascular ischemic event or died dur-
The results, expected in September 2011, and ing a 1-year follow-up period compared with
those of other large prospective trials should 10% of those without the variant (hazard ratio
provide adequate evidence for making recom- 2.42, P = .02).33
mendations about the clinical utility of rou- A 12-year prospective study investigat-
tine pharmacogenetic testing for guiding war- ing clopidogrel efficacy in 300 cardiac patients
farin therapy. under the age of 45 used cardiovascular death,
Several recent cost-utility and cost-effec- nonfatal myocardial infarction, and urgent coro-
tiveness studies have attempted to quantify the nary revascularization as end points. It conclud-
value of pharmacogenomic testing for warfarin ed that the only independent predictor of these
therapy,28–30 but their analyses are largely lim- events was the patient’s CYP2C19 status.34
ited because the benefit (clinical utility) is yet A study in 2,200 patients with recent myo-
to be sufficiently characterized. cardial infarction examined whether any of
The relevance of such analyses may soon the known allelic variations that modulate
be drastically diminished, as several non-vita- clopidogrel’s absorption (ABCB1), metabolic
min-K-dependent blood thinners such as riva- activation (CYP3A4/5 and CYP2C19), or bio-
roxaban (Xarelto), dabigatran (Pradaxa), and logic activity (P2RY12 and ITGB3) was asso-
apixaban are poised to enter clinical practice.31 ciated with a higher rate of the combined end
■■ CLOPIDOGREL IS ACTIVATED BY CYP2C19 point of all-cause mortality, nonfatal myocar-
Clopidogrel, taken by about 40 million pa- dial infarction, or stroke. None of the SNPs
tients worldwide, is used to prevent athero- in CYP3A4/5, P2RY12, or ITGB3 that were
thrombotic events and cardiac stent thrombo- evaluated was associated with a higher risk at
sis when given along with aspirin. 1 year. However, the allelic variations modu-
Clopidogrel is a prodrug, and to do its job lating clopidogrel’s absorption (ABCB1) and
it must be transformed to a more active me- metabolism (CYP2C19) were associated with
tabolite (FIGURE 2). CYP2C19 is responsible higher event rates. Patients with two variant
for its metabolic activation, and CYP2C19 ABCB1 alleles had a higher adjusted hazard
loss-of-function alleles appear to be associated ratio (95% confidence interval [CI] 1.2–2.47)
with higher rates of recurrent cardiovascular than those with the wild-type allele. Patients
events in patients receiving clopidogrel. At who had one or two CYP2C19 loss-of-func-
least one loss-of-function allele is carried by tion alleles had a higher event rate than those
24% of the white non-Hispanic population, with two wild-type alleles (95% CI 1.10–3.58
18% of Mexicans, 33% of African Americans, and 1.71–7.51, respectively).35
and 50% of Asians. Homozygous carriers, who Conversely, researchers from the Popula-
are poor CYP2C19 metabolizers, make up 3% tion Health Research Institute found no as-
to 4% of the population.32 sociation between poor-metabolizer status
Studies of clopidogrel pharmacogenomics and treatment outcomes when CYP2C19
A recent genome-wide association study analysis was retrospectively added to the find-
ings of two large clinical trials (combined N
> 5,000). However, patients with acute coro-
nary syndrome benefited more from clopido-

248  CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

KITZMILLER AND COLLEAGUES

MM Why some drugs don’t work in some patients

Pharmacogenomic research investigates how patients’ genetics influence their response to medication. Genetic vari-
ability can influence how quickly a drug is metabolized, how well a drug works, or even the likelihood of side effects.
The example below depicts how genetic variation can influence the conversion of a prodrug.

The normal “wild-type” allele (CYP2C19*1) results in a normal-
functioning CYP2C19 enzyme capable of converting the inactive prodrug
clopidogrel (Plavix) to its active metabolite.

Clopidogrel
is a prodrug

Active metabolite

Functional enzyme CYP2C19

Variant alleles (mainly CYP2C19*2, CYP2C19*3, and CYP2C19*17) result in
poor-metabolizer status (having little or no CYP2C19 function). For CYP2C19*2,
the single-nucleotide polymorphism (SNP) is substitution of adenine for guanine
at position 681 of CYP2C19 (681G > A). The resulting CYP2C19 protein is
unable to convert clopidogrel to its active form, and drug efficacy is reduced.

