Download CYP2D6 Pharmacogenomics of Tamoxifen Treatment Executive Summary

Technology Evaluation Center
CYP2D6 Pharmacogenomics
of Tamoxifen Treatment
Assessment
Program
Volume 28, No. 8
January 2014
Executive Summary
Background
Tamoxifen is prescribed as a component of adjuvant endocrine therapy to prevent hormone receptorpositive breast cancer recurrence, as treatment of metastatic breast cancer, and to prevent disease
in high-risk populations and in women with ductal carcinoma in situ. The cytochrome P450 2D6
(CYP2D6) metabolic enzyme has a major role in tamoxifen metabolism. The CYP2D6 gene is polymorphic; variant DNA gene sequences resulting in proteins with markedly reduced or absent enzyme
function may be associated with lower plasma levels of active tamoxifen metabolites, particularly
endoxifen, which is predominantly dependent on CYP2D6 for its production, and are thus hypothesized to have an impact on tamoxifen treatment efficacy. Patients who have little or no CYP2D6
enzyme function are called poor metabolizers (PMs) compared with patients with 2 fully functional
alleles, termed wild-type or extensive metabolizers (EMs). Those with enzyme activity in between are
called intermediate metabolizers.
Objective
This Assessment evaluates the evidence for CYP2D6 genotyping, compared with no testing, to direct
treatment regimen choices for patients at high risk for primary breast cancer or breast cancer recurrence, and to improve survival outcomes.
Search Strategy
MEDLINE® was searched (via PubMed) using the search string (“Breast Neoplasms”[MeSH®] AND
“Tamoxifen”[MeSH®]) AND “Cytochrome P-450 Enzyme System”[MeSH®] through November 2013.
Clinical trials, recent reviews (2008-2013), editorials, and letters related to the pharmacogenomics of
tamoxifen were retrieved. Text and reference lists of retrieved articles were examined for additional
relevant articles.
®
®
BlueCross
BlueShield
Association
An Association
of Independent
Blue Cross and
Blue Shield Plans
Selection Criteria
Full-length, peer-reviewed publications reporting studies of postmenopausal women undergoing
endocrine therapy whose treatment regimen selection is based on CYP2D6 genotyping versus usual
selection methods or studies of the association of CYP2D6 genotype with intermediate (e.g., tamoxifen-active metabolite levels) or final outcomes (e.g., time to recurrence, survival) were selected
for review. Studies were assessed for potential bias, such as survival bias (genotyping whole blood
from surviving participants of retrospective studies); misclassification of metabolizer phenotype; and
adjustment for variables not considered confounders of the genotype-outcome association.
Main Results
One U.S. Food and Drug Administration-cleared test for CYP2D6 genotyping has consistent evidence
of analytic validity (i.e., technical accuracy and reliability).
NOTICE OF PURPOSE: TEC Assessments are scientific opinions, provided solely for informational purposes. TEC Assessments
should not be construed to suggest that the Blue Cross Blue Shield Association, Kaiser Permanente Medical Care Program or the
TEC Program recommends, advocates, requires, encourages, or discourages any particular treatment, procedure, or service; any
particular course of treatment, procedure, or service; or the payment or non-payment of the technology or technologies evaluated.
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
1
Technology Evaluation Center
There is a chain of evidence and direct evidence for clinical validity (i.e., association of CYP2D6
genotype with drug efficacy). The chain of evidence associates an intermediate outcome, tamoxifen metabolite plasma levels, with both CYP2D6 genotype and health outcomes (recurrence,
survival) in women treated with tamoxifen. Direct evidence compares outcomes in women treated
with tamoxifen or other endocrine therapy stratified by metabolizer status.
Chain of Evidence
n Association of genotype with plasma levels of active tamoxifen metabolites:
Five prospective cohort studies of adjuvant tamoxifen treatment provide consistent evidence
that CYP2D6 nonfunctional variant alleles are associated with significantly reduced plasma
endoxifen levels. However, endoxifen levels overlap across all genotypes, indicating that
CYP2D6 genetic variability only partly explains endoxifen level variability. Although generation of 4-hydroxy tamoxifen (4-OH tamoxifen), another active tamoxifen metabolite, does not
depend solely on CYP2D6, 3 of 4 studies showed that low CYP2D6 function was associated
with reduced plasma 4-OH tamoxifen levels. Coadministration of potent CYP2D6 inhibitors
to CYP2D6 homozygous wild-type patients is associated with endoxifen levels near those of
patients who are PMs.
n Association of in vivo endoxifen levels with clinical outcomes:
Two studies examined the relationship between CYP2D6 genotype and active tamoxifen
metabolites, and between genotype and clinical outcomes in the same patient population. Both
studies enrolled patients from Asian populations, focusing almost exclusively on the prevalent
reduced-function CYP2D6*10 variant in this population. Both studies reported reduced endoxifen and/or 4-OH tamoxifen concentrations in patients homozygous or heterozygous for variant
alleles, and in conjunction reported decreased disease- or recurrence-free survival. One or both
studies have design flaws likely resulting in selection bias. In addition both studies are small,
resulting in estimates of association with wide confidence intervals. Thus, the relationship
between endoxifen (or 4-OH tamoxifen) plasma concentrations and clinical outcomes has not
been established in Asian populations, nor has it been studied in white populations with null
function CYP2D6 genotypes.
Direct Evidence
Association of genotype with clinical outcomes:
One group* directly compared tamoxifen-treated women with those clinically eligible for but
not receiving tamoxifen, stratified by CYP2D6 genotype. This retrospective study used archived
tumor samples from a randomized controlled trial of tamoxifen treatment in the adjuvant setting.
Investigators found that tamoxifen-treated EMs obtained no significant clinical benefit compared
with EMs not treated with tamoxifen, and, paradoxically, that carriers of a CYP2D6*4 nonfunctional variant allele benefited from tamoxifen treatment. There were several limitations to this
study such that results are questionable.
Most included studies examined tamoxifen use in the adjuvant setting in postmenopausal women.
Most enrolled only tamoxifen-treated women and evaluated outcomes by CYP2D6 genotype. A few
separately evaluated a non-tamoxifen-treated control population but without direct comparison.
Seven small studies in Asian populations focused on the CYP2D6*10 reduced function allele; 4
reported significant results for the association of CYP2D6 genotype with outcomes of tamoxifen
treatment, but may be affected in unpredictable ways by different types of bias (survival bias and
adjustment for variables that are not confounders). Two of the 3 studies that reported no association may have less potential for bias. Sixteen studies evaluated samples from primarily white
patients. Of the 5 largest studies, 3 reported no significant association between CYP2D6 reduced
or null activity genotype and time to breast cancer recurrence. Two of the negative studies were
retrospective analyses of clinical trial samples and one was a matched case-control study nested
within a well-documented breast cancer registry. All 3 were designed to minimize the potential for
bias; their sizes allowed comparison of homozygous nonfunctional CYP2D6 genotypes (PMs) with
fully functional wild-type genotypes (EMs), that is, the most extreme comparison and most likely
* Wegman P, Vainikka L, Stal O et al. (2005). Genotype of metabolic enzymes and the benefit of tamoxifen in postmenopausal
breast cancer patients. Breast Cancer Res, 7(3):R284-90.
2
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
to reveal a true association. Two studies reported significant positive results; 1 study combined
samples from different sources, some of which had already been analyzed for this hypothesis.
In addition, it is unclear from the report whether nearly half of the samples were obtained from
patients who had survived and were available at a time distant from their diagnosis and surgery, a
survivorship selection bias that can unpredictably affect results. The other positive study matched
samples from a randomized controlled trial on several breast cancer prognostic variables, neither
causally related to CYP2D6 genotype nor surrogates for genotype, potentially biasing selection
of cases and controls. The remaining 9 studies report a variety of significant and nonsignificant
results; no pattern of bias, genotyping or group scheme, or accounting for CYP2D6 inhibitor use
(among possibilities) explains the differences in results. Heterogeneity of results across all studies,
and clear results of no genotype-tamoxifen treatment outcome in 2 large trials with the least
apparent potential for bias, suggest lack of support for clinical validity.
There is no direct evidence of clinical utility (whether use of CYP2D6 genotype testing for endocrine therapy regimen selection improves recurrence and survival outcomes). Without demonstrable clinical validity, there is no basis for changing management in patients with specific genotypes
to improve outcomes (clinical utility).
Author’s Conclusions and Comments
The question examined in this Assessment is whether patients with CYP2D6 gene variants that
result in markedly reduced or absent enzyme function have reduced tamoxifen metabolism and
lower endoxifen levels compared with genotypic wild-type extensive metabolizers, and as a result
have poorer clinical outcomes. This question rests on the assumption, not yet supported by evidence, that some level of endoxifen is sufficient and necessary for tamoxifen efficacy, and that
this level is not achieved in patients with markedly reduced or no CYP2D6 enzymatic function.
However, because tamoxifen metabolism is complex and CYP2D6 does not appear to account
for all variability in endoxifen levels, it is conceivable that polymorphisms in other tamoxifen
metabolic pathway enzymes may affect active metabolite levels, and in theory direct measurement of the metabolite(s) itself might be the better predictor of benefit from tamoxifen treatment.
However, measuring metabolite levels is not practical for clinical applications.
Whether lower endoxifen levels can affect the pharmacodynamics of tamoxifen, the interaction
of tamoxifen metabolites with estrogen receptors, and ultimately tamoxifen efficacy, is unclear.
Dissociation constants of even the more weakly binding molecules, including tamoxifen itself,
are reportedly still sufficient to effectively block estrogen binding. Moreover, it is estimated that
at doses used for adjuvant treatment, which are intended to saturate the estrogen receptor, more
than 99% of estrogen receptors are bound by tamoxifen and its metabolites.
Lacking the appropriate mechanistic evidence, it remains to examine the clinical evidence, the
bulk of which addresses clinical validity, the CYP2D6 genotype-tamoxifen treatment outcome
association. As noted, heterogeneous results are observed across studies. Heterogeneity in effect
estimates, both in magnitude and significance, is likely due to the lack of power in most studies
and potential biases. The analysis of archived samples from 2 large completed clinical trials
was undertaken to achieve adequate power, to more fully evaluate CYP2D6 genotype, to evaluate
aromatase inhibitor-treated control populations in tandem, and to avoid potential sources of
bias. That the results of these studies discovered no evidence of association between CYP2D6
genotype and either tamoxifen- or aromatase inhibitor-treated patient outcomes has suggested
that using results of CYP2D6 genetic testing to influence decisions about tamoxifen treatment is
not currently warranted.
Based on the available evidence, the Blue Cross and Blue Shield Association Medical Advisory
Panel made the following judgments about whether CYP2D6 genotyping for directing endocrine
therapy regimen selection for women at high risk for primary breast cancer or breast cancer
recurrence meets the Blue Cross and Blue Shield Association Technology Evaluation Center
(TEC) criteria.
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.3
Technology Evaluation Center
1. The technology must have final approval from the appropriate governmental
regulatory bodies.
The Roche AmpliChip CYP450 Test is cleared by the U.S. Food and Drug Administration (FDA) and
is “intended to identify a patient’s CYP2D6 and CYP2C19 genotype from genomic DNA extracted
from a whole blood specimen. Information about CYP2D6 and CYP2C19 genotype may be used as
an aid to clinicians in determining therapeutic strategy and treatment dose for therapeutics that
are metabolized by the CYP2D6 or CYP2C19 gene product” (http://molecular.roche.com/assays/
Pages/AmpliChipCYP450Test.aspx).
CYP2D6 genotyping assays are also available as laboratory-developed tests (LDT). Clinical laboratories may develop and validate tests in-house and market them as a laboratory service; laboratories offering LDTs as a clinical service must meet the general regulatory standards of the Clinical
Laboratory Improvement Act (CLIA) and must be licensed by CLIA for high-complexity testing.
FDA has considered updating the label for tamoxifen (brand and generics) with information or
recommendations regarding CYP2D6 genotyping and impact on tamoxifen efficacy. On October
18, 2006, FDA held an Advisory Committee meeting to answer specific questions regarding the
evidence and recommendations for the label update. Since that Advisory Committee meeting,
AstraZeneca, the brand name (Nolvadex®) manufacturer, has ceased producing tamoxifen and is
no longer maintaining the prescribing information. As of the date of this Assessment, no direction has come from FDA regarding revised labeling of generic versions of tamoxifen to include
CYP2D6 genotyping information.
2. The scientific evidence must permit conclusions concerning the effect of the technology
on health outcomes.
There are several limitations to the overall body of evidence, but the largest, most well-designed
studies do not support clinical validity of the test. In the absence of evidence for clinical validity,
evidence to support clinical utility is lacking.
3. The technology must improve the net health outcome.
Evidence for clinical utility is currently lacking.
4. The technology must be as beneficial as any established alternatives.
Because the available evidence does not clearly support a significant association between CYP2D6
genotype and tamoxifen treatment outcome, a chain of evidence supporting the clinical utility of
CYP2D6 genotyping for directing endocrine therapy regimen selection for women at high risk for
or with breast cancer cannot be constructed.
5. The improvement must be attainable outside the investigational settings.
The use of CYP2D6 genotyping for directing endocrine therapy regimen selection for women at
high risk for or with breast cancer to improve health outcomes has not been demonstrated in the
investigational setting.
Based on the above, CYP2D6 genotyping does not meet the TEC criteria for directing endocrine
therapy regimen selection for women at high risk for primary breast cancer or breast cancer
recurrence.
4
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
Contents
Assessment Objective
6
Background6
Methods12
Problem Formulation
13
Review of Evidence
14
Discussion32
Summary of Application of the
Technology Evaluation Criteria
33
References35
Appendix39
Published in cooperation with Kaiser Foundation Health Plan and
Southern California Permanente Medical Group.
TEC Staff Contributors
Author—Joan Glacy, M.D.; TEC Executive Director—Naomi Aronson, Ph.D.; TEC Director, Technology
Assessments—Mark D. Grant, M.D., M.P.H.; Director, Clinical Science Services—Kathleen M. Ziegler, Pharm.D.;
Research/Editorial Staff—Claudia J. Bonnell, B.S.N., M.L.S.; Kimberly L. Hines, M.S.
Acknowledgment
Staff would like to acknowledge the work of Margaret A. Piper, Ph.D., M.P.H., in the research and development of this Assessment.
Blue Cross and Blue Shield Association Medical Advisory Panel
Trent T. Haywood, M.D., J.D.—Chairman, Senior Vice President, Clinical Affairs/Medical Director, Blue Cross and Blue Shield
Association; Steven N. Goodman, M.D., M.H.S., Ph.D.—Scientific Advisor, Dean for Clinical and Translational Research, Stanford
University School of Medicine, Professor, Departments of Medicine, Health Research and Policy; Mark A. Hlatky, M.D.—Scientific
Advisor, Professor of Health Research and Policy and of Medicine (Cardiovascular Medicine), Stanford University School of Medicine.
 Panel Members Peter C. Albertsen, M.D., Professor, Chief of Urology, and Residency Program Director, University of
Connecticut Health Center; Sarah T. Corley, M.D., F.A.C.P., Chief Medical Officer, NexGen Healthcare Information Systems,
Inc.—American College of Physicians Appointee; Helen Darling, M.A., President, National Business Group on Health;
Josef E. Fischer, M.D., F.A.C.S., William V. McDermott Professor of Surgery, Harvard Medical School—American College
of Surgeons Appointee; I. Craig Henderson, M.D., Adjunct Professor of Medicine, University of California, San Francisco;
Jo Carol Hiatt, M.D., M.B.A., F.A.C.S., Chair, Inter-Regional New Technology Committee, Kaiser Permanente; Saira A. Jan,
M.S., Pharm.D., Associate Clinical Professor, Ernest Mario School of Pharmacy, Rutgers, The State University
of New Jersey, Residency Director and Director of Clinical Programs Pharmacy Management, Horizon Blue Cross and
Blue Shield of New Jersey; Thomas Kowalski, R.Ph., Clinical Pharmacy Director, Blue Cross Blue Shield of Massachusetts;
Lawrence Hong Lee, M.D., M.B.A., F.A.C.P., Vice President and Executive Medical Director for Quality and Provider Relations,
Blue Cross and Blue Shield of Minnesota; Bernard Lo, M.D., Professor of Medicine and Director, Program in Medical
Ethics, University of California, San Francisco; Randall E. Marcus, M.D., Charles H. Herndon Professor and Chairman,
Department of Orthopaedic Surgery, Case Western Reserve University School of Medicine; Barbara J. McNeil, M.D., Ph.D., Ridley
Watts Professor and Head of Health Care Policy, Harvard Medical School, Professor of Radiology, Brigham and Women’s Hospital;
William R. Phillips, M.D., M.P.H., Clinical Professor of Family Medicine, University of Washington—American Academy
of Family Physicians’ Appointee; Richard Rainey, M.D., Medical Director, Regence BlueShield of Idaho; Rita F. Redberg,
M.D., M.Sc., F.A.C.C., Professor of Medicine and Director, Women’s Cardiovascular Services, University of California San Francisco;
Maren T. Scheuner, M.D., M.P.H., F.A.C.M.G., Chief, Medical Genetics, VA Greater Los Angeles Healthcare System; Associate
Clinical Professor, Department of Medicine, David Geffen School of Medicine at UCLA, Affiliate Natural Scientist, RAND Corporation;
J. Sanford Schwartz, M.D., F.A.C.P., Leon Hess Professor of Medicine and Health Management & Economics, School of Medicine and
The Wharton School, University of Pennsylvania.
CONFIDENTIAL: This document contains proprietary information that is intended solely for Blue Cross and Blue Shield Plans
and other subscribers to the TEC Program. The contents of this document are not to be provided in any manner to any other
parties without the express written consent of the Blue Cross and Blue Shield Association.
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.5
Technology Evaluation Center
Assessment Objective
Tamoxifen is a nonsteroidal anti-estrogen
drug classified as a selective estrogen
receptor modulator. Current indications for
tamoxifen, approved by the U.S. Food and Drug
Administration (FDA), are:
n Treatment of metastatic (Stage IV) breast
cancer
n Adjuvant breast cancer therapy to prevent
breast cancer recurrence
n Reduction of the risk of invasive breast
cancer in women with ductal carcinoma
in situ (DCIS)
n Reduction in the incidence of breast cancer
in women at high risk for breast cancer
The cytochrome P450 2D6 (CYP2D6) metabolic
enzyme has a major role in tamoxifen metabolism. The CYP2D6 gene is polymorphic; variant
DNA gene sequences resulting in proteins with
reduced or absent enzyme function may be
associated with lower plasma levels of active
tamoxifen metabolites, which could have an
impact on tamoxifen treatment efficacy.
This Assessment evaluates the evidence for
CYP2D6 genotyping, compared with no testing,
to direct treatment choices for patients at high
risk for or with breast cancer and to improve
survival outcomes.
Background
Tamoxifen Metabolism
Tamoxifen undergoes extensive primary and
secondary metabolism. Tamoxifen metabolites, rather than tamoxifen itself, are likely
the primary effectors of tamoxifen benefit.
Unlike a classic prodrug, however, tamoxifen
retains some weak receptor binding activity
(Table 1). N-desmethyl tamoxifen is the most
abundant tamoxifen metabolite, but has receptor binding activity no greater than the parent
drug. The metabolite, 4-hydroxytamoxifen
(4-OH tamoxifen), has demonstrated 100-fold
greater affinity for the estrogen receptor and
30- to 100-fold greater potency in suppressing
estrogen-dependent in vitro cell proliferation
when compared with the parent drug (summarized in Goetz et al. 2008). However, 4-OH
tamoxifen represents less than 20% of the
product of tamoxifen primary metabolism
(Fabian et al. 1981). The secondary metabolite,
4-hydroxy-N-desmethyl tamoxifen (endoxifen),
has identical properties and potency compared
6
with 4-OH tamoxifen in terms of its binding
affinity to estrogen receptors, suppression
of in vitro estrogen-receptor-dependent cell
proliferation, and gene expression of progesterone receptors, a marker of estrogenic effect
(Stearns et al. 2003; Johnson et al. 2004; Lim et
al. 2005; Lim et al. 2006). Steady-state plasma
endoxifen concentrations are, on average, 5- to
10-fold higher than 4-OH tamoxifen (Stearns et
al. 2003), although variability among individuals may be high (2- to 23-fold; Jin et al. 2005). It
has been assumed that endoxifen is the major
active metabolite of tamoxifen.
Endoxifen is formed predominantly by the
CYP2D6-mediated oxidation of N-desmethyl
tamoxifen. The CYP2D6 enzyme has known
interindividual variability in activity and therefore has been of great interest in investigating
tamoxifen metabolism and variation in circulating active metabolite levels.
Metabolic Enzyme Genotypes
The CYP2D6 gene exhibits a high degree of
polymorphism, with more than 100 allelic variants identified (The Human Cytochrome P450
[CYP] Allele Nomenclature Database; available at http://www.cypalleles.ki.se). Although
the most prevalent CYP2D6 *1 and *2 alleles
(both termed “wild-type” for this Assessment)
produce an enzyme with normal activity, there
are several variant (V) alleles that result in
enzymes with no activity or reduced activity (see Table 2). Because individuals have 2
CYP2D6 alleles, various combinations of the
possible alleles result in a spectrum of CYP2D6
function; these have been categorized as extensive metabolizers (EM or “normal”), intermediate metabolizers (IM), and poor metabolizers
(PM), based originally on pharmacokinetic
studies of CYP2D6-dependent probe drugs
before the discovery of the genetic basis of
variable function. An additional, rare category
of ultra-rapid metabolizers (UM) is defined by
the possession of 3 or more functional alleles
due to gene duplication. UMs have greater
functional activity than EM genotypes because
of additional expression of enzyme from the
extra gene(s).
Griese et al (1998) studied the correlation of
CYP2D6 functional categories, determined by
metabolic capacity using a probe drug (sparteine), with genotypes in 195 white individuals in
Germany. Although all PMs were “unambiguously identified as carriers of two nonfunctional alleles ... the most frequent functional
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
Table 1. Comparison of Tamoxifen and 3 Metabolites: Metabolism, Serum Concentration, and Estrogen Receptor
Binding Affinity (Dowsett and Haynes 2003; Lee et al. 2003; Jin et al. 2005; Goetz et al. 2008)
Metabolite
Predominant CYP Isoform
for Formation
Mean Serum
Concentration at Steady
State, mean (SD), ng/mLa
Mean Estrogen Receptor
Affinity, mean (SD)b
TAM
NA
116 (28)
3 (2)
N-desmethyl TAM
CYP3A4/5
215 (62)
2 (1)
4-OH TAM
CYP2D6
3 (1)
181 (84)
CYP2D6
5 (3)c
181d
Primary metabolites
Secondary metabolite
4-hydroxy-N-desmethyl
TAM (endoxifen)
TAM: tamoxifen; NA: not available
a
In women receiving tamoxifen 20 mg daily.
b
Reported in multiples of estradiol receptor binding affinity in animal studies.
c
Derived from 17 women, genotype status, and concurrent use of CYP2D6 inhibitors unknown (Lee et al. 2003).
d
Reported in Lash et al. 2009 without citation. Johnson et al. (2004) reported that the relative affinity of 4-OH-tamoxifen and endoxifen
was approximately 35% and 25% that of estradiol, respectively, in in vitro radioligand competitive receptor binding assays. In fluorescence
receptor binding assays (i.e., comparison of 4-OH-tamoxifen and endoxifen to displace a fluorescent estrogen from recombinant estrogen
receptors), both compounds had apparently identical estrogen receptor affinity.
Table 2. Frequencies of the Most Prevalent CYP2D6 Alleles Across Ethnic Groups
Allele Type
CYP2D6 Allele
Caucasian (%)
African-American (%)
Asian (%)
Functional
*1
33–40
28–50
23–42
*2
22–34
11–78
9–20
*9
0–2.9
0
3.3
*10
1.9–8
3.1–8.6
38–70
*17
0.1–0.3
9–34
0.5
*41
8
–
–
*3
1–3.9
0–0.5
0.8–1
*4
12–23
1.2–7
0–2.8
*5
1.6–7.3
0.6–6.1
4.5–6.1
*6
0.7–1
0
–
*8
(rare)
(rare)
(rare)
*1 × 2
0.2–0.5
3.3
0.5
*2 × 2
0.7–1.6
1.6–2.5
0–1
*4 × 2
0.1–0.2
0.9
–
Reduced function
Nonfunctional
Duplication
Reproduced from Beverage et al. CYP2D6 polymorphisms and the impact on tamoxifen therapy. J Pharm Sci, 2007; 96(9):2224-31;
Copyright © 2007 American Pharmacists Association; Reprinted with permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons,
Inc. Modified with information on *8 from Sachse et al. (1997); Kubota et al. (2000); Ji et al. (2002); Griese et al. (1999).
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.7
Technology Evaluation Center
genotypes extensively overlapped.” The authors
concluded that, except for PMs, genotype was
not a useful predictor of function. Thus, fully
functional (homozygous wild-type) genotypes
are consistently assigned to the EM category
and homozygous inactive variant genotypes
are consistently assigned to the PM category
in pharmacogenomic studies. However,
assignment of other genotypes with function
in between these 2 is inconsistent among
authors and requires standardization so that
results may be better compared across studies
(Beverage et al. 2007).
The prevalence of CYP2D6 PMs, defined either
by metabolic function or by the detection of 2
nonfunctional alleles, has been estimated in
various ethnic populations in several studies,
summarized by Bernard et al. (2006). The PM
prevalence is approximately 7% to 10% in
whites of Northern European descent, 1.9%
to 7.3% in African Americans, and 1% or less
in most Asian populations studied. The PM
phenotype in whites is largely accounted for by
CYP2D6*3 and *4 nonfunctional variants (see
Table 2). In African American and Asian populations, the nonfunctional *5 variant allele is
present in at most 6% to 7%, but homozygous
genotypes are relatively rare. However, some
PMs may reflect the combination of a nonfunctional and a reduced function allele. Among
reduced function variants, *17, *10, and *41
are the most important in African Americans,
Asians, and whites, respectively. Few studies
have investigated the frequency of CYP2D6
variant alleles or of PMs in the Hispanic population (Bernard et al. 2006).
The impact of CYP2D6 genotype and phenotype is the most extensively studied
pharmacogenomic influence on tamoxifen
treatment, and for that reason is the sole topic
of this Assessment. Other enzymes involved
in tamoxifen metabolism include CYP2B6,
CYP2C9, CYP2C19, and CYP3A. Enzymes
involved in metabolite elimination include
sulfotransferase 1A1 (SULT1A1) and uridine
diphosphate glucuronosyltransferase 2B15
(UGT2B15). Endocrine therapy efficacy is
also affected by other proteins, such as those
that regulate estrogen receptor activity and
enzymes involved in the metabolism of aromatase inhibitors (AIs). Polymorphisms that may
affect overall efficacy of tamoxifen and other
endocrine treatments have been found in the
genes for these proteins and enzymes. Research
in these areas is in an earlier discovery stage
8
compared with the research on CYP2D6, and
will not be discussed further in this Assessment.
Genetic Variability and Endocrine
Therapy Regimens
Current National Comprehensive Cancer
Network (NCCN, 2013) breast cancer treatment
and risk reduction recommendations for FDAapproved uses of tamoxifen are shown in Table
3. Because a small, but significant, proportion of
most ethnic populations have markedly reduced
CYP2D6 metabolic capacity, there is concern
that similar proportions of patients treated with
tamoxifen may have poorer outcomes than
patients with full CYP2D6 activity. Some have
recommended that patients who are to receive
tamoxifen be genotyped for CYP2D6, and that
PMs be treated with alternative therapy, if possible. As shown in Table 3, tamoxifen is the only
adjuvant treatment approved for preventing
breast cancer in premenopausal women with
DCIS (about 20% of all new breast cancers;
American Cancer Society 2012) and in high-risk
premenopausal women. Therefore, pharmacogenomic evaluation would not change treatment in these women.
In the adjuvant setting, tamoxifen currently is
the most commonly prescribed treatment for
hormone receptor-positive breast cancer.
For women who are premenopausal at diagnosis, pharmacogenomic evaluation could
direct consideration of ovarian ablation or
suppression in CYP2D6 PMs. Ovarian ablation
(oophorectomy or irradiation) is an effective
treatment compared with no adjuvant therapy
(Adjuvant Breast Cancer Trials Collaborative
Group [ABCTCG] 2007), but does not appear
to add benefit to adjuvant chemotherapy
(ABCTCG 2007). Acute and chronic adverse
effects include hot flushes, sweats, and sleep
disturbance. Functional ovarian suppression
(with gonadotropin-releasing-factor analogs) in
women with hormone receptor-positive tumors
confers benefits comparable to chemotherapy.
In postmenopausal women, who make up the
majority of patients with breast cancer, raloxifene is an alternative treatment option, with
efficacy equal to that of tamoxifen but markedly reduced risk of endometrial hyperplasia.
Although the steroidal aromatase inactivator,
exemestane, is recommended for risk reduction in high-risk postmenopausal women, it is
not currently FDA-approved for this indication
(see full prescribing information at http://www.
aromasin.com/).
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
Table 3. NCCN-Recommended Endocrine Therapy for Women With or at Risk for Breast Cancer
Indication
Treatment
Class
SOR
Surgical or radiotherapeutic
2A
LHRH agonist (goserelin or leuprolide)
2A
–Tamoxifen
–Toremifene
ER modulatorb
2A
–Tamoxifen
–Toremifene
ER modulator
2A
–Anastrozole
–Letrozole
–Exemestane
AI
2A
Fulvestrant
ER down regulator
2A
Megestrol acetate
Progestin
2A
Fluoxymesterone
Androgen
2A
Ethinyl estrogen
High dose estrogen
2A
Metastatic disease
Premenopausal at diagnosis
Ovarian ablationa
Ovarian suppression
Postmenopausal at diagnosisb
Premenopausal or
postmenopausalb at diagnosis
a
Adjuvant therapy to prevent recurrence
Premenopausal at diagnosis
Tamoxifen for 5 y
1
± Ovarian ablation/suppression
2B
Followed by:
Postmenopausal at diagnosis
– AI for 5 y if postmenopausal
1
– Tamoxifen to complete 10 y (pre- or postmenopausal)
1
– No further endocrine therapy if premenopausal
2A
AI for 5 y
1
AI for 2–3 y plus tamoxifen to complete 5 y
1
Tamoxifen for 2–3 y plus AI to complete 5 y
1
Tamoxifen for 2–3 y plus AI for 5 y
2B
Tamoxifen for 4.5–6 y plus AI for 5 y
1
Tamoxifen for 9.5–10 y
1
Tamoxifen for 5–10 y c
1
Risk reduction after DCIS or in high-risk women
Premenopausal
Tamoxifen
Postmenopausal
Tamoxifen
1
Raloxifene
1
Exemestane
ER modulator
AI
1
2A
Estrogen receptor (ER)– and/or progesterone receptor (PR)–positive disease predicts likely benefit from tamoxifen. However, due to (1) the
possibility of false-negative ER and PR tumor assessments, (2) heterogeneity in receptor expression between primary and metastatic sites,
(3) low toxicity of endocrine therapy, and (4) potential efficacy, NCCN recommends consideration of endocrine therapy in patients with
hormone receptor-negative disease localized to bone or soft tissue or with asymptomatic visceral disease only. (NCCN, 2013). Guidelines
available at www.nccn.org.
AI, aromatase inhibitor; ER, estrogen receptor modulator; LHRH, luteinizing hormone releasing hormone; SOR, strength of
recommendation
a
Ovarian ablation/suppression is followed by endocrine therapy as for postmenopausal women.
b
If no prior endocrine therapy within 1 year.
c
For women with a contraindication to AI or who decline or are intolerant of AI therapy.
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.9
Technology Evaluation Center
In clinical practice, AIs may eventually replace
tamoxifen for adjuvant therapy in postmenopausal women with hormone receptor-positive
tumors because of fewer adverse effects and
equal or better efficacy. In the multicenter,
double-blind Arimidex, Tamoxifen, Alone or in
Combination (ATAC) trial, adjuvant treatment
with anastrozole at 1 mg daily was compared
with tamoxifen at 20 mg daily. Recurrence-free
survival (RFS) at approximately 8 years was
improved in the anastrozole arm compared
with the tamoxifen arm: hazard ratio (HR)=0.85,
95% confidence interval (CI), 0.76 to 0.94,
p=0.003 (ATAC Trialists’ Group 2008); there was
no significant effect on overall survival. Table 4
shows results for prespecified adverse events in
the ATAC trial at a median follow-up of 5 years;
in particular, patients receiving anastrozole had
a statistically significant decrease in hot flashes,
vaginal bleeding, vaginal discharge, endometrial cancer, venous thromboembolic events
(including deep venous thrombosis), and ischemic cerebrovascular events compared with
patients receiving tamoxifen (shaded areas of
Table 4). There were no significant differences
in risk of cardiovascular morbidity or mortality;
fracture rates were higher in patients treated
with anastrozole versus tamoxifen during active
treatment, but not after treatment completion
(ATAC Trialists’ Group 2008).
Results from the Breast International Group
(BIG) 1-98 trial, designed to compare AI alone
with sequential AI + tamoxifen and to tamoxifen alone, indicated that AI (letrozole) alone
results in significantly fewer early relapses
than tamoxifen (Mauriac et al. 2007). However,
sequential use of tamoxifen and AI did not
show improved disease-free survival (DFS)
compared with AI monotherapy at a median of
71 months follow-up (Mouridsen et al. 2009).