Clopidogrel
remains inactive

SNP CYP2C19*2 protein
FIGURE 2 is enzymatically inactive

 CCF

Medical Illustrator: Beth Halasz ©2011

  249CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

PHARMACOGENOMIC TESTING

TABLE 3 pert Consensus Documents and the American
Drugs that inhibit CYP2C19 Heart Association collectively pronounced
(and therefore can reduce the the current evidence base insufficient for rec-
effect of clopidogrel) ommending routine pharmacogenomic test-
ing.39
Chloramphenicol Needed are large-scale studies examin-
Cimetidine (Tagamet) ing the cost-effectiveness and clinical util-
Felbamate (Felbatol) ity of genotype-guided clopidogrel therapy
Fluoxetine (Prozac) compared with other therapy options such as
Fluvoxamine (Luvox) prasugrel (Effient), an analogue not metabo-
Indomethacin (Indocin) lized by CYP2C19. One such study, sponsored
Ketoconazole (Nizoral) by Medco Health Solutions, plans to enroll
Lansoprazole (Prevacid) 14,600 cardiac patients and has an estimated
Modafinil (Provigil) completion date in June 2011.40 The expec-
Omeprazole (Priloxec) tation that clopidogrel will be available in
Oxcarbazepine (Trileptal) generic form in 2012 adds to the uncertainty
Pantoprazole (Protonix) regarding the cost-effectiveness of CYP2C19
Probenecid (Benemid) testing for clopidogrel therapy.
Rabeprazole (AcipHex) ■■ STATINS:
Ticlopidine (Ticlid)
Topiramate (Topamax) SLC01B1*5 INCREASES MYOPATHY RISK
Statins lower the concentration of low-den-
BASED ON INFORMATION IN FLOCKHART DA. DRUG INTERACTIONS: sity lipoprotein cholesterol (LDL-C), result-
CYTOCHROME P450 DRUG INTERACTION TABLE. ing in a relative-risk reduction of about 20%
for each 1 mmol/L (39 mg/dL) decrement in
INDIANA UNIVERSITY SCHOOL OF MEDICINE (2007). LDL-C.41 They are one of the most commonly
http://medicine.iupui.edu/clinpharm/ddis/table.asp. prescribed classes of drugs, but their side ef-
fects can limit their appeal: statin-induced
grel treatment if they were ultra-rapid metab- myopathy occurs in about 1:1,000 to 1:10,000
olizers (possessing the gain-of-function allele patients and is difficult to predict.
CYP2C19*17).36 SLC01B1. The Study of the Effectiveness
of Additional Reductions in Cholesterol and
Current status of clopidogrel testing: Homocysteine (SEARCH), a genome-wide
Uncertain association study, recently found a SNP (SL-
A current FDA boxed warning states that poor CO1B1*5) in the SLC01B1 gene to be associ-
CYP2C19 metabolizers may not benefit from ated with a higher risk of statin-induced myop-
clopidogrel and recommends that prescribers athy in cardiac patients receiving simvastatin
consider alternative treatment for patients in (Zocor) 40 or 80 mg daily.42 The SLC01B1
this category.37 However, routine CYP2C19 gene, located on chromosome 12, influences
testing is not recommended, and no firm rec- the extent of the drug’s hepatic uptake and its
ommendations have been established regard- serum concentration. Only the SLC01B1*5
ing dose adjustments for CYP2C19 status. SNP emerged as a predictor of statin-induced
Clinicians should be aware that the low myopathy across the entire genome.42
exposure seen in poor metabolizers also occurs The authors believe the findings are likely
in patients taking drugs that inhibit CYP2C19 to apply to other statins. The mechanisms
(TABLE 3).38 leading to statin-induced myopathy and the
In 2010, the American College of Cardi- impact of statin pharmacogenomics are still
ology Foundation Task Force on Clinical Ex- unclear.43
CYP3A4. Other genetic variants may play
a vital role in determining response to statin
therapy. Carriers of a newly identified CYP3A4

250  CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

KITZMILLER AND COLLEAGUES

polymorphism (intron 6 SNP, rs35599367, paired ability to metabolize the CYP2D6 sub-
C>T) required significantly lower statin doses strates sparteine and debrisoquine. Later work
(0.2–0.6 times less) for optimal lipid control. expanded this system to include categories for
The analyses included atorvastatin (Lipitor), intermediate (between poor and extensive)
simvastatin, and lovastatin (Mevacor), and and ultra-rapid (better than extensive) me-
the association was robust (P = .019).44 tabolizers.
Statin pharmacogenomic testing The genetic basis for these categories in-
is not routinely recommended cludes homozygosity for dysfunctional vari-
Routine pharmacogenomic testing for statin ants (the poor-metabolizer group) or extra
therapy is not recommended. Additional stud- copies of normal functioning variants (the
ies are needed to determine the clinical utility ultra-rapid-metabolizer group).
and cost-effectiveness of pharmacogenomic Newer systems have been described for
testing (involving a combination of several characterizing the CYP2D6 activity pheno-
polymorphisms) in various patient popula- type whereby CYP2D6 variants are assigned
tions delineated by type of statin, dose, and activity scores.53–56 The various scoring sys-
concomitant use of other drugs. tems have been reviewed by Kirchheiner.57
■■ TAMOXIFEN IS ACTIVATED BY CYP2D6 A recent version of the activity scoring
Tamoxifen is prescribed to prevent the recur- system also takes into consideration the many
rence of estrogen-receptor-positive breast can- drugs that inhibit CYP2D6, such as amioda-
cer, to treat metastatic breast cancer, to pre- rone (Cordarone) and fluoxetine (Prozac) that
vent cancer in high-risk populations, and to can reduce the action of tamoxifen if given
treat ductal carcinoma in situ. with it (TABLE 4).58 For example, the tamoxifen
Tamoxifen is metabolized to form en- exposure (as predicted by the CYP2D6-activi-
doxifen, which has much higher potency and ty score) experienced by a CYP2D6 extensive
higher systemic levels than tamoxifen.45 Both metabolizer taking a CYP2D6-inhibiting drug
CYP2D6 and CYP3A4/5 are required to pro- may be similar to the exposure experienced
duce endoxifen via two intermediates, but by a CYP2D6 poor metabolizer receiving the
CYP2D6 catalyzes the critical step leading to same tamoxifen dose but not taking a CY- The pharmaco-
metabolic activation.
The CYP2D6 gene is highly polymorphic, P2D6-inhibiting drug. genomics
with more than 75 allelic variants identi- Likewise, the activity score of a CYP2D6 of codeine
fied. Extensive literature is available describ- intermediate metabolizer taking a CYP2D6-
ing the influence of CYP2D6 polymorphisms inducing drug may be similar to that of a has become
on tamoxifen metabolism and therapy out- CYP2D6 ultra-rapid metabolizer not taking
comes.46–52 Several CYP2D6 variants result a CYP2D6-inducing drug. Examples of CY- a hot topic,
in reduced or no enzyme activity, and people P2D6 inducers are dexamethasone, rifampin, especially in
who have more than two normally function- and hyperforin (St. John’s wort).
ing alleles have exaggerated enzyme activity While the newer systems are reported to breastfeeding
(gene amplification). provide better correlations between genotype mothers
Classification of CYP2D6 status
Several systems have been developed to cat- and phenotype scores, the older scoring sys-
egorize the phenotypic activity of CYP2D6 tems and the categorical names are still widely
based on genotype. used (eg, in the FDA-approved AmpliChip
A genetic basis for the observed diversity CYP450 test from Roche,59 which includes
in the metabolism of cytochrome P450 sub- genotype data for CYP2D6 and CYP2C19).
strates was recognized more than 30 years ago.
People were categorized as either extensive No firm recommendations
or poor metabolizers, reflecting normal vs im- for CYP2D6 testing in tamoxifen users
The different genotypes and phenotypes
vary in prevalence in different ethnic
groups, and significantly different activ-
ity levels for endoxifen formation are ob-
served. Tamoxifen lacks efficacy in those
who are poor CYP2D6 metabolizers—ie,