Thus, tamoxifen may not be the first choice
for adjuvant treatment in postmenopausal
women. However, because AIs are contraindicated in premenopausal women, and because
tamoxifen remains important in the treatment
of metastatic cancer where either tamoxifen
or AI resistance may develop, the use of pharmacogenomics to improve the likelihood of
tamoxifen benefit remains of interest.
Pharmacologic Inhibitors of Metabolic Enzymes
CYP2D6 activity may be affected not only by
genotype, but also by coadministration of drugs
Table 4. Number (%) of Patients With Prespecified Adverse Events in the ATAC Triala
Adverse Events
Anastrozole
(n=3,092), n (%)
Tamoxifen
(n=3,094), n (%)
Odds Ratio
95% CI
All fractures
315 (10)
209 (7)
1.57
1.30 to 1.88
Fractures of spine, hip, wrist
133 (4)
91 (3)
1.48
1.13 to 1.95
Musculoskeletal eventsb
1100 (36)
911 (29)
1.32
1.19 to 1.47
Ischemic cardiovascular disease
127 (4)
104 (3)
1.23
0.95 to 1.60
Mood disturbances
597 (19)
554 (18)
1.10
0.97 to 1.25
Fatigue/asthenia
575 (19)
544 (18)
1.07
0.94 to 1.22
Nausea and vomiting
393 (13)
384 (12)
1.03
0.88 to 1.19
Cataracts
182 (6)
213 (7)
0.85
0.69 to 1.04
Hot flashes
1104 (36)
1264 (41)
0.80
0.73 to 0.89
Ischemic cerebrovascular events
62 (2)
88 (3)
0.70
0.50 to 0.97
Deep venous thromboembolic events
48 (2)
74 (2)
0.64
0.45 to 0.93
Venous thromboembolic events
87 (3)
140 (5)
0.61
0.47 to 0.80
Vaginal bleeding
167 (5)
317 (10)
0.50
0.41 to 0.61
Vaginal discharge
109 (4)
408 (13)
0.24
0.19 to 0.30
Endometrial cancer
4 (0.2)
13 (0.6)
0.31
0.10 to 0.94
a
b
Reproduced from anastrozole (Arimidex®) prescribing information (available at http://www.arimidex.com/).
Refers to joint symptoms, including joint disorder, arthritis, arthrosis, and arthralgia.
10
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
that inhibit the metabolic activity of CYP2D6.
For example, in healthy volunteers found to
be EM phenotype, CYP2D6 activity was determined by measuring the metabolic products
of CYP2D6-dependent probe drugs before and
after administration of potential CYP2D6 inhibitory medications. Studies of selective serotonin
reuptake inhibitors (SSRIs) in particular have
shown that fluoxetine and paroxetine, but not
sertraline, fluvoxamine, or venlafaxine, are
potent CYP2D6 inhibitors (Alfaro et al. 1999;
Alfaro et al. 2000; Lam et al. 2002). Some individuals treated with fluoxetine or paroxetine
changed from EM phenotype to PM (Alfaro et
al. 1999). The degree of inhibition may depend
on the SSRI dose; for example, sertraline may
be weakly if at all inhibitory at 50 mg but can
become a potent inhibitor at higher doses
(Sproule et al. 1997).
SSRIs are often prescribed to alleviate hot
flashes, which can be an adverse effect of
tamoxifen therapy. However, research has suggested that hot flashes accompanying tamoxifen
treatment predict a lower likelihood of breast
cancer recurrence (Mortimer et al. 2007).
Reduction in hot flashes with SSRI coadministration may reflect CY2D6 inhibition, reduction
of tamoxifen metabolism, and reduced levels
of active tamoxifen metabolites, resulting in
poorer outcomes. Contrasting with this hypothesis are results from a study indicating that
venlafaxine, at doses previously shown not to
inhibit CYP2D6 activity, can also reduce hot
flashes during tamoxifen treatment (Loprinzi
et al. 2000).
Currently, both NCCN breast cancer guidelines
(2013) and an American Society of Clinical
Oncology (ASCO) clinical practice guideline
update on adjuvant endocrine therapy for
women with hormone receptor–positive breast
cancer (Burstein et al. 2010) recommend
caution regarding coadministration of certain
serotonin reuptake inhibitors that are strong
inhibitors of CYP2D6 (fluoxetine and paroxetine). Because CYP2D6 inhibitors may have
the potential to change the CYP2D6 phenotype, it has been recommended that studies of
CYP2D6 genotype and tamoxifen treatment outcomes account for the use of CYP2D6 inhibitors
in assigning CYP2D6 functional status.*
Guidelines
Regarding the use of CYP2D6 genetic testing
before prescribing tamoxifen, NCCN breast
cancer guidelines (NCCN 2013) state, “At this
time, based on current data the panel does not
endorse routine CYP2D6 testing for women
being considered for tamoxifen therapy.”
The ASCO 2009 guidelines update (Burstein et
al. 2010) states, “The Update Committee recommends against using CYP2D6 genotype to
select adjuvant endocrine therapy.”
Regulatory Status
The Roche AmpliChip CYP450 Test is cleared
by FDA and is “intended to identify a patient’s
CYP2D6 and CYP2C19 genotype from genomic
DNA extracted from a whole blood specimen. Information about CYP2D6 and CYP2C19
genotype may be used as an aid to clinicians
in determining therapeutic strategy and treatment dose for therapeutics that are metabolized
by the CYP2D6 or CYP2C19 gene product”
(http://molecular.roche.com/assays/Pages/
AmpliChipCYP450Test.aspx).
CYP2D6 genotyping assays are also available
as laboratory developed tests (LDTs). Clinical
laboratories may develop and validate tests
in-house and market them as laboratory services; laboratories offering an LDT as a clinical
service must meet the general regulatory standards of the Clinical Laboratory Improvement
Act (CLIA) and must be licensed by CLIA for
high-complexity testing. Although FDA has
technical authority to regulate LDTs, to date
there has been no active oversight.
FDA has been considering updating the label
for tamoxifen (brand and generics) with information or recommendations regarding CYP2D6
genotyping and impact on tamoxifen efficacy.
On October 18, 2006, FDA held an Advisory
Committee meeting to answer specific questions regarding the evidence and recommendations for the label update; the questions and a
summary of the Advisory Committee responses
are presented in Table 5. Since the Advisory
Committee meeting, AstraZeneca, the brand
name (Nolvadex®) manufacturer, has ceased
producing tamoxifen and is no longer maintaining the prescribing information. As of the
* CYP2D6 is not considered to be an inducible enzyme in vivo (Ingelman-Sundberg 2005; Lynch and Price 2007).
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.11
Technology Evaluation Center
Table 5. Questions Submitted to the FDA Advisory Committee for Pharmaceutical Science, Clinical
Pharmacology Subcommittee (October 18, 2006) and Summaries of Subcommittee Answers from the Final
Minutes of the Advisory Committee Meeting
Questions
Answers
1.The scientific evidence on the metabolism of
tamoxifen demonstrates that CYP2D6 is an
important pathway in the formation of endoxifen.
No disagreement.
2.The pharmacologic and clinical evidence
are sufficient to demonstrate that endoxifen
significantly contributes to the pharmacologic
Although the Subcommittee felt that CYP2D6
contributed clinically to the level of endoxifen in in
vitro data, there was no direct concentration/response
information to indicate that endoxifen is a major
contributor to the clinical effect of tamoxifen.
(anti-estrogenic) effect of tamoxifen.
3.Does the clinical evidence demonstrate that
postmenopausal women with ER-positive breast
cancer who are CYP2D6 poor metabolizers (by
genotype or drug interaction) are at increased
risk for breast cancer recurrence? If yes, should
the tamoxifen label include information about
increased risk for breast cancer recurrence in
CYP2D6 poor metabolizers prescribed tamoxifen?
The label should be updated to reflect the increased
risk for breast cancer along with the mechanistic data
presented.
4.Is there sufficient scientific and clinical evidence
to support revisions of the tamoxifen label
that recommends CYP2D6 genotype testing
for postmenopausal patients before they are
prescribed tamoxifen for adjuvant treatment?
The Subcommittee did not reach consensus on this
question. Some members felt that the genetic test
should be RECOMMENDED while others felt that it
should be mentioned in the label as an OPTION for
discussion between the health care provider and
patient. However, the majority indicated that it should
be included in an appropriate section of the package
insert.
Available at http://www.fda.gov/ohrms/dockets/ac/cder06.html.
ER, estrogen receptor
date of this Assessment, FDA has delivered
no direction on revised labeling of generic
versions of tamoxifen to include CYP2D6
genotyping information.
Methods
In June 2009, after publication of a study indicating that certain SSRIs in combination with
tamoxifen increased breast cancer recurrence
rates (Aubert et al. 2009), FDA indicated that
changes to the label regarding the interaction
of tamoxifen and certain SSRIs were likely.
However, these changes have yet to appear in
product labels.
Search Methods
MEDLINE® was searched (via PubMed) using
the search string (“Breast Neoplasms”[MeSH®]
AND “Tamoxifen”[MeSH®]) AND “Cytochrome
P-450 Enzyme System”[MeSH®] through
November 2013. Clinical trials, recent reviews
(2008-2013), editorials, and letters related to
the pharmacogenomics of tamoxifen were
retrieved. Additionally, we searched text and
reference lists of retrieved papers for other relevant articles.
Clinical Trials of CYP2D6 Pharmacogenomics
and Tamoxifen in Progress
The ClinicalTrials.gov database was searched
for ongoing studies related to tamoxifen
and CYP2D6 genotyping; Table 6 lists those
that specified evaluation of the association of
genotype and patient outcomes as an intended
result of the trial.
Study Selection
We selected full-length, peer-reviewed articles
reporting studies of postmenopausal women
undergoing endocrine therapy whose treatment regimen selection is based on CYP2D6
genotyping versus usual selection methods; or
studies of the association of CYP2D6 genotype
with intermediate (tamoxifen-active metabolite
12
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
Table 6. Studies Listed in ClinicalTrials.gov That Will Report the Association of CYP2D6 Genotypes
With Patient Outcomes
Trial No.
Country
Target n
Status
Comment
NCT00973037
Korea
922
Recruiting
Using samples from prospective randomized
multicenter study
NCT00815555
Israel
500
Unknown*
Prospective
NCT01169792
Korea
NR
Completed
No publication
NCT00532454
Korea
21
Terminated
Metastatic breast cancer
NCT01124695
United States
240
Recruiting
Patients with metastatic breast cancer treated with
single agent tamoxifen
NCT01220076
France
265
Unknown*
Nonmetastatic, hormone receptor-positive breast
cancer
NCT01181518
Korea
1,000
Recruiting
Prospective, observational
NCT00963209
Switzerland
140
Recruiting
Genotype correlation with tamoxifen metabolites
and tumor recurrence
NCT01357772
Italy
1,400
Recruiting
Low-dose tamoxifen in women with lobular and
ductal intraepithelial neoplasia
NR: not reported
* Information has not been verified in more than 2 years.
levels) or final outcomes (time to recurrence, survival) for review. Study quality was
evaluated by considering consistency of patient
populations; thoroughness of CYP2D6 allele
genotyping; whether CYP2D6 inhibitors were
taken into account when assigning patients to
metabolizer status categories; possible sources
of bias (e.g., survival bias, genotyping whole
blood from surviving participants of retrospective studies); misclassification of metabolizer
phenotype; and adjustment for variables not
considered confounders of the genotype-outcome association.
Medical Advisory Panel Review
The Assessment was reviewed by the Blue
Cross and Blue Shield Association Medical
Advisory Panel (MAP) on October 2, 2013.
This Assessment was initially reviewed by
the Panel on February 17, 2011. To maintain
the timeliness of the scientific information in
this Assessment, literature searches were performed subsequent to the Panel’s review (see
Search Methods section). If the search updates
identified additional studies that met the criteria for detailed review, the results of these
studies were included in the tables and text
where appropriate.
Problem Formulation
Patient Indications
Indications for CYP2D6 pharmacogenomic
testing include patients who are to be treated
with tamoxifen for prevention of breast cancer
in high-risk women or women with DCIS,
for adjuvant treatment to prevent breast
cancer recurrence, or for treatment of metastatic disease, and for women who have no
contraindications to treatment with AIs (for
treatment of existing disease) or raloxifene
(for prevention of disease). Postmenopausal
patients, determined to be CYP2D6 PMs, could
avoid tamoxifen therapy and be treated with
AIs alone. Premenopausal patients with breast
cancer, who are determined to be CYP2D6 PMs,
could consider ovarian ablation or suppression
in place of or in addition to tamoxifen treatment. All indications of tamoxifen use will be
considered in this Assessment, because the biologic effects of tamoxifen treatment should not
differ by treatment indication (Dieudonne et
al. 2013). For any indication, coadministration
of drugs that inhibit CYP2D6 activity should be
taken into account.
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.13
Technology Evaluation Center
Technologies to Be Compared
CYP2D6 testing and consequent alteration of
treatment regimen in CYP2D6 PMs is compared
with no testing (and no alteration of decision
to treat with tamoxifen). Tamoxifen pharmacogenomics have been most often studied in
the adjuvant setting in postmenopausal women
with hormone receptor-positive tumors.
Health Outcomes
n Breast cancer
–Occurrence
– Time to recurrence
– Disease-free survival
n Overall survival
Main health outcomes are disease prevention,
improved time to recurrence, recurrencefree survival, and/or overall survival for
patients whose treatment decisions are altered
by CYP2D6 testing. As discussed further
below, intermediate outcomes, such as
tamoxifen-active metabolite levels (endoxifen,
4-hydroxytamoxifen) do not provide sufficient
evidence unless linked to evidence of main
health outcomes.
Specific Assessment Questions
1. Are CYP2D6 genotyping assays accurate and
reliable (i.e., analytic validity)?
2. In women who need endocrine therapy,
are CYP2D6 variant genotypes significantly
associated with (i.e., do they predict) intermediate and/or main health outcomes (i.e.,
clinical validity)?
3. In women who need endocrine therapy, does
use of CYP2D6 genotype testing result in
selecting endocrine therapy regimens that
improve health outcomes compared with
selection of endocrine therapy regimens
without testing (i.e., clinical utility)?
The analytic framework for questions 2 and 3 is
shown in Figure 1.
Review of Evidence
1. Are CYP2D6 genotyping assays accurate
and reliable (i.e., analytic validity)?
The Roche AmpliChip CYP450 Test for detecting CYP2D6 variants in whole blood samples
has been fully validated for analytic validity; a summary of results submitted for FDA
14
clearance is provided in FDA’s decision
summary (available at http://www.accessdata.
fda.gov/cdrh_docs/reviews/K042259.pdf).
Compared with sequencing (criterion standard), the specificity of the AmpliChip CYP450
Test for detection of wild-type samples (n=100;
3 different wild-type alleles) was 100%; sensitivity for 22 different variant alleles in 492
alleles tested was 99.2%, with no mis-calls
and 4 no-calls. Reproducibility (99.9%) and
other measures of assay robustness indicate
highly reliable performance. The AmpliChip kit
contains positive and negative quality control
samples that must yield specified results for an
assay run to be valid.
Although comparable information on the analytic validity of LDTs is usually not available,
in an experienced laboratory and with validation of in-house results compared with either
sequencing or to AmpliChip, accurate and
reliable performance should be achievable. For
example, Heller et al. (2006) compared their
in-house method with AmpliChip and achieved
95.6% agreement; 7 discordant samples contained CYP2D6 gene amplifications not detected
by the in-house method (discordant results
were verified by sequencing).
2. In tamoxifen-eligible women who need
endocrine therapy, are CYP2D6 variant
genotypes significantly associated with
(i.e., do they predict) intermediate and/
or main health outcomes (i.e., clinical
validity)?
The analytic framework for assessing the clinical validity of CYP2D6 genotyping is shown
in Figure 1. There is direct evidence and a
chain of evidence. Direct evidence compares
outcomes in women treated with tamoxifen or
other endocrine therapy stratified by metabolizer status. The chain of evidence associates
an intermediate outcome, tamoxifen metabolite
plasma levels, with both CYP2D6 genotype and
main health outcomes in women treated with
tamoxifen.
Chain of Evidence
Association of Genotype with TamoxifenActive Metabolite Concentrations
Six studies evaluated the association of CYP2D6
genotypes with circulating 4-OH tamoxifen
and/or endoxifen levels (Table 7). All were prospective cohort studies of adjuvant treatment;
tamoxifen was administered at the standard
dose of 20 mg daily (or dose was not reported).