  251CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

PHARMACOGENOMIC TESTING

TABLE 4 about 7% of the white population.
However, the FDA has not made firm
Drugs that inhibit CYP2D6 recommendations about CYP2D6 testing for
(and therefore can reduce the effects prescribing tamoxifen because the evidence of
of tamoxifen and codeine) benefit, although suggestive, has been consid-
ered insufficient.
Amiodarone (Cordarone, Pacerone) Clinicians should be aware that tamoxi-
Bupropion (Wellbutrin) fen’s efficacy is greatly reduced by concomi-
Celecoxib (Celebrex) tant therapy with CYP2D6-inhibiting drugs
Chlorpheniramine (TABLE 4).
Chlorpromazine (Thorazine) Other genes affecting tamoxifen:
Cimetidine (Tagamet) CYP3A4/5, SULT1A1, and UGT2B15
Cinacalcet (Sensipar) Some investigators propose that polymor-
Citalopram (Celexa) phisms in additional genes encoding enzymes
Clemastine (Tavist) in the tamoxifen metabolic and elimination
Clomipramine (Anafranil) pathways (eg, CYP3A4/5, SULT1A1, and
Cocaine UGT2B15) also need to be considered to ac-
Diphenhydramine (Benadryl) count adequately for interindividual variation
Doxepin (Sinequan) in drug response.
Doxorubicin (Adriamycin) For example, CYP3A4 and CYP3A5 are
Duloxetine (Cymbalta) also polymorphic, and large interindividual
Escitalopram (Lexapro) variation exists in their enzyme activities.
Fluoxetine (Prozac) These enzymes have overlapping substrate
Halofantrine (Halfan) specificities, represent the most abundant
Haloperidol (Haldol) drug-metabolizing enzymes in the human liv-
Hydroxyzine (Vistaril) er, and are involved in the biotransformation
Levomepromazine of a broad range of endogenous substrates and
Methadone most drugs.60
Metoclopramide (Reglan) Clinical studies evaluating the impact of
Mibefradil (Posicor) CYP3A4/5 polymorphisms have been incon-
Midodrine (ProAmatine) sistent in their conclusions, which is gener-
Moclobemide ally attributed to the relatively low functional
Paroxetine (Paxil) impact or the low prevalence of the SNPs
Perphenazine (Trilafon) evaluated. Many of the nearly 100 CYP3A4/5
Quinidine polymorphisms identified have not yet been
Ranitidine (Zantac) characterized regarding their functional im-
Ritonavir (Norvir) pact on enzyme expression or activity. CYP-
Sertraline (Zoloft) 3A4/5 enzyme activity is highly variable be-
Terbinafine (Lamisil) tween individuals and warrants further study
Ticlopidine (Ticlid) of its role in outcomes of tamoxifen therapy.
Tripelennamine (Pyribenzamine) Ongoing and future prospective clinical trials
evaluating CYP2D6, CYP3A4/5, and other
relevant polymorphisms are necessary to de-
fine their clinical relevance before routine ge-
netic testing for tamoxifen can be justified.
■■ CODEINE IS ALSO ACTIVATED BY CYP2D6

BASED ON INFORMATION IN FLOCKHART DA. DRUG INTERACTIONS: Codeine also depends on the CYP2D6 gene,
CYTOCHROME P450 DRUG INTERACTION TABLE. INDIANA UNIVERSITY SCHOOL OF MEDICINE as it must be activated to its more potent
opioid metabolites, including morphine.
(2007). http://medicine.iupui.edu/clinpharm/ddis/table.asp.