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
Clinical Utility (Question 3)
No Evidence Identified
CYP2D6
Genotyping
Genotype Directed
Treatment
Outcomes
TamoxifenEligible Women
Genotype
Breast cancer:
– Occurrence
– Time to recurrence
– Disease-free survival
Overall survival
Tamoxifen
Treatment
Metabolite
(Endoxifen) Levels
Clinical Validity (Question 2)
– Table 7
– Table 8
Clinical Validity (Question 2)
Observational Data:
– Table 9 (Asian Populations)
– Table 10
For question 2 (dashed lines), evidence links CYP2D6 genotype with main health outcomes in women treated with tamoxifen. A chain of evidence links CYP2D6 genotype with
main health outcomes through the intermediate outcome of tamoxifen metabolite plasma levels. For question 3 (solid lines), no evidence was identified.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.15
Figure 1. Analytic Framework for Assessment Questions 2 and 3
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
– TAM Use
–Dose
Study
n
Study Design,
Patients
Stearns et
al. (2003)
12
Prospective cohort
Breast cancer
patients taking
TAM and taking
paroxetine (SSRI)
for hot flashes
Jin et al.
(2005)
80
Prospective cohort
Newly diagnosed,
postsurgery breast
cancer patients
starting TAM
–Duration
Before
Evaluation
–Adjuvant
– 20 mg/d
– >4 wk
–Adjuvant
–(Dose not
reported)
– 4 mo
Study
Accounted
for CYP2D6
Inhibitors?
Yes
Gene
(Typed
Variants)
Comparison
Results
CYP2D6
(*4, *6, *8)
wt/V or V/V
vs. wt/wt
Baseline endoxifen concentrations were lower in V carriers
(p=0.002)
(Patients
taking other
CYP2D6
inhibitors
excluded)
No
CYP2D6 genotype
wt/wt (n=7)
wt/V or V/V (n=5)
CYP2D6
(*1,*3, *4,
*5,*6)
wt/wt vs. wt/V
vs. V/V
Pre- vs. Post-SSRI
Endoxifen Decrease %
64
24
p Value
0.03
Mean concentration after 4 mo:
Genotype (n) Endoxifen 95% CI
4-OH TAM 95% CI
wt/wt (48)
78.0 nM
65.9–90.1 9.5
8.4–10.6
wt/V (29)
43.1
33.3–52.9 8.3
6.7–9.9
V/V (3)20.011.1–28.9
7.11.2–13.0
p-value <0.001
0.86
Distribution of individual patient results shows overlap among all
groups.
Yes
No CYP2D6
inhibitor
vs.
CYP2D6 inhibitor
by genotype
Genotype ± Inhibitor (n)
wt/wt – inhib. (34)
wt/wt + inhib. (13)
wt/V – inhib. (17)
wt/V + inhib. (11)
V/V – inhib. (3)
V/V + inhib (0)
Mean Endoxifen Percent
Level (nM)
Decrease
91.4
38.6
57.8
51.7
31.0
40.0
20.0
–
p Value
0.003
0.08
–
Technology Evaluation Center
16
Table 7. Studies Reporting Association of CYP2D6 Genotype With Tamoxifen-Active Metabolite Levels
Study
n
Study Design,
Patients
Borges et al.
(2006)
158
Prospective cohort
Newly diagnosed,
postsurgery breast
cancer patients
starting TAM
– TAM Use
–Dose
–Duration
Before
Evaluation
Study
Accounted
for CYP2D6
Inhibitors?
–Adjuvant
No
– 20 mg/d
– 4 mo
Gene
(Typed
Variants)
CYP2D6
(33 alleles by
AmpliChip a
also *3, *4,
*6, *7, *8,
*10, *17 by
alternate
methods)
Comparison
Results
EM
vs.
IM
vs.
PM
Phenotype,c n
EM (68)
IM (71)
PM (19)
p Value
<0.001
Each functional category comprised a variety of possible
genotypes, listed in Borges et al. 2006. Although mean ratios
differed significantly, individual patient data points for all
groups overlapped.
c
d
Yes
Plasma Ratio, mean (SD)d
0.18 (0.09)
0.09 (0.04)
0.04 (0.02)
EM ± CYP2D6
inhibitor
Phenotype determined by genotype classification.
Plasma ratio = endoxifen to N-desmethyl tamoxifen ratio.
Plasma Percent
p Value
PhenotypeEndoxifena Decrease
EM or
84.1 (39.4)
EM + weak inhibitor vs. 93.6 (38.6)
EM + potent inhibitor
23.5 (9.5)
72.1
<0.001
PM (no inhibitors)
19.4 (6.10
a
Values are mean (SD), nM.
(Similar trends found in other major phenotypes.)
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.17
Table 7. Studies Reporting Association of CYP2D6 Genotype With Tamoxifen-Active Metabolite Levels (cont’d)
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
Study
n
Study Design,
Patients
Lim (2007)
202
Prospective cohort,
Korean women with
early or metastatic
breast cancer taking
TAM
– TAM Use
–Dose
–Duration
Before
Evaluation
–Adjuvant or
treatment
of
metastatic
disease,
– 0 mg/d,
– > 8 wk
Study
Accounted
for CYP2D6
Inhibitors?
Yes
(patients
using CYP2D6
inhibitors
excluded)
Gene
(Typed
Variants)
Comparison
Results
CYP2D6
(*2xN, *5,
*10)
*10/*10
vs.
wt/wt or wt/*10
Genotype (n)
wt/wt (64) or
wt/*10 (89) vs.
*10/*10 (49)
37
Kiyotani et
al. (2010)
98
Pre- and
postmenopausal
women with newly
diagnosed breast
cancer starting TAM
–Adjuvant
– 20 mg/d
– >4 wk
Independently
recruited patients
with breast cancer
taking TAM
–Adjuvant
– 20 mg/d
– >4 wk
Yes
No patients
were taking
CYP2D6
inhibitors
Yes
(patients taking
SSRIs not
included)
4-OH TAMa
2.8
2.5
1.5 (p<0.001)
Results were similar if *10 and *5 were combined as variant
alleles; *5 allele frequency was 5%.
a
Xu et al.
(2008)
Endoxifena 19.9 18.1
7.9 (p<0.001)
Values are ng/mL.
CYP2D6
(*10)
*10/*10
or
wt/*10
vs.
wt/wt
4-HydroxyPercent
Genotype
n
Tamoxifen
Decrease
p Value
185.3
wt/wt
7
5.2
2.0
0.96
wt/*10
*10/*10 124.1
22.6
0.04
CYP2D6
(*1; variant
alleles V:
*4, *6, *10,
*14B, *18,
*21, *36,
*41)
V/V or wt/V
vs.
wt/wt
Genotype
PlasmaPercent
Endoxifena Decrease
p Value
wt/wt35.4
27.223.2
wt/V
15.556.2 <.001
V/V
4-HydroxyPercent
Tamoxifena Decrease
p Value
Genotype
wt/wt5.3
wt/V
4.221
3.142 <.001
V/V
a
Values are ng/mL.
TAM, tamoxifen; 4-OH TAM, 4-hydroxytamoxifen; chemo, chemotherapy; V, variant; wt, wild type (*1 or *2); EM, extensive metabolizer; IM, intermediate metabolizer; PM, poor metabolizer;
SSRI, selective serotonin reuptake inhibitor
a
AmpliChip detects CYP2D6 alleles *1 to *10AB, *11, *14A, *14b, *15, *17, *19, *20, *25, *26, *29 to *31, *35, *36, *40, *41, *1xN, *2xN *4xN, *10xN, *17xN, *35xN, and *41xN.
Technology Evaluation Center
18
Table 7. Studies Reporting Association of CYP2D6 Genotype With Tamoxifen-Active Metabolite Levels (cont’d)
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
Metabolite levels were determined at a
minimum of 4 weeks from the start of tamoxifen treatment.
In 5 studies, the presence of nonfunctional
variant alleles (Stearns et al. 2003; Jin et al.
2005; Borges et al. 2006) or the presence
of homozygous reduced function variant
alleles (Lim et al. 2007; Kiyotani et al. 2010)
was associated with significantly reduced
endoxifen levels.
Although mean endoxifen levels for each genotype group differed significantly, individual
patient data points for all groups overlapped
considerably in 2 studies that reported these
data (Jin et al. 2005; Borges et al. 2006).
Four studies also reported 4-OH tamoxifen
levels by CYP2D6 genotype, finding nonsignificant differences in one (Jin et al. 2005) and significant differences in 3 (Lim et al. 2007; Xu et
al. 2008; Kiyotani et al. 2010). Because the production of 4-OH tamoxifen can be effected by
several different CYP450 enzymes, it is unclear
whether major differences by CYP2D6 genotype
are expected.
When examined, coadministration of a potent
CYP2D6 inhibitor medication (the SSRIs paroxetine or fluoxetine) to homozygous wild-type
(EM) patients resulted in endoxifen levels close
to those of homozygous nonfunctional genotype (PM) patients (Jin et al. 2005; Borges et
al. 2006). Coadministration of a weak CYP2D6
inhibitor (various, e.g., the SSRIs sertraline or
citalopram; celecoxib) was not associated with
decreased endoxifen in 1 study (Borges et al.
2006). Thus, it is important to consider the use
of potent CYP2D6 inhibitors when assigning
metabolizer status to patients in studies of the
association of CYP2D6 genotype with clinical
outcome.* When inhibitor medication is not
taken into account in comparing endoxifen
levels by genotype, the results represent the
weakest case, because some genotypic EMs
may be functional PMs.
Summary. Five prospective cohort studies of
adjuvant tamoxifen treatment provide consistent evidence that CYP2D6 nonfunctional
variant alleles are associated with significantly
reduced plasma endoxifen levels. However,
endoxifen levels overlap across all genotypes,
suggesting that CYP2D6 genetic variability
does not explain all variability in endoxifen
levels. Somewhat surprisingly, because generation of 4-OH tamoxifen does not depend
predominantly on CYP2D6 as does generation of endoxifen, 3 of 4 studies report a significant association of low CYP2D6 function
with reduced plasma 4-OH tamoxifen levels.
Coadministration of a potent CYP2D6 inhibitor to CYP2D6 homozygous wild-type patients
(EMs) is associated with endoxifen levels near
those of patients who are PMs; thus use of
CYP2D6 inhibitors should be taken into account
in assigning metabolizer status in studies of the
association of CYP2D6 genotype with clinical
outcomes.
Association of Tamoxifen-Active Metabolite
Levels With Clinical Outcomes
Two studies (Xu et al. 2008 and Kiyotani et al.
2010; see Table 10) measured 4-OH tamoxifen
and/or endoxifen levels as well as patient
outcomes and correlated CYP2D6 genotypes
with both effector metabolite levels and with
clinical outcome. Both studies were small and
were conducted in Asian populations in which
the CYP2D6*10 reduced function variant is
common. Xu et al. (2008) genotyped patients
only for the *10 variant; Kiyotani et al. (2010)
genotyped for *10 and for several other variants, but the *10 allele accounted for the vast
majority of variant alleles detected.
Xu et al. (2008) measured only 4-OH tamoxifen levels, lacking an endoxifen standard for
measurement of that metabolite. The authors
reported a significant decrease in 4-OH tamoxifen levels for patients with the *10/*10 genotype
(see Table 7) and a corresponding significant
reduction in 5-year DFS (HR for distant metastasis or death from breast cancer = 4.7; 95% CI,
1.1 to 20) relative to wild-type homo- or heterozygotes, but not in disease-specific survival,
defined as time from diagnosis to death from
breast cancer (Table 8). Kiyotani et al. reported
a significant trend for decreased endoxifen and
4-OH tamoxifen across wild type/variant and
variant/variant genotypes compared with wild
type/wild type, and reported significant and
increasing HRs for RFS for variant genotypes
(wild type/variant: HR for breast cancer recurrence [locoregional, contralateral, or distant
metastasis] = 4.4; 95% CI, 1.3 to 15.0; variant/
variant: HR = 9.5; 95% CI, 2.8 to 32.5) compared
with homozygous wild type.
* It may also be important to avoid use of potent CYP2D6 inhibitors when prescribing tamoxifen; however, this Assessment does
not address that question.
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.19
Table 8. Studies Reporting Both Association of CYP2D6 Genotype With Tamoxifen Active Metabolites and With Clinical Outcomes
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
n
Xu et al
(2008)
152
[Asian
population]
Kiyotani et
al. (2010a),
update of
Kiyotani et
al. (2008)
[Asian
population]
282
Pre- and
postmenopausal
women with newly
diagnosed breast
cancer receiving
surgery and
starting tamoxifen
monotherapy Dec
1994 to Nov 2005
–Adjuvant
– 20 mg/d
Cross-sectional
selection
of pre- and
postmenopausal
surgical patients
with ER+ invasive
breast cancer
receiving TAM
1986-2007, seen
again Sep 2007 to
Apr 2009
–Adjuvant
– 20 mg/d
– 5 y
–7.1 y
(range:
0.8–23.5)
–NR
– 5.25 y
Study
Accounted
for CYP2D6
Inhibitors?
No patients
were treated
with known
inhibitors of
CYP2D6
Gene
(Typed
Variants)
Comparison
Results
CYP2D6
(*10)
PM (n=28)
[*10/*10]
Whole blood,
fresh-frozen
tumor, or
vs.
4-OH TAM
Genotype n
wt/wt18 5.3
wt/*10
7
5.2
*10/*1012 4.1
paraffinembedded
negative
axillary LN
No patients
were treated
with SSRIs
CYP2D6
(*1 = wt;
variant
alleles V:
*4, *6, *10,
*14B, *18,
*21, *36,
*41)
Whole blood
collected Sep
2007 to Apr
2009
% Decrease p Value
2.0
22.6
0.96
0.04
EM (n=124)
[*1/*1, *1/*10]
5-Year
HRa for
HRa for
5-Year DSS
Genotype DFS, % 5-Year DFS
89
4.7 (1.1 to 20)
2.7 (0.4 to 17.3)
PM
EM 96Reference Reference
a
Adjusted for age, clinical stage, LN, tumor size, adjuvant
therapy, surgery, HER-2, and hormone receptor status. Values
are HR (95% CI).
V/V or wt/V
p Value
Genotype
Plasma Endoxifena % Decrease
wt/wt35.4
wt/V27.2
23.2
V/V 15.5
56.2<.001
vs.
wt/wt
Genotype
4-OH TAMa
wt/wt5.3
wt/V4.2
V/V 3.1
% Decrease
p Value
21
42<.001
Genotype
HR (95% CI) for RFSb
V/V
9.52 (2.79 to 32.45)
wt/V
4.44 (1.31 to 15.0)
wt/wt (Reference)1.0
a
Values are ng/mL.
b
Adjusted for tumor size and LN status
TAM, tamoxifen; 4-OH TAM, 4-hydroxytamoxifen; chemo, chemotherapy; V, variant; wt, wild type (*1 or *2); EM, extensive metabolizer; IM, intermediate metabolizer; PM, poor metabolizer;
SSRI, selective serotonin reuptake inhibitor; HR, hazard ratio; OR, odds ratio; RFS, recurrence-free survival; DSS, disease-specific survival; LN, lymph node; ER, estrogen receptor.
Technology Evaluation Center
20
Study
Study Design,
Patients
– TAM Use
– Dose and
–Duration
Before
Evaluation
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
A difficulty with the Kiyotani et al. (2010)
study is that CYP2D6 genotypes were determined from whole blood samples collected at
a time distant from breast cancer diagnosis
and surgery, such that patients selected for
this study had to have survived in some cases
for extended periods of time before sampling.
Patients who died or were lost to follow-up
could not be considered; thus patient selection is likely biased toward survivors, with
unpredictable effects on the genotype-clinical
outcome association (Lash 2010). In the Xu
et al. (2008) study, 25 of 152 included patients
were genotyped from blood samples. It is
unclear when blood was drawn, but if also at a
time distant from surgery, similar survivorship
bias may apply.
Summary. Two studies report on the relationship between CYP2D6 genotype and active
tamoxifen metabolites, and between genotype
and clinical outcomes in the same patient
population. Both studies enrolled patients from
Asian populations, focusing almost exclusively
on the reduced function CYP2D6*10 variant.
Both studies reported reduced endoxifen and/
or 4-OH tamoxifen concentrations in patients
homozygous or heterozygous for variant
alleles, and in conjunction reported decreasing
DFS or recurrence-free survival. One or both
studies have study design flaws likely resulting in selection bias. In addition, both studies
are small, resulting in estimates of association
with wide CIs. Thus, the relationship between
endoxifen (or 4-OH tamoxifen) plasma concentrations and clinical outcomes has not been
established in Asian populations, nor has it
been studied in white populations with null
function CYP2D6 genotypes.
Direct Evidence
Association of Genotype
with Clinical Outcomes
We included in this Assessment 23 studies (in
24 publications) evaluating the association
between CYP2D6 genotype and clinical outcomes in women treated with tamoxifen. Most
of these studies examined tamoxifen use in the
adjuvant setting in postmenopausal women.
Study details are presented in Appendix Tables
A and B. Summary tables follow.