252  CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

KITZMILLER AND COLLEAGUES

Poor CYP2D6 metabolizers do not benefit by any expert panel because sufficient clini-
from codeine therapy. cal utility and cost-effectiveness have not
The pharmacogenomics of codeine has be- been demonstrated. A brief summary of
come a hot topic, especially regarding breast- study findings and a few practical sugges-
feeding mothers. The debate was ignited with tions follow.
the publication in 2006 of a case report of an Polymorphisms in metabolizing enzymes
infant’s death, apparently the result of meta- have been investigated in patients receiving
bolic polymorphisms.61 The evolution of this psychotropic drugs.
debate and the outcome of the case may be
noteworthy to clinicians, as they illustrate the CYP2D6 and antidepressants
gravity of public and patient interest in phar- Many antidepressants show significant differ-
macogenomic testing. In this case, the breast- ences in plasma drug levels with CYP2D6 poly-
feeding mother had taken codeine regularly morphisms (in descending order of influence)55:
for about 14 days when her 13-day-old infant • Imipramine (Tofranil)
died from toxic levels of morphine. Unknown • Doxepin (Adapin, Silenor, Sinequan)
to her and the prescriber, both the mother and • Maprotiline (Deprilept, Ludiomil,
infant were ultra-rapid CYP2D6 metabolizers, Psymion)
resulting in a more rapid and extensive con- • Trimipramine (Surmontil)
version of codeine to morphine. • Desipramine (Noraprim)
A logical strategy for preventing similar • Nortriptyline (Aventyl, Pamelor)
deaths would be routine CYP2D6 genotyp- • Clomipramine (Anafranil)
ing when prescribing codeine to breast- • Paroxetine (Paxil)
feeding mothers. However, after several • Venlafaxine (Effexor)
investigations examined the metabolic and • Amitriptyline (Elavil)
excretion pathways of codeine in their en- • Mianserin
tirety, the FDA did not recommend routine • Trazadone (Desyrel)
CYP2D6 testing when prescribing codeine to • Bupropion (Wellbutrin)
breastfeeding mothers because several other • Nefazodone (Serzone) Consider
factors, including rare genetic variations of
other enzymes, proved necessary for reach- • Citalopram (Celexa) non-SSRI
ing the opioid toxicity leading to the infant’s • Sertraline (Zoloft). antidepressants
death.62
■■ PHARMACOGENOMICS CYP2D6 and antipsychotics for patients
Several antipsychotics are also influenced by
OF PSYCHOTROPIC Drugs CYP2D6 polymorphisms (also in descending with the
Pharmacogenomic testing has clinical utility order of influence)55: SLC6A4 variant
for some psychotropic drugs. • Perphenazine (Trilafon)
HLA-B and carbamazepine • Thioridazine (Mellaril)
Considered a standard of care, HLA-B geno- • Olanzapine (Zyprexa)
typing is appropriate before prescribing carba- • Zuclopenthixol (Cisordinol, Clopixol,
mazepine (Tegretol, Equetro) to patients in Acuphase)
populations in which HLAB*1502 is likely to • Aripiprazole (Abilify)
be present, such as Asians. Carriers of HLA- • Flupentixol (Depixol, Fluanxol)
B*1502 are at higher risk of life-threatening • Haloperidol (Haldol)
skin reactions such as Stevens-Johnson syn- • Perazine (Taxilan)
drome.11 • Risperidone (Risperdal)
Several other pharmacogenomic appli- • Pimozide (Orap).
cations for psychotropic medications have
been suggested, but routine testing has not CYP2C19 and antidepressants
been recommended by the FDA or endorsed CYP2C19 polymorphisms are likewise associ-
ated with differences in drug metabolism for
many antidepressants, such as (in descending
order of CYP2C19-mediated influence)55:

  253CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

PHARMACOGENOMIC TESTING

TABLE 5
Suggested dose adjustments based on CYP2D6 phenotype

Percent of standard dose

DRUG POOR INTERMEDIATE EXTENSIVE ULTRA-RAPID
METABOLIZERS METABOLIZERS METABOLIZERS METABOLIZERS

Imipramine (Tofranil) 30 80 130 180

Doxepin (Sinequan) 35 80 130 175

Trimipramine (Surmontil) 35 90 130 180

Despramine (Norpramin) 40 80 125 170

Nortriptyline (Pamelor) 55 95 120 150

Clomipramine (Anafranil) 60 90 110 145

Paroxetine (Paxil) 65 85 115 135

Venlafaxine (Effexor) 70 80 105 130

Amitriptyline (Elavil) 75 90 105 130

Bupropion (Wellbutrin) 90 95 105 110

Perphenazine (Trilafon) 30 80 125 175

Haloperidol (Haldol) 75 95 100 115

Knowing the Olanzapine (Zyprexa) 60 105 120 150
frequency
of pharmaco- Risperidone (Risperdal) 85 90 100 110
genomic
variants Dose-adjustment values were determined by comparing drug concentration, clearance, or exposure data across phenotypes. Values
in a given have been approximated to the nearest 5%.
population
can be helpful ADAPTED FROM INFORMATION IN KIRCHHEINER J, NICKCHEN K, BAUER M, ET AL. PHARMACOGENETICS OF ANTIDEPRESSANTS AND ANTIPSYCHOTICS:
in prescribing THE CONTRIBUTION OF ALLELIC VARIATIONS TO THE PHENOTYPE OF DRUG RESPONSE. MOL PSYCHiatry 2004; 9:442–473.

• Trimipramine significant associations.69–72 Likewise, several
• Doxepin studies have found an association between
• Amitriptyline ultra-rapid CYP2D6 metabolizer status and
• Imipramine diminished response to antidepressants,65,73,74
• Citalopram (Celexa) but no association was found in a larger ret-
• Clomipramine rospective study.75
• Moclobemide (Aurorix, Manerix) Routine CYP2D6 and CYP2C19 screening
• Sertraline is not recommended when prescribing psycho-
• Fluvoxamine (Luvox). tropic drugs. However, reviews of the pharma-
Clinical relevance of CYP2D6 and CYP2C19 cokinetic data have suggested a few practical
Several studies have demonstrated that applications when genetic status is already
poor and intermediate CYP2D6 metaboliz- known. In general, clinicians can consider re-
ers have a higher incidence of adverse ef- ducing the dose of tricyclic antidepressants by
fects when taking CYP2D6-dependent an- about 50% when prescribing to CYP2D6-poor-
tidepressants63–68; however, an almost equal metabolizers.55,76–78
number of studies did not find statistically TABLE 5 gives examples of specific dose ad-
justments of antidepressants and antipsychot-

254  CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

KITZMILLER AND COLLEAGUES

ics based on CYP2D6-mediated influence. ■■ THE FUTURE
Kirchheiner’s review article55 includes several OF PHARMACOGENOMIC TESTING
similar tables and charts based on CYP2D6
status as well as several based on CYP2C9 The examples discussed in this article provide
status. Clinicians should consider using these some insight about how pharmacogenomic test-
types of pharmacokinetic-derived charts and ing is maturing and slowly being integrated into
tables when prescribing to patients whose ge- the practice of medicine. They also illustrate the
netic status is known. complexity of the multiple stages of research that
Genes that affect serotonin metabolism pharmacogenomic applications must go through
Several genes in the serotonin pathway have in order to be adopted as standard practice.
been investigated to determine whether they In the future, pharmacogenomic data will
influence patients’ susceptibility to depres- continue to accumulate, and the clinical util-
sion and adverse effects and response to psy- ity of many other pharmacogenomic tests
chotropic medications. may be uncovered. The FDA provides infor-
SLC6A4. Polymorphisms in the promot- mation on emerging pharmacogenomic tests
er region of the serotonin transporter gene at its Web site, www.fda.gov.11 Its up-to-date
SLC6A4 appear to influence the treatment “Table of Valid Genomic Biomarkers in the
response and side-effect profiles of selective Context of Approved Drug Labels” includes
serotonin reuptake inhibitors (SSRIs). Car- boxed warnings, recommendations, research
riers of the SLC6A4 5-HTTLPR L alleles outcomes, and relevant population genetics.
have fewer side effects79 and better response If the FDA continues its current policy, pro-
to SSRI treatment, and carriers of the S allele spective randomized trials that show improve-
have a higher incidence of antidepressant- ment in patient outcomes will remain the gold
induced mania80 and poorer response to SSRI standard for determining the clinical signifi-
treatment.81 cance of a pharmacogenomic test. Furthermore,
5-HT. Polymorphisms in serotonin recep- cost-benefit analyses are likely to continue dic-
tors (2A and 2C subtypes) appear to influ- tating policy regarding pharmacogenomic test-
ence SSRI response and side effects. Carriers ing, and cost-benefit profiles should improve as
of 5-HT 2A C alleles had more severe adverse technology advances and as information gath-
effects from paroxetine,71 but another 5-HT ered from a single test becomes applicable to
2A polymorphism common to Asians is asso- multiple medications and clinical scenarios.
ciated with better response to antidepressant In the meantime, physicians should be-
therapy.82 A 5-HT 2C polymorphism was as- come familiar with the terms used in medical
sociated with a lower incidence of antipsy- genetics and pharmacogenomics and begin
chotic-induced weight gain.83 to understand genetic contributions to the
Although the understanding of these re- outcomes of drug therapy. For example, un-
lationships is incomplete and routine phar- derstanding the consequences of metabolizer
macogenomic testing is not currently recom- status and the frequency of variants in a given
mended, reviews of the pharmacodynamic population can be tremendously helpful when
data have suggested a few practical applica- advising our patients about anticipating po-
tions when a patient’s genetic status is already tential problems when taking specific medi-
known. One should consider: cations and about making informed decisions
• Selecting treatments other than SSRIs for about pharmacogenomic testing.
This exchange of information alone may go
depressed patients known to possess the a long way in improving therapy outcomes even
SLC6A4 variant when prospective pharmacogenomic testing is
• Selecting citalopram for depressed pa- not routinely performed. Furthermore, an in-
tients known to carry the 5-HT 2A poly- creasing number of patients will already have ge-
morphism notyping information available when they come
• Avoiding treatment with antipsychotic to us, and clinicians need to be aware of the many
drugs for patients known to possess the pharmacogenomic applications recommended by
5-HT 2C polymorphism. the FDA when genetic status is known.10 ■

  255CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

PHARMACOGENOMIC TESTING

■■ REFERENCES alters required warfarin dose. Blood 2008; 111:4106–4112.
24. Bristol-Myers Squibb. Coumadin (warfarin sodium) Prescribing
1. King HC, Sinha AA. Gene expression profile analysis by DNA micro-
arrays: promise and pitfalls. JAMA 2001; 286:2280–2288. Information. January 2010.
25. Barnes-Jewish Hospital at Washington University Medical Center.
2. Nuckols TK, Paddock SM, Bower AG, et al. Costs of intravenous
adverse drug events in academic and nonacademic intensive care Warfarin dosing. http://warfarindosing.org Accessed 1/20/2011.
units. Med Care 2008; 46:17–24. 26. Flockhart DA, O’Kane D, Williams MS, et al. Pharmacogenetic test-