We searched for data that evaluated CYP2D6
genotype as a prognostic marker to ensure that
a prognostic association would not confound
the effect of genotype on tamoxifen outcomes.
Four studies (Nowell et al. 2005; Schroth et
al. 2007; Wegman et al. 2005; Xu et al. 2008)
evaluated outcomes stratified by genotype for
women not treated with tamoxifen (data not
shown). Although there were limitations in
study quality or reporting, none of the studies
found that outcome varied by CYP2D6 genotype. Additionally, 3 studies (Toyama et al.
2009; Okishiro et al. 2009; Ramon Y Cajal et
al. 2010) evaluated the association between
CYP2D6 genotype and various breast cancer
prognostic factors, such as tumor size, lymph
node status, and histologic grade, but in all
cases found no significant associations (data
not shown). Due to Mendelian randomization
(natural randomization of genotype), it may be
expected that CYP2D6 genotypes are distributed
similarly across all prognostic groups (Dezentje
et al. 2009; Lash et al. 2009). Therefore, this
Assessment assumed that CYP2D6 genotype,
independent of tamoxifen treatment, was not
a confounding factor in the evaluation of the
association of CYP2D6 genotype and clinical
outcomes.
Several studies included non-tamoxifen-treated
groups (Nowell et al. 2005; Rae et al. 2012;
Regan et al. 2012; Schroth et al. 2007; Wegman
et al. 2005). However, only one directly compared tamoxifen-treated with non-tamoxifentreated patients within genotype strata
(Wegman et al. 2005). This group conducted a
retrospective study of archived tumor samples
from a randomized controlled trial of tamoxifen treatment and found that EMs (defined as
non-CYP2D6*4 carriers) treated with tamoxifen
received no statistically significant clinical
benefit compared with EMs not treated with
tamoxifen (relative risk [RR] of distant recurrence=0.9; 95% CI, 0.5 to 1.6). Paradoxically,
carriers of 1 or 2 CYP2D6*4 nonfunctional
variant alleles obtained significant benefit
from tamoxifen treatment (RR = 0.3; 95% CI,
0.1 to 0.7). This study has several limitations:
Tissue samples were available for only 226
(33%) of the originally enrolled patients; only
47 patients carried a *4 allele (of whom, only 4
were homozygous PMs) and no other variants
were tested; tamoxifen dose was 40 mg daily
instead of the recommended 20 mg daily and
was administered for only 2 years instead of
the recommended 5 years. In addition, analysis
of results did not take variable chemotherapy
use or CYP2D6 inhibitors into account, but did
adjust inappropriately for prognostic factors
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.21
Technology Evaluation Center
(i.e., tumor size, lymph node status). Because of
these limitations, study results are questionable.
In the included studies, we note the possibility
for several types of bias in estimating the
magnitude of association, defined as follows
(Lash et al. 2009):
n Survival (survivorship) bias: Most studies
are retrospective and use archived tumor
samples taken at surgery for CYP2D6 determination. Some studies, however, recall
available patients at a later date and draw
whole blood. These studies select only from
among survivors not lost to follow up.
n Overfitting: Multivariate analyses that adjust
for a large number of variables in studies
that included a relatively small number of
patients may yield an erroneously increased
estimate of association due to sparse data.
n Inappropriate confounders: As discussed
above, breast cancer prognostic factors are
neither causally related to CYP2D6 genotype
nor are they surrogates for genotype (or
for the actual exposures of interest, plasma
endoxifen and 4-OH tamoxifen levels).
Thus, they do not satisfy criteria for confounding. Adjustment for factors not considered confounders of the association of
interest (the genotype-outcome association) may introduce bias or reduce precision of effect estimates (Lash et al. 2009).
When provided, univariate Cox proportional
hazards or Kaplan-Meier/log-rank analyses
(rather than multivariate analyses) were
preferentially abstracted in data tables to
avoid potential bias.
n Misclassification: Classification of similar
genotypes into different metabolizer status
categories may erroneously reduce estimates
of the strength of association. Similarly, different genotypes grouped together (e.g., IMs
plus EMs) may bias results toward the null.
As mentioned, in most studies, CYP2D6 germline genotype was determined from archived
tumor tissue. To determine if CYP2D6 genotype
sequences are changed in tumor samples compared with normal tissue samples, Toyama et
al. (2009) confirmed that CYP2D6*10 genotype
frequencies from peripheral whole blood and
from fresh-frozen tumors were the same in
50 patients. Similarly, Lash et al. (2011) found
100% concordance for CYP2D6*4 genotyping of
DNA extracted from 106 normal tissue samples
compared with DNA extracted from paired
tumor samples. Rae et al. (2013) analyzed white
blood cell CYP2D6 genotypes (alleles *1, *2, *3,
22
*4, *6, *10, and *41) and reported concordance
of 94% and 93% with DNA extracted from
matched, formalin-fixed, paraffin-embedded
tumor tissue and unaffected lymph nodes,
respectively, in 122 breast cancer patients.
Several studies enrolled patients of Asian ethnicity, in which nonfunctional CYP2D6 variants
are rare, but the CYP2D6*10 variant is very
common (see Background section and Table 2).
CYP2D6*10 produces an unstable enzyme
resulting in decreased activity. Studies of the
metabolic activity of CYP2D6 in patients with
the *10/*10 genotype administered an enzymespecific probe, compared with patients with
the *1/*1 genotype, confirm reduced activity,
but also show overlap between the 2 groups
(Myrand et al. 2008). Nevertheless, it has been
hypothesized that breast cancer patients with
the CYP2D6*10 genotype may have worse
outcomes after tamoxifen treatment compared with patients with wild-type sequences.
Because of the differing CYP2D6 genotype
variants by ethnicity, studies enrolling patients
of Asian descent will be considered together
as a group for ease of review, followed by
review of all other studies.
Studies of Patients of Asian Ethnicity and
CYP2D6*10 Major Variant
Seven retrospective cohort studies reported the
association of CYP2D6 genotype or phenotype
with clinical outcomes in Asian populations
(see Table 9 for summary, and Appendix Table
A for details). Six studied patients who were
administered tamoxifen as adjuvant therapy
after surgery; the other study selected patients
with metastatic breast cancer (Lim et al. 2007).
All studies included a mix of pre- and postmenopausal patients.
Four of the 7 studies reported a significant
reduction for *10 homozygotes in RFS or DFS
or time to progression odds ratio compared
with wild-type homozygotes (Lim et al. 2007;
Kiyotani et al. 2008; Kiyotani et al. 2010a; Xu et
al. 2008). CYP2D6 heterozygotes (wt/V) were
usually combined with wt/wt, but in 1 study
were separately significantly associated with
reduced RFS (Kiyotani et al. 2010a). This last
study included women taking only tamoxifen,
without concomitant chemotherapy. In a companion study, Kiyotani et al. (2010b) studied a
similar population of women who differed only
in that their treatment also included chemotherapy; in this group, no significant association between CYP2D6 genotype and clinical
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
Table 9. Clinical Validity Studies of Asian Populations and CYP2D6*10 Variant (Significant Results Highlighted, With Detail in Appendix Table A)
n
Lim (2007
21
1.6
x
Kiyotani
(2008)
67
10
x
x
Xu (2008)
152
5.25
x
Toyama
(2009)
154
7.9
Yes
Study
Median
Follow
Up, y
No
Misclassification
Inappropriate
Overfittinga
Survival Bias
Potential for Bias
Comparator
Referent
Genotype
n
Genotype
n
*10/*10
12
wt/wt +
wt/*10
9
*10/*10
15
wt/wt +
wt/*10
43
x
*10/*10
28
wt/wt +
wt/*10
NR
*10/*10
28
wt/wt
x
Other Outcome
HR for RFS/DFS
(95% CI)
Type
HR (95% CI)
TTP
3.7 (1.3 to 10.7)
8.7 (1.1 to 71.1)
124
4.7 (1.1 to 20.0)
DSS
2.7 (0.4 to 17.3)
64
NR
OS
NR
(NS KM log-rank) (NS KM log-rank)
wt/*10
62
wt/wt
64
NR
(NS)
Okishiro
(2009)
173
4.7
x
*10/*10
Kiyotani
(2010a)
(no chemo)
282
7.1
x
x
x
V/V
Kiyotani
(2010b)
(plus chemo)
167
NR
x
NR
49
wt/wt +
wt/*10
133
NR
63
wt/wt
83
9.5 (2.8 to 32.4)
wt/V
136
wt/wt
83
4.4 (1.3 to 15.0)
V/V
63
wt/wt
83
0.64 (0.2 to 1.99)
wt/V
136
wt/wt
83
1.1 (0.5 to 2.3)
(NS KM log-rank)
wt, wild type; V, variant; HR, hazard ratio; CI, confidence interval; RFS, recurrence-free survival; DFS, disease-free survival; DSS, disease-specific survival; OS overall survival; TTP, time to (disease) progression;
NR, not reported; NS, not significant; KM, Kaplan-Meier
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.23
CYP2D6
Metabolizer
Status
Accounted
for CYP2D6
Inhibitors?
Technology Evaluation Center
outcome was found. Chemotherapy does not
explain results from the other 2 studies, which
did not report a significant association between
genotype and clinical outcome, given that 58%
of the patients studied by Okishiro et al. (2009)
received chemotherapy and/or goserelin, but
the patients studied by Toyama et al. (2009)
received tamoxifen alone.
Results from 5 of the 7 studies may be affected
in unpredictable ways by different types of bias,
as indicated in Table 9; these included 3 of the
4 studies reporting a significant association
between the CYP2D6*10 genotype and reduced
DFS. Two of the studies that reported no association (Toyama et al. 2009; Okishiro et al.
2009) were designed with no obvious potential
for bias.
All studies were relatively small, with numbers
of homozygous variants ranging from 12 to 63
(see Appendix Table A) and likely few events
within each group, resulting in extremely wide
CIs for all estimates (see Table 9). For example,
Goetz (2010) estimated that the Okishiro et
al. (2009; n=173) study had a power of 26% to
detect a 2-fold increase in the hazard of disease
recurrence in the *10/*10 carriers relative to
wt/wt carriers. Kiyotani et al. (2010b) reported
a statistical power of 63% to detect an HR of
5.4 in their earlier study (Kiyotani et al. 2010a),
which reported an HR of 9.4 for RFS for patients
with the lowest activity CYP2D6 allele combinations versus wild type.
All Other Studies
We included 13 studies (14 publications), primarily of white patients administered tamoxifen for adjuvant treatment of invasive breast
cancer (12 studies) and 1 study of tamoxifentreated metastatic breast cancer patients, in
this group (Table 10). Goetz et al. (2005) and
Goetz et al. (2007), the latter a reanalysis of the
earlier study, are counted as a single study. One
of the studies of patients given adjuvant tamoxifen exclusively selected patients with BRCA1
or BRCA2 mutations (Newman et al. 2008). One
study included only patients with metastatic
breast cancer treated with a higher dose of
tamoxifen (40 mg daily) than the dose used for
adjuvant treatment (20 mg daily) (Lammers et
al. 2010). Two studies assessed tamoxifen prophylaxis therapy in high-risk patients (Goetz et
al. 2011; Sestak et al. 2012).
Of the 16 studies, one is a matched case-control
nested within a breast cancer registry (Lash et
24
al. 2011), 3 are matched case-control studies
derived from completed randomized controlled
trials (Goetz et al. 2011; Goetz et al. 2013;
Sestak et al. 2012), 4 are other types of retrospective analyses of samples from completed
clinical trials (Regan et al. 2012; Rae et al. 2012;
Goetz et al. 2005/2007; Wegman et al. 2005),
one evaluated relevant patients within a population-based cohort (Bijl et al. 2009), and the
rest are cohort or consecutive series studies.
Lash et al. (2011) genotyped archived tumor
samples from a well-documented Danish breast
cancer registry that uses the same 10-year
follow-up protocol for all patients. The authors
conducted a matched, nested case-control study
in which estrogen receptor-positive patients
treated with tamoxifen, who had a recurrence,
were cases and those without a recurrence
served as the reference population. Estrogen
receptor-negative patients who were not treated
with tamoxifen also were genotyped to evaluate
any direct association between CYP2D6 inhibition (i.e., presence of 1 or 2 CYP2D6*4 nonfunctional alleles) and breast cancer recurrence.
Several safeguards were prospectively built into
the study to avoid bias, and the sample size was
chosen to achieve a statistical power of 90% to
detect an odds ratio of 1.5 associating reduced
CYP2D6 activity with risk of breast cancer
recurrence. There were no significant associations found for either homozygous or heterozygous CYP2D6*4 variant patients and breast
cancer recurrence (Table 10). Point estimates
and CIs for both CYP2D6 variant groups were
very close to those found in estrogen receptornegative patients not treated with tamoxifen. To
address potential bias introduced by genotyping
only 1 variant allele, the authors conducted a
probabilistic bias analysis (sensitivity analysis)
using comprehensive genotyping information
from another study and found little change in
results.
The largest studies reporting results for
CYP2D6 genotype and clinical outcomes were
Schroth et al. (2009; n=1345), Regan et al.
(2012; n=1243), Goetz et al. (2013; n=876),
Wegman et al. (2007; n=677), and Rae et al.
(2012; n=588).
Schroth et al. (2009) estimated that a minimum
sample size of 1279 and a minimum of 166
events would be needed to detect an HR for
breast cancer recurrence of 1.85 for PMs compared with EMs with a power of 90% and a
1-sided alpha level of 0.05. This group reported
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
10b
Schroth
et al. (2009)
1345
Regan
(2012)
Rae (2012)
6.3
x
x
xc
973
6.3
no
chemo
588
10
No
991
Misclassification
Lash et al.
(2011)
Inappropriate
Confounders
n
Overfittinga
Study
Median
Follow
Up, y
Survival Bias
Potential for Bias
Yes
CYP2D6
Metabolizer
Status
Accounted
for CYP2D6
Inhibitors?
x
x
Comparator
Referent
Other Outcome
Genotype
n
Genotype
n
HR for RFS/DFS
(95% CI)
*4/*4
71
wt/wt
607
1.0 (0.8 to 1.3)
wt/*4
313
wt/wt
607
1.4 (0.8 to 2.3)
V/V for *3,
*4, or *5
79
wt/wt
*10*10,
*41/*41 or
*10 or *41 +
*3, *4, or *5
637
V/V for *3,
*4, *6 or *7
Type
HR (95% CI)
629
TTR
2.1 (1.3 to 3.5)
wt/wt
629
TTR
1.5 (1.1 to 2.0)
86
wt/wt
610
0.6 (0.3 to 1.2)
BCFI
(KM)
0.6 (0.2 to 1.2)
IM (see
Table B)
277
wt/wt
610
0.95 (0.5 to 1.4)
BCFI
(KM)
0.95 (0.50 to 1.40)
IM-0.5
31
PM
38
2.8 (0.9 to 8.5)
IM-1
151
PM
38
1.3 (0.5 to 3.5)
IM-1.5
51
PM
38
0.8 (0.2 to 2.8)
EM
317
PM
38
1.2 (0.5 to 3.2)
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.25
Table 10. Clinical Validity Studies in Primarily White Populations (Significant Results Highlighted; Detail in Appendix Table B)
Table 10. Clinical Validity Studies in Primarily White Populations (Significant Results Highlighted; Detail in Appendix Table B) (cont’d)
Wegman
et al. (2007)
677
7.1
Goetz et al.
(2005)
223
11.4
Goetz
et al. (2007),
reanalysis of
Goetz (2005)
171
11.4
Wegman
et al. (2005)
226
10.7
x
x
Referent
Other Outcome
HR for RFS/DFS
(95% CI)
Genotype
n
Genotype
n
x
V/V for *3,
*4, or *6
NR
wt/wt
NR
OR = 2.4 (1.05 to
5.7)
V/v or V/wt,
V=*3, *4, or
*6 and
v=*10 or *41
NR
wt/wt
NR
OR = 1.7 (1.0 to 2.9)
wt/v or v/v,
v=*10 or *41
NR
wt/wt
NR
OR = 1.2 (0.6 to 2.6)
xd
*4/*4
35
wt/wt
475
Univariate NR
(K-M *4/*4 RFS =
89% vs. wt/wt 74%,
log-rank p=0.05)
x *4/*4
13
wt/wt or
wt/*4
177
*4/*4
16
wt/wt
wt/*4
40
Tam + wt/wt
Tam + wt/4*
or 4*/4*
x
x
Comparator
No
5
Yes
876
Misclassification
Goetz (2013)
Inappropriate
Confounders
n
Overfittinga
Study
Median
Follow
Up, y
Survival Bias
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
Potential for Bias
x
Type
HR (95% CI)
2.7 (1.2 to 6.4) OS
1.7 (0.8 to 3.8)
115
2.7 (1.3 to 5.4)
TTR
OS
3.2 (1.4 to 7.6)
2.0 (0.9 to 4.2)
wt/wt
115
1.6 (0.95 to 2.8)
TTR
OS
1.4 (0.7 to 3.0)
1.4 (0.8 to 2.4)
52
Tam– wt/wt
55
0.9 (0.5 to 1.6)
24
Tam– wt/4*
or 4*/4*
23
0.3 (0.1 to 0.7)
Technology Evaluation Center
26
CYP2D6
Metabolizer
Status
Accounted
for CYP2D6
Inhibitors?