3. Gurwitz JH, Field TS, Judge J, et al. The incidence of adverse ing of CYP2C9 and VKORC1 alleles for warfarin. Genet Med 2008;
events in two long-term care facilities. Am J Med 2005; 118:251– 10:139–150.
258. 27. National Institutes of Health. ClinicalTrials.gov. http://clinicaltrials.
gov. Accessed January 20, 2011.
4. Vargas E, Terleira A, Hernando F, et al. Effects of adverse drug re- 28. Eckman MH, Rosand J, Greenberg SM, Gage BF. Cost-effectiveness
actions on length of stay in surgical intensive care units. Crit Care of using pharmacogenetic information in warfarin dosing for pa-
Med 2003; 31:694–698. tients with nonvalvular atrial fibrillation. Ann Intern Med 2009;
150:73–83.
5. Lazarou J, Pomeranz BH, Corey PN. Incidence of adverse drug 29. Meckley LM, Gudgeon JM, Anderson JL, Williams MS, Veenstra
reactions in hospitalized patients: a meta-analysis of prospective DL. A policy model to evaluate the benefits, risks, and costs of
studies. JAMA 1998; 279:1200–1205. warfarin pharmacogenomic testing. Pharmacoeconomics 2010;
28:61–74.
6. Field TS, Gilman BH, Subramanian S, Fuller JC, Bates DW, Gurwitz 30. Patrick AR, Avron J, Choudhry NK. Cost-effectiveness of genotype-
JH. The costs associated with adverse drug events in older adults guided warfarin dosing for patients with atrial fibrillation. Circ
in the ambulatory setting. Med Care 2005; 43:1171–1176. Cardiovasc Qual Outcomes 2009; 2:429–436.
31. Haas S. New oral Xa and IIa inhibitors: updates on clinical trial
7. Classen DC, Pestotnik SL, Evans RS, Lloyd JF, Burke JP. Adverse results. J Thromb Thrombolysis 2008; 25:52–60.
drug events in hospitalized patients. Excess length of stay, extra 32. Bhatt DL. Tailoring antiplatelet therapy based on pharmacoge-
costs, and attributable mortality. JAMA 1997; 277:301–306. nomics: how well do the data fit? JAMA 2009; 302:896–897.
33. Shuldiner AR, O’Connell JR, Bliden KP, et al. Association of cyto-
8. Ernst FR, Grizzle AJ. Drug-related morbidity and mortality: updat- chrome P450 2C19 genotype with the antiplatelet effect and
ing the cost-of-illness model. J Am Pharm Assoc (Wash) 2001; clinical efficacy of clopidogrel therapy. JAMA 2009; 302:849–857.
41:192–199. 34. Collet JP, Hulot JS, Pena A, et al. Cytochrome P450 2C19 polymor-
phism in young patients treated with clopidogrel after myocar-
9. Bradford LD. CYP2D6 allele frequency in European Caucasians, dial infarction: a cohort study. Lancet 2009; 373:309–317.
Asians, Africans and their descendants. Pharmacogenomics 2002; 35. Simon T, Verstuyft C, Mary-Krause M, et al; French Registry of
3:229–243. Acute ST-Elevation and Non-ST-Elevation Myocardial Infarc-
tion (FAST-MI) Investigators. Genetic determinants of response
10. Sadee W. Measuring cis-acting regulatory variants genome-wide: to clopidogrel and cardiovascular events. N Engl J Med 2009;
new insights into expression genetics and disease susceptibility. 360:363–375.
Genome Med 2009; 1:116. 36. Paré G, Mehta SR, Yusuf S, et al. Effects of CYP2C19 genotype
on outcomes of clopidogrel treatment. N Engl J Med 2010;
11. US Food and Drug Administration. Table of pharmacogenomic 363:1704–1714.
biomarkers in drug labels. http://www.fda.gov/Drugs/Scien- 37. Bristol-Myers Squibb/Sanofi Pharmaceutical Partnership. Plavix
ceResearch/ResearchAreas/Pharmacogenetics/ucm083378.htm. (clopidogrel bisulfate) prescribing information. August 2010.
Accessed 1/18/2011. 38. P450 Drug Interaction Table. Indiana University School of Medi-
cine. http://medicine.iupui.edu/clinpharm/ddis/table.asp. Ac-
12. Budnitz DS, Shehab N, Kegler SR, Richards CL. Medication use cessed 1/21/2011.
leading to emergency department visits for adverse drug events 39. Society for Cardiovascular Angiography and Interventions; Society
in older adults. Ann Intern Med 2007; 147:755–765. of Thoracic Surgeons; Writing Committee Members, Holmes DR
Jr, Dehmer GJ, Kaul S, Leifer D, O’Gara PT, Stein CM. ACCF/AHA
13. Elias DJ, Topol EJ. Warfarin pharmacogenomics: a big step forward clopidogrel clinical alert: approaches to the FDA “boxed warn-
for individualized medicine: enlightened dosing of warfarin. Eur ing”: a report of the American College of Cardiology Founda-
J Hum Genet 2008; 16:532–534. tion Task Force on Clinical Expert Consensus Documents and the
American Heart Association. Circulation 2010; 122:537–557.
14. Epstein RS, Moyer TP, Aubert RE, et al. Warfarin genotyping 40. National Institutes of Health. Genotype Guided Comparison of
reduces hospitalization rates results from the MM-WES (Medco- Clopidogrel and Prasugrel Outcomes Study. http://clinicaltrials-
Mayo Warfarin Effectiveness study). J Am Coll Cardiol 2010; feeds.org/clinical-trials/show/NCT00995514. Accessed 1/20/2011.
55:2804–2812. 41. Amarenco P, Labreuche J. Lipid management in the prevention of
stroke: Review and updated meta-analysis of statins for stroke
15. Rettie AE, Wienkers LC, Gonzalez FJ, Trager WF, Korzekwa KR. prevention. Lancet Neurol 2009; 8:453–463.
Impaired (S)-warfarin metabolism catalysed by the R144C allelic 42. SEARCH Collaborative Group, Link E, Parish S, Armitage J, et al.
variant of CYP2C9. Pharmacogenetics 1994; 4:39–42. SLCO1B1 variants and statin-induced myopathy—a genomewide
study. N Engl J Med 2008; 359:789–799.
16. Higashi MK, Veenstra DL, Kondo LM, et al. Association between 43. Romaine SP, Bailey KM, Hall AS, Balmforth AJ. The influence of
CYP2C9 genetic variants and anticoagulation-related outcomes SLC01B1 (OATP1B1) gene polymorphisms on response to statin
during warfarin therapy. JAMA 2002; 287:1690–1698. therapy. Pharmacogenomics J 2010; 10:1–11.
44. Wang D, Guo Y, Wrighton SA, Cooke GE, Sadee W. Intronic poly-
17. Sanderson S, Emery J, Higgins J. CYP2C9 gene variants, drug morphism in CYP3A4 affects hepatic expression and response to
dose, and bleeding risk in warfarin-treated patients: a HuGEnet statin drugs. Pharmacogenomics J 2010; Apr 13 [Epub ahead of
systematic review and meta-analysis. Genet Med 2005; 7:97–104. print].
45. Goetz MP, Rae JM, Suman VJ, et al. Pharmacogenetics of tamoxifen
18. Au N, Rettie AE. Pharmacogenomics of 4-hydroxycoumarin antico- biotransformation is associated with clinical outcomes of efficacy
agulants. Drug Metab Rev 2008; 40:355–375.

19. García-Martín E, Martínez C, Ladero JM, Agúndez JA. Interethnic
and intraethnic variability of CYP2C8 and CYP2C9 polymorphisms
in healthy individuals. Mol Diagn Ther 2006; 10:29–40.

20. Sullivan-Klose TH, Ghanayem BI, Bell DA, et al. The role of the
CYP2C9-Leu359 allelic variant in the tolbutamide polymorphism.
Pharmacogenetics 1996; 6:341–349.

21. Wen MS, Lee M, Chen JJ, et al. Prospective study of warfarin dos-
age requirements based on CYP2C9 and VKORC1 genotypes. Clin
Pharmacol Ther 2008; 84:83–89.

22. Larramendy-Gozalo C, Yang JQ, Verstuyft C, et al. Genetic poly-
morphism of vitamin K epoxide reductase (VKORC1) 1173C>T in a
Chinese and a Caucasian population. Basic Clin Pharmacol Toxicol
2006; 98:611–613.

23. Caldwell MD, Awad T, Johnson JA, et al. CYP4F2 genetic variant

256  CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011

KITZMILLER AND COLLEAGUES

and hot flashes. J Clin Oncol 2005; 23:9312–9318. pressants—a pilot study. Clin Pharm Ther 2004; 75:386–393.
46. Beverage JN, Sissung TM, Sion AM, Danesi R, Figg WD. CYP2D6 66. McAlpine DE, O’Kane DJ, Black JL, Mrazek DA. Cytochrome P450