5.4
x
Comparator
Referent
Other Outcome
No
162
Misclassification
Nowell
et al. (2005)
Inappropriate
Confounders
n
Overfittinga
Study
Median
Follow
Up, y
Survival Bias
Potential for Bias
Yes
CYP2D6
Metabolizer
Status
Accounted
for CYP2D6
Inhibitors?
Genotype
n
Genotype
n
HR for RFS/DFS
(95% CI)
Type
HR (95% CI)
x
*4/*4 or
wt/*4
49
wt/wt
112
0.7 (0.3 to 1.4)
OS
0.8 (0.3 to 1.8)
*1/*4
29
*1/*1
52
DSS
OS
1.9 (0.9 to 3.9)
1.5 (0.8 to 2.8)
V/V,
wt/*4 or
wt/*5
79
wt/wt,
wt/*10 or
wt/*41
118
1.9 (1.1 to 3.3)
TTR
OS
2.2 (1.2 to 4.3)
NR (NS)
2.1 (0.8 to 5.4)
OS
2.5 (0.8 to 8.2)
DSS
OS
4.1 (1.1 to 15.9)
1.9 (0.6 to 4.6)
Schroth
et al. (2007)
206
5.9
x
Newman
et al. (2008),
with
BRCA
patients
115
10
x
x
V/V for *3,
*4, or *5
12
wt/wt/ or
wt/*3, *4,
*5, or *41
96
Bijl et al.
(2009)
85
10
x
x
*4/*4
4
*1/*1
52
Ramon y
Cajal et al.
(2010)
91
10
*4/*4,
*4/*41,
*1/*5, *2/*5
16
NR
(mean DFS = 95,
p=0.016)
All others
75
NR
(mean DFS = 119)
x
x
x
x
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.27
Table 10. Clinical Validity Studies in Primarily White Populations (Significant Results Highlighted; Detail in Appendix Table B) (cont’d)
NR
Goetz et al.
(2011)
913
5
x
x
x
x
No
102
Yes
Lammers
et al. (2010)
Misclassification
n
Inappropriate
Confounders
Study
Median
Follow
Up, y
Overfittinga
Potential for Bias
Survival Bias
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6
Metabolizer
Status
Accounted
for CYP2D6
Inhibitors?
Comparator
Referent
Other Outcome
Genotype
n
Genotype
n
V/V for *3,
*4, or *6
13
wt/wt + IM
(see Table
B)
48
V/V
V/v
V/wt
V = *3, *4,
*5, *6
v = *10, *17,
*41
48
52
292
wt/wt
wt/wt
wt/wt
HR for RFS/DFS
(95% CI)
Type
TTR
372
372
372
HR (95% CI)
Univariate NR
(K-M V/V 1.4 y, 0.7
to 2.2, vs. wt/wt +
IM 1.8, 1.4 to 2.3,
p=0.089)
OS
(K-M V/V 5.0 y, 4.1
to 5.9, vs. wt/wt+
IM 7.9, 6.2 to 9.5,
p=0.012)
BCI
OR = 0.9 (0.5 to 1.7)
OR = 1.5 (0.8 to 2.7)
OR = 0.9 (0.7 to 1.3)
Technology Evaluation Center
28
Table 10. Clinical Validity Studies in Primarily White Populations (Significant Results Highlighted; Detail in Appendix Table B) (cont’d)
8
x
x
No
265
Misclassification
Sestak et al.
(2012)
Inappropriate
Confounders
n
Overfittinga
Study
Median
Follow
Up, y
Survival Bias
Potential for Bias
Yes
CYP2D6
Metabolizer
Status
Accounted
for CYP2D6
Inhibitors?
Comparator
Referent
Other Outcome
Genotype
n
Genotype
n
PM=V/V
IM=V/v or v/v
PM + IM
18
29
47
wt/wt
wt/wt
wt/wt
218
218
218
HR for RFS/DFS
(95% CI)
Type
HR (95% CI)
BCI
OR = 1.0 (0.3 to 3.3)
OR = 0.8 (0.3 to 2.2)
OR = 1.1 (0.5 to 2.3)
V=
nonfunctional
allele
V = reduced
function allele
(AmpliChipe
CYP2D6 array)
BCI: breast cancer incidence; wt: wild type; V: variant; IM: intermediate metabolizer; :HR: hazard ratio; CI: confidence interval; RFS: recurrence-free survival; DFS: disease-free survival; DSS: disease-specific survival;
OS: overall survival; TTP: time to (disease) progression; TTR: time to (disease) recurrence; RR: relative risk; NR: not reported; KM: Kaplan-Meier; chemo: chemotherapy; TAM: tamoxifen; BCFI: breast cancer-free interval;
OR: odds ratio
a
Authors report that selection bias (other than survival bias) could not be excluded.
b
Case-control study from a Danish national registry (the Danish Breast Cancer Cooperative Group) in which nearly all newly diagnosed breast cancer incident cases have been reported and registered since 1978
(Blichert-Toft et al. 2008). Per Lash et al. (2011), the same 10-year follow-up protocol was used for all patients.
c
CYP2D6 genotype determined from whole blood for 601 patients; unclear from publications when blood was drawn in relation to diagnosis/surgery. In addition, results from 22% of patients have been reported
in previous studies.
d
Authors report SSRIs were “rarely used” in the patient population.
e
AmpliChip detects CYP2D6 alleles *1 to *10AB, *11, *14A, *14b, *15, *17, *19, *20, *25, *26, *29 to *31, *35, *36, *40, *41, *1xN, *2xN *4xN, *10xN, *17xN, *35xN, and *41xN.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.29
Table 10. Clinical Validity Studies in Primarily White Populations (Significant Results Highlighted; Detail in Appendix Table B) (cont’d)
Technology Evaluation Center
significant HR estimates for time to breast
cancer recurrence for patients with CYP2D6
homozygous nonfunctional alleles (PMs) and
for patients with CYP2D6 reduced function
(IMs) compared with patients with fully functional alleles (EMs). Several CYP2D6 allele
variants were tested, and the study sample
exceeded the number recommended by the
power analysis. The unadjusted HR for time to
recurrence for CP2D6 nonfunctional homozygotes (PMs) compared with fully functional wild
types (EMs) was 2.1 (95% CI, 1.3 to 3.5; Table
10). This study combined patient samples from
different sources, including some (22% of total)
for which results had already been reported,
although not as extensively genotyped. In addition, 601 of 1345 samples were genotyped from
whole blood, but the report does not clarify
when blood samples were collected relative
to disease diagnosis or surgery, suggesting the
potential for survival bias.
Regan (2012) and Rae (2012) each reported
analyses of samples from the single-agent
treatment arms of completed clinical trials, the
Breast International Group (BIG) 1-98 trial, and
Arimidex, Tamoxifen, Alone or in Combination
(ATAC) trial, respectively. These studies had the
advantage of known and consistent treatment
within each population, compared with the
samples of convenience of many other studies,
as well as large sample sizes and testing for
multiple CYP2D6 variants. In neither study
were HRs for recurrence significant or trending
toward significance, either in tamoxifen-treated
patients (Table 10) or in patients treated with
AIs, tested for comparison (Appendix Table B).
Regan et al. (2012) separately analyzed the
majority of patients who did not receive chemotherapy (Table 10) as well as those who did
(Appendix Table B), but found no difference
in results. Both studies have been criticized
for sampling tumor tissue rather than normal
tissue and for genotyping techniques that may
have increased misclassification (Pharoah et al.
2012; Stanton 2012; Brauch et al. 2013; Goetz
2013) or invalidated the results (Nakamura
et al. 2012). However, the authors calculated
that misclassification of 75% of EMs would
have been required to change the results of
the Regan et al. (2012) study and observed that
similar results across all 3 phenotype groups
in the Rae et al. (2012) study would likely be
unaffected by misclassification. Other authors
have supported the validity of the results
(Berry 2013).
30
Goetz et al. (2013) evaluated a total of 319
cases and 557 matched controls (i.e., patients
without a recurrence event) from a randomized
controlled trial comparing 5 years of tamoxifen with 2 years of tamoxifen plus 3 years of
anastrozole. Cases were matched to controls
on several variables, including variables associated with breast cancer prognosis. Because
such variables are neither causally related to
CYP2D6 genotype nor surrogates for genotype,
their use may introduce bias into comparative
analyses. Exact numbers of patients in each
genotypic category for comparative calculations
were not reported. Results, reported as odds
ratios rather than HRs, indicate that CYP2D6
PMs had significantly higher odds of disease
recurrence in the 5-year tamoxifen arm, but not
in the 2-year tamoxifen arm, relative to EMs.
Wegman et al. (2007) genotyped 677 samples
for only the *4 variant, but found no significant
difference in RFS between *4/*4 PMs and *4
noncarriers. Only 2 events occurred in 35 *4/*4
patients, highlighting the necessity of large
numbers for these studies.
The remaining studies evaluated much smaller
sample sizes (range, 85-223 for primary outcomes). Goetz et al. (2005; n=223) and Schroth
et al. (2007; n=486) retrospectively enrolled
relatively homogeneous populations of patients
who had been treated with tamoxifen alone
(no chemotherapy) following tumor resection. Goetz et al. (2005) only genotyped for
the CYP2D6*4 nonfunctional variant (most
common), whereas Schroth et al. (2007) also
tested for the *5 nonfunctional variant and the
*10 and *41 reduced function variants. Goetz
et al. (2007) reported a reanalysis of Goetz
et al. (2005), incorporating information on
CYP2D6 inhibitor medication use to assign
patient metabolizer status (not done in Schroth
et al. [2007]). Both studies reported significantly reduced RFS and time to recurrence for
patients homozygous for nonfunctional CYP2D6
alleles (Table 12). The analysis by Schroth et al.
(2007) was inappropriately adjusted for tumor
size and nodal status, which may have biased
results. However, these results represent a
potentially weaker effect than might otherwise
have been found, because IMs are included in
the decreased CYP2D6 activity group.
Other studies reported both significant and
not significant results for CYP2D6 genotype
association with clinical outcomes of tamoxifen
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
treatment. Difficulties include low sample
numbers in all studies; analyses adjusted for
parameters that do not meet the definition of
a confounder of the genotype-outcome association in some studies; and study designs
potentially affected by survivor bias in 2 studies
(see Table 10). One study grouped genotypes
for comparison with no apparent rationale
other than that the results became significant
(Ramon y Cajal et al. 2010). The only study
that reported a significant association between
genotype and overall survival was in the metastatic setting (Lammers et al. 2010). This was
a retrospective analysis of 102 patients with
hormone receptor-positive metastatic breast
cancer who were treated with tamoxifen 40 mg
daily. In analyses inappropriately adjusted for
age, overall survival was significantly shorter
for PMs than EMs (HR for death = 2.1; 95% CI,
1.1 to 4.1).
Two studies in the prophylactic setting found
no association between CYP2D6 genotype
and breast cancer occurrence (Goetz et al.
2011; Sestak et al. 2012). Goetz et al. (2011)
conducted a case-control study of 913 women
(93%) in the tamoxifen arms of the National
Surgical Adjuvant Breast and Bowel Project
(NSABP) P1 and P2 prevention trials. Cases
developed invasive or noninvasive (DCIS)
breast cancer after prophylactic treatment
with tamoxifen. The study had 80% power
to detect an odds ratio of 1.54 for PMs versus
EMs. Sestak et al. (2012) assessed 265 women
randomized to tamoxifen (20 mg daily) in
the International Breast Cancer Intervention
Study I (IBIS-I). IBIS-I compared 5 years of
prophylactic tamoxifen therapy with placebo
in high-risk women. Cases were women who
developed estrogen receptor-positive invasive
breast cancer. The study had 74% power to
detect a 2-fold increase in cancer incidence for
combined PMs and IMs compared with EMs.
Both studies matched controls on prognostic
variables, such as age and 5-year predicted
breast cancer risk, which may have introduced
selection bias.
Reviewing all studies, no pattern for potential
bias and significant versus nonsignificant
results is apparent. Additionally, not all studies
accounted for CYP2D6 inhibitor use in assigning metabolizer status, which would be most
important for nonsignificant studies, as it would
bias toward the null. However, studies that
did not make this correction are not uniquely
associated with negative results. Across all
studies there was considerable variation in
alleles genotyped, from only the *4 allele in
some studies to as many as 7 specific variant
alleles or 33 using the AmpliChip. Studies also
used different schemes for assigning genotypes
to metabolizer categories (Appendix Table B).
In addition, smaller studies often grouped nonfunctional homozygotes and heterozygotes (i.e.,
PMs and IMs) together in order to have sufficient numbers for comparison, which would
also bias toward the null. However, the largest
studies that compared only homogeneous nonfunctional genotypes with fully functional genotypes (the most extreme comparison) reported
predominately negative results.
Summary. Twenty-three studies (in 24 publications) assessed the association between
CYP2D6 genotype and clinical outcomes in
women treated with tamoxifen. Most of these
examined tamoxifen use in the adjuvant setting
in postmenopausal women. Seven small studies
in Asian populations focused on the CYP2D6*10
reduced function allele, and 5 reported significant results for the association of CYP2D6 genotype with outcomes of tamoxifen treatment;
these studies may be affected in unpredictable
ways by different types of bias. Two studies that
reported no association may have less potential for bias. Of 14 studies (in 15 publications)
that evaluated samples from primarily white
patients, 3 large studies reported no significant
association for time to recurrence. Two of these
were retrospective analyses of clinical trial
samples and were designed to minimize the
potential for bias; their size allowed for comparison of homozygous nonfunctional CYP2D6
genotypes with fully functional wild-type genotypes, i.e., the most extreme comparison and
most likely to reveal a true association. Two
large positive studies introduced potential bias
due to uncertain timing of tissue (blood) collection (survival bias) and inappropriate matching
on prognostic factors. Smaller studies reported
a variety of significant and nonsignificant
results; no pattern of bias, genotyping or group
scheme, or accounting for CYP2D6 inhibitor use
(among possibilities) explains the differences
in results. There are several limitations to this
evidence:
n Most studies were insufficiently powered
to detect a significant association between
genotype and disease recurrence comparing
homozygous nonfunctional genotypes (PMs)
with fully functional wild types (EMs).
n There is much variability across studies in
the variant alleles genotyped, in genotype
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.31
Technology Evaluation Center
grouping into CYP2D6 metabolizer category,
and in how categories are grouped for
comparison in statistical analyses, making it
difficult to compare results across studies.
n In several studies, there is potential for bias,
and/or only adjusted estimates of effect
are reported from analyses incorporating
parameters that do not meet the definition
of a confounder of the genotype-outcome
association.
n Only 1 small study in patients with metastatic breast cancer reported a significant
association between genotype and overall
survival; larger studies are needed to
determine effect.
Two studies in the prophylactic setting found
no association between CYP2D6 genotype and
breast cancer incidence. The heterogeneity of
results across all studies, and clear results of
no genotype-tamoxifen treatment outcome in
2 large trials with the least apparent potential
for bias, suggest a lack of support for clinical
validity.
metabolism and lower endoxifen levels compared with genotypic wild-type EMs, and as
a direct result have poorer clinical outcomes.
This question rests on the assumption, not
yet supported by evidence, that some level
of endoxifen is sufficient and necessary for
tamoxifen efficacy, and that this level is not
achieved in genotypic and functional CYP2D6
PMs, and possibly not in some IMs. However,
because tamoxifen metabolism is complex
and CYP2D6 does not appear to account for all
variability in endoxifen levels, it is conceivable that polymorphisms in other tamoxifen
metabolic pathway enzymes may affect active
metabolite levels, and in theory direct measurement of the metabolite(s) itself might be
the better predictor of benefit from tamoxifen
treatment. However, because it takes 8 weeks
for tamoxifen metabolites to reach steady-state
concentrations, measuring metabolite levels is
not practical for clinical applications.
Discussion
More to the point is a cautionary note by Lash
et al. (2009) regarding the likely effect of
CYP2D6 nonfunctional allele products on the
pharmacodynamics of tamoxifen. The estrogen
receptor is the binding target of tamoxifen
and its metabolites, and dissociation constants
of even the more weakly binding molecules,
including tamoxifen itself, are still sufficient
to effectively block estrogen binding (Ratliff
et al. 2004). Moreover, it is estimated that at
doses used for adjuvant treatment, which are
intended to saturate the estrogen receptor,
more than 99% of estrogen receptors are bound
by tamoxifen and its metabolites (Dowsett and
Haynes 2003). Lash et al. (2009) modeled the
effect of CYP2D6 variant alleles on estrogen
receptor binding by tamoxifen and metabolites
and found negligible effect. As the authors
note, however, modeling cannot account for
many metabolic complexities, and mechanistic
data are needed “to show how the change in
metabolite-concentration profile associated
with inheriting the variant alleles reduces the
protection against recurrence conferred by
tamoxifen therapy, despite evidence to the contrary.” For example, is there a critical threshold
for endoxifen concentrations associated with
tamoxifen efficacy, and do CYP2D6 variants
reduce endoxifen concentrations below that
threshold?