polymorphisms and the impact on tamoxifen therapy. J Pharm Sci 2D6 genotype variation and venlafaxine dosage. Mayo Clin Proc
2007; 96:2224–2231. 2007; 82:1065–1068.
47. Flockhart D. CYP2D6 genotyping and the pharmacogenetics of 67. Chen S, Chou WH, Blouin RA, et al. The cytochrome P450 2D6
tamoxifen. Clin Adv Hematol Oncol 2008; 6:493–494. (CYP2D6) enzyme polymorphism: screening costs and influence
48. Goetz MP, Kamal A, Ames MM. Tamoxifen pharmacogenomics: the on clinical outcomes in psychiatry. Clin Pharmacol Ther 1996;
role of CYP2D6 as a predictor of drug response. Clin Pharmacol 60:522–534.
Ther 2008; 83:160–166. 68. Shams ME, Arneth B, Hiemke C, et al. CYP2D6 polymorphism and
49. Stearns V, Rae JM. Pharmacogenetics and breast cancer endocrine clinical effect of the antidepressant venlafaxine. J Clin Pharm
therapy: CYP2D6 as a predictive factor for tamoxifen metabolism Ther 2006; 31:493–502.
and drug response? Expert Rev Mol Med 2008; 10:e34. 69. Whyte EM, Romkes M, Mulsant BH, et.al. CYP2D6 genotype and
50. Dezentjé VO, Guchelaar HJ, Nortier JW, van del Velde CJ, Gelderb- venlafaxine-XR concentrations in depressed elderly. Int J Geriatr
lom H. Clinical implications of CYP2D6 genotyping in tamoxifen Psychiatry 2006; 21:542–549.
treatment for breast cancer. Clin Cancer Res 2009; 15:15–21. 70. Roberts RL, Mulder RT, Joyce PR, Luty SE, Kennedy MA. No evi-
51. Higgins MJ, Rae JM, Flockhart DA, Hayes DF, Stearns V. Pharma- dence of increased adverse drug reactions in cytochrome P450
cogenetics of tamoxifen: who should undergo CYP2D6 genetic CYP2D6 poor metabolizers treated with fluoxetine or nortripty-
testing? J Natl Compr Canc Netw 2009; 7:203–213. line. Hum Psychopharmacol 2004; 19:17–23.
52. Hoskins JM, Carey LA, McLeod HL. CYP2D6 and tamoxifen: DNA 71. Murphy GM Jr, Kremer C, Rodrigues HE, Schatzberg AF. Phar-
matters in breast cancer. Nat Rev Cancer 2009; 9:576–586. macogenetics of antidepressant medication intolerance. Am J
53. Steimer W, Zöpf K, von Amelunxen S, et al. Allele-specific change Psychiatry 2003; 160:1830–1835.
of concentration and functional gene dose for the prediction of 72. Gillman PK. Re: no evidence of increased adverse drug reactions in
steady-state serum concentrations of amitriptyline and nor- cytochrome P450 CYP2D6 poor metabolizers treated with fluox-
triptyline in CYP2C19 and CYP2D6 extensive and intermediate etine or nortriptyline. Hum Psychopharmacol 2005; 20:61–62.
metabolizers. Clinical Cancer 2004; 50:1623–1633. 73. Gex-Fabry M, Eap CB, Oneda B, et al. CYP2D6 and ABCB1 genetic
54. Gaedigk A, Simon SD, Pearce RE, Bradford LD, Kennedy MJ, Leeder variability: influence on paroxetine plasma level and therapeutic
JS. The CYP2D6 activity score: translating genotype information into a response. Ther Drug Monit 2008; 30:474–482.
qualitative measure of phenotype. Clin Pharm Ther 2008; 83:234–242. 74. Kawanishi C, Lundgren S, Agren H, Bertilsson L. Increased inci-
55. Kirchheiner J, Nickchen K, Bauer M, et al. Pharmacogenetics of dence of CYP2D6 gene duplication in patients with persistent
antidepressants and antipsychotics: the contribution of allelic mood disorders: ultrarapid metabolism of antidepressants as a
variations to the phenotype of drug response. Mol Psychiatry cause of nonresponse. A pilot study. Eur J Clin Pharmacol 2004;
2004; 9:442–473. 59:803–807.
56. Kirchheiner J, Schmidt H, Tzvetkov M, et al. Pharmacokinetics of 75. Serretti A, Calati R, Massat I, et al. Cytochrome P450 CYP1A2,
codeine and its metabolite morphine in ultra-rapid metabolizers CYP2C9, CYP2C19 and CYP2D6 genes are not associated with
due to CYP2D6 duplication. Pharmacogenomics J 2007; 7:257– response and remission in a sample of depressive patients. Int
265. Clin Psychopharmacol 2009; 24:250–256.
57. Kirchheiner J. CYP2D6 phenotype prediction from genotype: which 76. de Leon J, Armstrong SC, Cozza KL. Clinical guidelines for psychia-
system is the best? Clin Pharmacol Ther 2008; 83:225–227. trists for the use of pharmacogenetic testing for CYP450 2D6 and
58. Borges S, Desta Z, Jin Y, et al. Composite functional genetic and CYP450 2C19. Psychosomatics 2006; 47:75–85.
comedication CYP2D6 activity score in predicting tamoxifen drug 77. de Leon J, Susce MT, Johnson M, et al. DNA microarray technol-
exposure among breast cancer patients. J Clin Pharmacol 2010; ogy in the clinical environment: the AmpliChip CYP450 test for
50:450–458. CYP2D6 and CYP2C19 genotyping. CNS Spectr 2009; 14:19–34.
59. Hoffmann-La Roche Ltd. AmpliChip CYP450 Test. http://www. 78. Thuerauf N, Lunkenheimer J. The impact of the CYP2D6-polymor-
roche.com/assays/Pages/AmpliChipCYP450Test.aspx. Accessed phism on dose recommendations for current antidepressants. Eur
1/21/2011. Arch Psychiatry Clin Neurosci 2006; 256:287–293.
60. Anzenbacher P, Anzenbacherová E. Cytochromes P450 and me- 79. Horstmann S, Binder EB. Pharmacogenomics of antidepressant
tabolism of xenobiotics. Cell Mol Life Sci 2001; 58:737–747. drugs. Pharmacol Ther 2009; 124:57–73.
61. Koren G, Cairns J, Chitayat D, Gaedigk A, Leeder SJ. Pharmaco- 80. Ferreira Ade A, Neves FS, da Rocha FF, et al. The role of 5-HTTLPR
genetics of morphine poisoning in a breastfed neonate of a polymorphism in antidepressant-associated mania in bipolar
codeine-prescribed mother. Lancet 2006; 368:704. disorder. J Affect Disord 2009; 112:267–272.
62. Willmann S, Edginton AN, Coboeken K, Ahr G, Lippert J. Risk to 81. Serretti A, Kato M, De Ronchi D, Kinoshita T. Meta-analysis of
the breast-fed neonate from codeine treatment to the mother: a serotonin transporter gene promoter polymorphism (5-HTTLPR)
quantitative mechanistic modeling study. Clin Pharm Ther 2009; association with selective serotonin reuptake inhibitor efficacy in
86:634–643. depressed patients. Mol Psychiatry 2007; 12:247–257.
63. Grzesiak M, Beszłej A, Lebioda A, Jonkisz A, Dobosz T, Kienja A. 82. Kato M, Serretti A. Review and meta-analysis of antidepressant
[Retrospective assessment of the antidepressants tolerance in pharmacogenetic findings in major depressive disorder. Mol
the group of patients with diagnosis of depression and different Psychiatry 2010; 15:473–500.
CYP2D6 genotype.] [In Polish] Psychiatr Pol 2003; 37:433–444. 83. Reynolds GP, Zhang Z, Zhang X. Polymorphism of the promoter
64. Laika B, Leucht S, Heres S, Steimer W. Intermediate metabolizer: region of the serotonin 5-HT(2C) receptor gene and clozapine-
increased side effects in psychoactive drug therapy. The key to induced weight gain. Am J Psychiatry 2003; 160:677–679.
cost-effectiveness of pretreatment CYP2D6 screening? Pharma-
cogenomics J 2009; 9:395–403. ADDRESS: Joseph P. Kitzmiller, MD, PhD, Department of Pharmacology,
65. Rau T, Wohlleben G, Wuttke H, et al. CYP2D6 genotype: Impact on The Ohio State University, 5072C Graves Hall, 333 West 10th Avenue,
adverse effects and nonresponse during treatment with antide- Columbus, OH 43210; e-mail [email protected].

  257CLEVELAND CLINIC JOURNAL OF MEDICINE   VOLUME 78  •  NUMBER 4   APRIL  2011