The question examined in this Assessment
is whether patients with CYP2D6 gene variants that result in markedly reduced or absent
enzyme function have reduced tamoxifen
Thus it remains to examine the clinical evidence, the bulk of which addresses clinical
validity, the CYP2D6 genotype-tamoxifen treatment outcome association. As noted, study
3. In women who need endocrine therapy,
does use of CYP2D6 genotype testing
result in selection of endocrine therapy
regimens that improve health outcomes
compared with selection of endocrine
therapy regimens without testing (i.e.,
clinical utility)?
No clinical trials that provide direct evidence
of clinical utility have been conducted. Such a
trial might prospectively enroll patients who
would be prescribed endocrine therapy including tamoxifen, but who would also be eligible
for AI treatment. Patients would be randomized
to usual methods of treatment selection or to
CYP2D6 genotyping after which PMs would
receive AI treatment alone. Currently, however,
without a foundation in clinical validity, there is
no basis for considering a change in management for patients with specific genotypes to
improve outcomes (clinical utility).
Summary. Evidence for clinical utility is
currently lacking.
32
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
results were heterogeneous. Lash et al. (2009)
also suggests that the estimates of effect of
many studies are implausible, implying either
that tamoxifen therapy is more effective than
AI therapy in women with 2 functional alleles,
or that tamoxifen therapy is less effective than
placebo in women with no functional allele,
depending on the direction of the association.
Heterogeneity in effect estimates, both in size
and significance or lack thereof, is likely due
to the lack of power in most studies, and to
potential sources of bias in some. The analysis
of archived samples from 2 large completed
clinical trials was undertaken to achieve
adequate power, to evaluate CYP2D6 genotype more fully, to evaluate AI-treated control
populations in tandem, and to avoid potential
sources of bias. That the results of these studies
discovered no evidence of association between
CYP2D6 genotype and either tamoxifen- or
AI-treated patient outcomes suggests that using
the results of CYP2D6 genetic testing to influence decisions about tamoxifen treatment is
not currently warranted. Strong evidence of
clinical utility will likely be needed to change
this recommendation.
Summary of Application of the
Technology Evaluation Criteria
Based on the available evidence, the Blue Cross
and Blue Shield Medical Advisory Panel made
the following judgments about whether CYP2D6
genotyping for directing endocrine therapy
regimen selection for women at high risk for
primary breast cancer or breast cancer recurrence meets the Blue Cross and Blue Shield
Association Technology Evaluation Center
(TEC) criteria.
1. The technology must have final approval
from the appropriate governmental
regulatory bodies.
(http://molecular.roche.com/assays/Pages/
AmpliChipCYP450Test.aspx).
CYP2D6 genotyping assays are also available
as laboratory developed tests (LDT). Clinical
laboratories may develop and validate tests inhouse and market them as a laboratory service;
laboratories offering LDTs as a clinical service
must meet the general regulatory standards
of the Clinical Laboratory Improvement Act
(CLIA) and must be licensed by CLIA for highcomplexity testing.
FDA has considered updating the label for
tamoxifen (brand and generics) with information or recommendations regarding CYP2D6
genotyping and impact on tamoxifen efficacy.
On October 18, 2006, FDA held an Advisory
Committee meeting to answer specific questions regarding the evidence and recommendations for the label update. Since that Advisory
Committee meeting, AstraZeneca, the brand
name (Nolvadex®) manufacturer, has ceased
producing tamoxifen and is no longer maintaining the prescribing information. As of the
date of this Assessment, no direction has come
from FDA regarding revised labeling of generic
versions of tamoxifen to include CYP2D6 genotyping information.
2. The scientific evidence must permit
conclusions concerning the effect of the
technology on health outcomes.
There are several limitations to the overall
body of evidence, but the largest, most welldesigned studies do not support clinical validity
of the test. Without demonstrable clinical validity, there is no basis for changing management
in patients with specific genotypes to improve
outcomes (clinical utility).
3. The technology must improve
the net health outcome.
Evidence for clinical utility is currently lacking.
The Roche AmpliChip CYP450 Test is cleared
by the US Food and Drug Administration
(FDA) and is “intended to identify a patient’s
CYP2D6 and CYP2C19 genotype from genomic
DNA extracted from a whole blood specimen. Information about CYP2D6 and CYP2C19
genotype may be used as an aid to clinicians
in determining therapeutic strategy and treatment dose for therapeutics that are metabolized
by the CYP2D6 or CYP2C19 gene product”
4. The technology must be as beneficial as
any established alternatives.
Because the available evidence does not clearly
support a significant association between
CYP2D6 genotype and tamoxifen treatment
outcome, a chain of evidence supporting the
clinical utility of CYP2D6 genotyping for directing endocrine therapy regimen selection for
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.33
Technology Evaluation Center
women at high risk for or with breast cancer
cannot be constructed.
5. The improvement must be attainable
outside the investigational settings.
The use of CYP2D6 genotyping for directing endocrine therapy regimen selection for
women at high risk for or with breast cancer to
improve health outcomes has not been demonstrated in the investigational setting.
Based on the above, CYP2D6 genotyping does
not meet the TEC criteria for directing endocrine therapy regimen selection for women at
high risk for primary breast cancer or breast
cancer recurrence.
NOTICE OF PURPOSE: TEC Assessments are scientific opinions, provided solely for informational purposes. TEC Assessments
should not be construed to suggest that the Blue Cross Blue Shield Association, Kaiser Permanente Medical Care Program or the
TEC Program recommends, advocates, requires, encourages, or discourages any particular treatment, procedure, or service; any
particular course of treatment, procedure, or service; or the payment or non-payment of the technology or technologies evaluated.
CONFIDENTIAL: This document contains proprietary information that is intended solely for Blue Cross and Blue Shield Plans
and other subscribers to the TEC Program. The contents of this document are not to be provided in any manner to any other
parties without the express written consent of the Blue Cross and Blue Shield Association.
34
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
References
Adjuvant Breast Cancer Trials Collaborative Group.
(2007). Ovarian ablation or suppression in premenopausal
early breast cancer: Results from the International
Adjuvant Breast Cancer Ovarian Ablation or Suppression
Randomized Trial. J Natl Cancer Inst, 99(4):516-525.
Alfaro CL, Lam YW, Simpson J et al. (1999). CYP2D6
status of extensive metabolizers after multiple-dose
fluoxetine, fluvoxamine, paroxetine, or sertraline.
J Clin Psychopharmacol, 19(2):155-63.
Alfaro CL, Lam YW, Simpson J et al. (2000). CYP2D6
inhibition by fluoxetine, paroxetine, sertraline, and
venlafaxine in a crossover study: intraindividual variability
and plasma concentration correlations. J Clin Pharmacol,
40(1):58-66.
American Cancer Society. (2008). Breast Cancer Facts &
Figures 2011-2012. Atlanta: American Cancer Society, Inc.
Available at: http://www.cancer.org/cancer/breastcancer/
index. Last accessed August, 2013.
ATAC (Arimidex, Tamoxifen, Alone or in Combination)
Trialists’ Group, Forbes JF, Cuzick J et al. (2008). Effect
of anastrozole and tamoxifen as adjuvant treatment for
early-stage breast cancer: 100-month analysis of the ATAC
trial. Lancet Oncol, 9(1):45-53.
Aubert RE, Stanek EJ, Yao J et al. (2009). Risk of breast
cancer recurrence in women initiating tamoxifen with
CYP2D6 inhibitors. American Society of Clinical Oncology
2009 Annual Meeting, Abstract CRA508.
Bernard S, Neville KA, Nguyen AT et al. (2006).
Interethnic differences in genetic polymorphisms of
CYP2D6 in the U.S. population: clinical implications.
Oncologist, 11(2):126-35.
Berry D. (2013). CYP2D6 genotyping and the use
of tamoxifen in breast cancer. J Natl Cancer Inst,
105(17):1267-9.
Beverage JN, Sissung TM, Sion AM et al. (2007).
CYP2D6 polymorphisms and the impact on tamoxifen
therapy. J Pharm Sci, 96(9):2224-31.
Bijl MJ, van Schaik RH, Lammers LA et al. (2009). The
CYP2D6*4 polymorphism affects breast cancer survival in
tamoxifen users. Breast Cancer Res Treat, 118(1):125-30.
Blichert-Toft M, Christiansen P, Mouridsen HT. (2008).
Danish Breast Cancer Cooperative Group--DBCG: history,
organization, and status of scientific achievements at
30-year anniversary. Acta Oncol, 47(4):497-505.
Bonanni B, Macis D, Maisonneuve P et al. (2006).
Polymorphism in the CYP2D6 tamoxifen-metabolizing gene
influences clinical effect but not hot flashes: data from the
Italian Tamoxifen Trial. J Clin Oncol, 24(22):3708-9
Borges S, Desta Z, Li L et al. (2006). Quantitative effect of
CYP2D6 genotype and inhibitors on tamoxifen metabolism:
implication for optimization of breast cancer treatment.
Clin Pharmacol Ther, 80(1):61-74.
Brauch H, Schroth W, Goetz MP et al. (2013). Tamoxifen
use in postmenopausal breast cancer: CYP2D6 matters.
J Clin Oncol, 31(2):176-80.
Burstein HJ, Prestrud AA, Seidenfeld J et al. (2010).
American Society of Clinical Oncology clinical practice
guideline: update on adjuvant endocrine therapy for
women with hormone receptor-positive breast cancer.
J Clin Oncol, 28(23):3784-96.
Coombes RC, Kilburn LS, Snowdon CF et al. (2007).
Survival and safety of exemestane versus tamoxifen
after 2-3 years’ tamoxifen treatment (Intergroup
Exemestane Study): a randomised controlled trial.
Lancet, 369(9561):559-70.
Desta Z, Ward BA, Soukhova NV et al. (2004).
Comprehensive evaluation of tamoxifen sequential
biotransformation by the human cytochrome P450 system
in vitro: prominent roles for CYP3A and CYP2D6.
J Pharmacol Exp Ther, 310(3):1062-75.
Dezentjé VO, Guchelaar HJ, Nortier JW, et al. (2009).
Clinical implications of CYP2D6 genotyping in tamoxifen
treatment for breast cancer. Clin Cancer Res, 15(1):15-21.
Dieudonne AS, Lambrechts D, Wildiers H et al. (2013).
Why should results from metastatic trials no longer matter
for early-stage disease? J Clin Oncol, 31(21):2753.
Dowsett M, Haynes BP. (2003). Hormonal effects of
aromatase inhibitors: focus on premenopausal effects and
interaction with tamoxifen. J Steroid Biochem Mol Biol,
86(3-5):255-63.
Fabian C, Tilzer L, Sternson L. (1981). Comparative
binding affinities of tamoxifen, 4-hydroxytamoxifen, and
desmethyltamoxifen for estrogen receptors isolated from
human breast carcinoma: correlation with blood levels
in patients with metastatic breast cancer. Biopharm Drug
Dispos, 2(4):381-90.
Goetz MP. (2013). Update on CYP2D6 and tamoxifen.
Clin Adv Hematol Oncol, 11(3):178-80.
Goetz MP, Kamal A, Ames MM. (2008). Tamoxifen
pharmacogenomics: the role of CYP2D6 as a predictor of
drug response. Clin Pharmacol Ther, 83(1):160-6.
Goetz MP, Knox SK, Suman VJ et al. (2007). The impact
of cytochrome P450 2D6 metabolism in women receiving
adjuvant tamoxifen. Breast Cancer Res Treat, 101(1):113-21.
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.35
Technology Evaluation Center
Goetz MP, Rae JM, Suman VJ et al. (2005).
Pharmacogenetics of tamoxifen biotransformation is
associated with clinical outcomes of efficacy and hot
flashes. J Clin Oncol, 23(36):9312-8.
Goetz MP, Schaid DJ, Wickerham DL et al. (2011).
Evaluation of CYP2D6 and efficacy of tamoxifen
and raloxifene in women treated for breast cancer
chemoprevention: results from the NSABP P1 and P2
clinical trials. Clin Cancer Res, 17(21):6944-51.
Jonat W, Gnant M, Boccardo F et al. (2006).
Effectiveness of switching from adjuvant tamoxifen to
anastrozole in postmenopausal women with hormonesensitive early-stage breast cancer: a meta-analysis.
Lancet Oncol, 7(12):991-6.
Kaufmann M, Jonat W, Hilfrich J et al. (2007). Improved
overall survival in postmenopausal women with early
breast cancer after anastrozole initiated after treatment
with tamoxifen compared with continued tamoxifen: the
ARNO 95 Study. J Clin Oncol, 25(19):2664-70.
Goetz M, Suman V. (2010). Genetic polymorphisms of
CYP2D6*10 and CYP2C19*2, *3 are not associated with
prognosis, endometrial thickness, or bone mineral density
in Japanese breast cancer patients treated with adjuvant
tamoxifen. Cancer, 116(4):1007.
Kiyotani K, Mushiroda T, Hosono N, et al. (2010b).
Lessons for pharmacogenomics studies: association study
between CYP2D6 genotype and tamoxifen response.
Pharmacogenet Genomics, 20(9):565-8.
Goetz MP, Suman VJ, Hoskin TL, et al. (2013). CYP2D6
metabolism and patient outcome in the Austrian Breast
and Colorectal Cancer Study Group trial (ABCSG) 8.
Clin Cancer Res, 19(2):500-7.
Kiyotani K, Mushiroda T, Imamura CK et al. (2010a).
Significant effect of polymorphisms in CYP2D6 and ABCC2
on clinical outcomes of adjuvant tamoxifen therapy for
breast cancer patients. J Clin Oncol, 28(8):1287-93.
Griese EU, Asante-Poku S, Ofori-Adjei D et al. (1999).
Analysis of the CYP2D6 gene mutations and their
consequences for enzyme function in a West African
population. Pharmacogenetics, 9(6):715-23.
Kiyotani K, Mushiroda T, Sasa M et al. (2008). Impact of
CYP2D6*10 on recurrence-free survival in breast cancer
patients receiving adjuvant tamoxifen therapy. Cancer Sci,
99(5):995-9.
Griese EU, Zanger UM, Brudermanns U et al. (1998).
Assessment of the predictive power of genotypes for
the in-vivo catalytic function of CYP2D6 in a German
population. Pharmacogenetics, 8(1):15-26.
Kubota T, Yamaura Y, Ohkawa N et al. (2000).
Frequencies of CYP2D6 mutant alleles in a normal
Japanese population and metabolic activity of
dextromethorphan O-demethylation in different CYP2D6
genotypes. Br J Clin Pharmacol, 50(1): 31–34.
Grimm SW, Dyroff MC. (1997). Inhibition of human drug
metabolizing cytochromes P450 by anastrozole, a potent
and selective inhibitor of aromatase. Drug Metab Dispos,
25:598-602.
Hartman AR, Helft P. (2007). The ethics of CYP2D6
testing for patients considering tamoxifen. Breast Cancer
Res, 9(2):103.
Heller T, Kirchheiner J, Armstrong VW et al. (2006).
AmpliChip CYP450 GeneChip: a new gene chip that allows
rapid and accurate CYP2D6 genotyping. Ther Drug Monit,
28(5):673-7.
Ingelman-Sundberg M. (2005). Genetic polymorphisms
of cytochrome P450 2D6 (CYP2D6): clinical consequences,
evolutionary aspects and functional diversity.
Pharmacogenomics J, 5(1):6-13.
Ji L, Pan S, Marti-Jaun J et al. (2002). Single-step assays
to analyze CYP2D6 gene polymorphisms in Asians: allele
frequencies and a novel *14B allele in mainland Chinese.
Clin Chem, 48(7):983-8.
Jin Y, Desta Z, Stearns V et al. (2005). CYP2D6 genotype,
antidepressant use, and tamoxifen metabolism during
adjuvant breast cancer treatment. J Natl Cancer Inst,
97(1):30-9.
Johnson MD, Zuo H, Lee KH et al. (2004).
Pharmacological characterization of 4-hydroxy-Ndesmethyl tamoxifen, a novel active metabolite of
tamoxifen. Breast Cancer Res Treat, 85(2):151-9.
36
Lam YW, Gaedigk A, Ereshefsky L et al. (2002). CYP2D6
inhibition by selective serotonin reuptake inhibitors:
analysis of achievable steady-state plasma concentrations
and the effect of ultrarapid metabolism at CYP2D6.
Pharmacotherapy, 22(8):1001-6.
Lammers LA, Mathijssen RH, van Gelder T, et al.
(2010). The impact of CYP2D6-predicted phenotype on
tamoxifen treatment outcome in patients with metastatic
breast cancer. Br J Cancer, 103(6):765-71.
Lash TL, Cronin-Fenton D, Ahern TP, et al. (2011).
CYP2D6 inhibition and breast cancer recurrence in a
population-based study in Denmark. J Natl Cancer Inst,
2011 Feb 15. [Epub ahead of print]
Lash TL, Lien EA, Sorensen HT, Hamilton-Dutoit S.
(2009). Genotype-guided tamoxifen therapy: time to pause
for reflection? Lancet Oncol, 10(8):825-33.
Lash TL, Rosenberg CL. (2010). Evidence and practice
regarding the role for CYP2D6 inhibition in decisions about
tamoxifen therapy. J Clin Oncol, 28(8):1273-5.
Lee K-H, Ward BA, Desta Z et al. (2003). Quantification
of tamoxifen and three metabolites in plasma by highperformance liquid chromatography with fluorescence
detection: application to a clinical trial. Journal of
Chromatography B, 791(1–2):245-53.
Lim HS, Ju Lee H, Seok Lee K et al. (2007). Clinical
implications of CYP2D6 genotypes predictive of tamoxifen
pharmacokinetics in metastatic breast cancer. J Clin Oncol,
25(25):3837-45.
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
Lim YC, Desta Z, Flockhart DA et al. (2005).
Endoxifen (4-hydroxy-N-desmethyl-tamoxifen) has
anti-estrogenic effects in breast cancer cells with potency
similar to 4-hydroxy-tamoxifen. Cancer Chemother
Pharmacol, 55(5):471-8.
Lim YC, Li L, Desta Z et al. (2006). Endoxifen, a
secondary metabolite of tamoxifen, and 4-OH-tamoxifen
induce similar changes in global gene expression patterns
in MCF-7 breast cancer cells. J Pharmacol Exp Ther,
318(2):503-12.
Lonning PE. (2007). Adjuvant endocrine treatment
of early breast cancer. Hematol Oncol Clin North Am,
21(2):223-38.
Loprinzi CL, Kugler JW, Sloan JA et al. (2000).
Venlafaxine in management of hot flashes in survivors
of breast cancer: a randomised controlled trial. Lancet,
356(9247):2059-63.
Lynch T, Price A. (2007). The effect of cytochrome P450
metabolism on drug response, interactions, and adverse
effects. Am Fam Physician, 76(3):391-6.
Newman WG, Hadfield KD, Latif A et al. (2008).
Impaired tamoxifen metabolism reduces survival
in familial breast cancer patients. Clin Cancer Res,
14(18):5913-8.
Nowell SA, Ahn J, Rae JM et al. (2005). Association of
genetic variation in tamoxifen-metabolizing enzymes with
overall survival and recurrence of disease in breast cancer
patients. Breast Cancer Res Treat, 91(3):249-58.
Okishiro M, Taguchi T, Jin Kim S et al. (2009). Genetic
polymorphisms of CYP2D6 10 and CYP2C19 2, 3 are not
associated with prognosis, endometrial thickness, or bone
mineral density in Japanese breast cancer patients treated
with adjuvant tamoxifen. Cancer, 115(5):952-61.
Pharoah PD, Abraham J, Caldas C. (2012). Re: CYP2D6
genotype and tamoxifen response in postmenopausal
women with endocrine-responsive breast cancer: the
Breast International Group 1-98 trial and re: CYP2D6 and
UGT2B7 genotype and risk of recurrence in tamoxifentreated breast cancer patients. J Natl Cancer Inst,
104(16):1263-4; author reply 6-8.
Marsh S, McLeod HL. (2007). Pharmacogenetics and
oncology treatment for breast cancer. Expert Opin
Pharmacother, 8(2):119-27.
Rae JM, Drury S, Hayes DF, et al. (2012). CYP2D6 and
UGT2B7 genotype and risk of recurrence intamoxifentreated breast cancer patients. J Natl Cancer Inst,
104(6):452-60.
Mauriac L, Keshaviah A, Debled M et al. (2007).
Predictors of early relapse in postmenopausal women with
hormone receptor-positive breast cancer in the BIG 1-98
trial. Ann Oncol, 18(5):859-67.
Rae JM, Regan MM, Thibert JN et al. (2013).
Concordance between CYP2D6 genotypes obtained from
tumor-derived and germline DNA. J Natl Cancer Inst,
105(17):1332-4.
Mortimer JE, Flatt SW, Parker BA et al. (2007).
Tamoxifen, hot flashes and recurrence in breast cancer.
Breast Cancer Res Treat, 108(3):421-6.
Ramon y Cajal T, Altes A, Pare L et al. (2010). Impact
of CYP2D6 polymorphisms in tamoxifen adjuvant breast
cancer treatment. Breast Cancer Res Treat, 119(1):33-8.
Mouridsen H, Giobbie-Hurder A, Goldhirsch A, et al
for the BIG 1-98 Collaborative Group. (2009). Letrozole
therapy alone or in sequence with tamoxifen in women
with breast cancer. N Engl J Med, 361(8):766-76.
Reddy P. (1998). A review of the newer aromatase
inhibitors in the management of metastatic breast cancer.
J Clin Pharm Ther, 23(2):81-90.
Myrand SP, Sekiguchi K, Man MZ, et al. (2008).
Pharmacokinetics/genotype associations for major
cytochrome P450 enzymes in native and first- and thirdgeneration Japanese populations: comparison with Korean,
Chinese, and Caucasian populations. Clin Pharmacol Ther,
84(3):347-61.
Nakamura Y, Ratain MJ, Cox NJ et al. (2012).
Re: CYP2D6 genotype and tamoxifen response in
postmenopausal women with endocrine-responsive breast
cancer: The Breast International Group 1-98 Trial. Journal
of the National Cancer Institute, 104(16):1264.
National Comprehensive Cancer Network (NCCN).
(2013). Clinical Practice Guidelines in Oncology™: Breast
Cancer V.3.2013. Available at http://www.nccn.org/
professionals/physician_gls/pdf/breast.pdf. Last accessed
August 2013.
National Comprehensive Cancer Network (NCCN).
(2013). Clinical Practice Guidelines in Oncology™: Breast
Cancer Risk Reduction V.1.2013. Available at http://www.
nccn.org/professionals/physician_gls/pdf/breast_risk.pdf.
Last accessed August 2013.
Regan MM, Leyland-Jones B, Bouzyk M, et al.
(2012). CYP2D6 genotype and tamoxifen response in
postmenopausal women with endocrine-responsive breast
cancer: The Breast International Group 1-98 Trial. J Natl
Cancer Inst, 104(6):441-51.
Sachse C, Brockmöller J, Bauer S et al. (1997).
Cytochrome P450 2D6 variants in a Caucasian population:
allele frequencies and phenotypic consequences.
Am J Hum Genet, 60(2):284-95.
Schroth W, Antoniadou L, Fritz P et al. (2007). Breast
cancer treatment outcome with adjuvant tamoxifen
relative to patient CYP2D6 and CYP2C19 genotypes.
J Clin Oncol, 25(33):5187-93.
Schroth W, Goetz MP, Hamann U et al. (2009).
Association between CYP2D6 polymorphisms and
outcomes among women with early stage breast cancer
treated with tamoxifen. JAMA, 302(13):1429-36.
Sestak I, Kealy R, Nikoloff M et al. (2012). Relationships
between CYP2D6 phenotype, breast cancer and hot
flushes in women at high risk of breast cancer receiving
prophylactic tamoxifen: results from the IBIS-I trial.
Br J Cancer, 107(2):230-3.
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.37
Technology Evaluation Center
Sproule BA, Otton SV, Cheung SW et al. (1997). CYP2D6
inhibition in patients treated with sertraline. J Clin
Psychopharmacol, 17(2):102-6.
Stanton V, Jr. (2012). Re: CYP2D6 genotype and tamoxifen
response in postmenopausal women with endocrineresponsive breast cancer: the Breast International
Group 1-98 trial. J Natl Cancer Inst, 104(16):1265-6;
author reply 6-8.
Stearns V, Johnson MD, Rae JM et al. (2003). Active
tamoxifen metabolite plasma concentrations after
coadministration of tamoxifen and the selective serotonin
reuptake inhibitor paroxetine. J Natl Cancer Inst,
95(23):1758-64.
Toyama T, Yamashita H, Sugiura H et al. (2009).
No association between CYP2D6*10 genotype and
survival of node-negative Japanese breast cancer patients
receiving adjuvant tamoxifen treatment. Jpn J Clin Oncol,
39(10):651-6.
Wegman P, Elingarami S, Carstensen J et al. (2007).
Genetic variants of CYP3A5, CYP2D6, SULT1A1, UGT2B15
and tamoxifen response in postmenopausal patients with
breast cancer. Breast Cancer Res, 9(1):R7.
Wegman P, Vainikka L, Stal O et al. (2005). Genotype
of metabolic enzymes and the benefit of tamoxifen in
postmenopausal breast cancer patients. Breast Cancer Res,
7(3):R284-90.
Xu Y, Sun Y, Yao L et al. (2008). Association between
CYP2D6 *10 genotype and survival of breast cancer
patients receiving tamoxifen treatment. Ann Oncol,
19(8):1423-9.
Young D. (2006). Genetics examined in tamoxifen’s
effectiveness: recurrence warning urged for labeling.
Am J Health Syst Pharm, 63(23):2286, 2296.
38
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
Appendix
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.39
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
– TAM Use
–Dose
Study
n
Lim et al.
(2007)
AHRQ
21
Study Design,
Patients
Retrospective
cohort
Korean pre- and
postmenopausal
women with
metastatic breast
cancer taking TAM
+ prior chemo and/
or AI
Kiyotani
et al. (2008)
67
Pre- and
postmenopausal
patients with ER+
invasive breast
cancer. Crosssectional selection
from surgical
patients receiving
tamoxifen monotherapy 1986 2006, seen again
Sep–Nov 2007
–Duration
–Median
Patient
Follow-Up
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
–Treatment
of
metastatic
disease
– 20 mg/d
–Median
Yes
CYP2D6
(*5,*10)
Comparison, n
Results
*10/*10 (12)
vs. wt/wt
or wt/*10 (n=9)
Patients experiencing complete or partial response, or stable
disease at ≥24 mo:
wt/wt or wt/*10 patients: 9 of 9 (100%)
6 of 12 (50%)
*10/*10 patients:
Median time to progression:
wt/wt or wt/*10 patients: 21.8 mo
5.03 (p=0.003)
*10/*10 patients:
Whole blood
collected at
follow-up
9 (range,
2–23+) mo
–Median
1.6 (range,
0.6–4.5) y
–Adjuvant
– 20 mg/d
– 5 y
– 10 y
Univariate HR for time to progression 3.69 ( 95% CI, 1.28 to
10.67)
No patients
were treated
with SSRIs
CYP2D6
(*4, *5, *6,
*10, *14,
*18, *21,
*41)
Whole blood
collected
Sep- Nov
2007
PM (n=15)
[*10/*10]
Frequency of all other tested genotypes except those listed as
PM or EM was collectively 13% and not further studied
EM (n=43)
[wt/wt, wt/*10]
PM vs. EM, OR for recurrence = 6.65 (95% CI, 1.68 to 26.4)
PM vs. EM, HR for RFS = 8.67 (95% CI, 1.06 to 71.09) in
univariate analysis
Technology Evaluation Center
40
Table A. Association of Genotype With Clinical Outcome in Patients of Asian Ethnicity and CYP2D6*10 Major Variant
Study
n
Xu et al
(2008)
152
Toyama
et al. (2009)
154
Study Design,
Patients
Pre- and
postmenopausal
women with newly
diagnosed breast
cancer receiving
surgery and
starting tamoxifen
monotherapy
Dec 1994–
Nov 2005,
Pre- and
postmenopausal
consecutive
patients
with primary
LN-negative
invasive breast
cancer receiving
tamoxifen
monotherapy
– TAM Use
–Dose
–Duration
–Median
Patient
Follow-Up
–Adjuvant
– 20 mg/d
–NR
– 5.25 y
–Adjuvant
– 20 mg/d
–3.2 (range,
2–5) y
–7.9 (range,
2.1–20.8) y
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
Comparison, n
Results
No patients
were treated
with known
inhibitors of
CYP2D6
CYP2D6
(*10)
PM (n=28)
[*10/*10]
Whole
blood,
fresh-frozen
tumor, or
paraffinembedded
negative
axillary LN
EM (n=124)
[wt/wt, wt/*10]
Multivariate analyses adjusted for age, clinical stage, LN,
tumor size, adjuvant therapy, surgery, HER-2, and hormone
receptor status; in separate analyses, CYP2D6 genotype was not
significantly associated with any of these features
CYP2D6
(*10)
*10/*10 (n=28)
vs. wt/*10 (n=62)
vs. wt/wt (n=64)
NR
Fresh-frozen
tumor tissue
Genotype
PM
EM
5-Year
DFS, %
89
96
HR for
HR for
5-Year DSSa
5-Year DFSa
4.7 (1.1 to 20) 2.7 (0.4 to 17.3)
ReferenceReference
Among 141similar patients who did not receive tamoxifen, the
CYP2D6*10 genotype was not associated with DFS or DSS
a
95% CI.
No correlation between CYP2D6*10 genotype and prognostic
features
No association between CYP2D6*10 genotype and DFS or OS by
Kaplan-Meier analysis and log-rank verification
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.41
Table A. Association of Genotype With Clinical Outcome in Patients of Asian Ethnicity and CYP2D6*10 Major Variant (cont’d)
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
– TAM Use
–Dose
Study
n
Okishiro
et al. (2009)
173
Kiyotani
et al. (2010a)
(update of
Kiyotani
et al. 2008)
282
–Duration
–Median
Patient
Follow-Up
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
Pre- and
postmenopausal
consecutive
hormone receptor
positive breast
cancer surgical
patients treated
with TAM from
1998–2004; 58%
also received
chemo and/or
goserelin
–Adjuvant
– 20 mg/d
–4.3 (range,
0.75–5) y
–4.7 (range,
Patients
receiving
SSRIs were
excluded
CYP2D6
(*10)
Cross-sectional
selection
of pre- and
postmenopausal
surgical patients
with ER+ invasive
breast cancer
receiving TAM
1986–2007, seen
again Sep 2007–
Apr 2009
–Adjuvant
– 20 mg/d
– 5 y
–7.1 (range,
0.8–23.5) y
Study Design,
Patients
0.7–9.1) y
No patients
were treated
with SSRIs
Whole
blood,
drawn at
surgery
CYP2D6
(*1 = wt;
variant
alleles V:
*4, *6, *10,
*14B, *18,
*21, *36,
*41)
Whole blood
collected
Sep 2007–
Apr 2009
Comparison, n
Results
*10/*10 (40)
vs.
wt/wt
or
wt/*10 (n=133)
None of the genotypes (wt/wt, *10/wt, or *10/*10) showed any
significant association with various prognostic features e.g.
tumor size, lymph node status, HR status, histologic grade
V/V (n=63)
or
wt/V (n=136)
Genotype
Adjusted RFS HR (95% CI)a
V/V
9.52 (2.79 to 32.45)
wt/V
4.44 (1.31 to 15.0)
wt/wt (reference)
1.0
a
Adjusted for tumor size and nodal status.
vs.
wt/wt (n=83)
RFS not significantly different between *10/*10 and other
patients (log-rank p=0.98); adjustment for well-established
prognostic factors did not change conclusions
Technology Evaluation Center
42
Table A. Association of Genotype With Clinical Outcome in Patients of Asian Ethnicity and CYP2D6*10 Major Variant (cont’d)
– TAM Use
–Dose
Study
n
Kiyotani
et al. (2010b)
167
(compare
with no
chemo, see
Kiyotani
et al. (2010 a)
Study Design,
Patients
Pre- and
postmenopausal,
hormone receptor
positive, LN +/breast cancer
patients treated
with TAM + chemo
combo therapy
–Duration
–Median
Patient
Follow-Up
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
NR
NR
CYP2D6
(*1, *4, *5,
*10, *21,
*36, *41)
Sample NR
Comparison, n
Results
V/V: *5/*5, *10,
*21; *10/*10, *21;
*21/*41 (n=33)
Genotype
V/V
wt/V
wt/wt (reference)
wt/V: *1/*4, *5,
*10, *36, *41
(n=85)
wt/wt: *1/*1
(n=49)
Adjusted RFS HR (95% CI)a
0.64 (0.2 to 1.99)
1.05 (0.48 to 2.27)
1.0
No significant association of CYP2D6 genotype with RFS in either
LN+ or LN- subgroups.
No significant association of CYP2D6 genotype with RFS in
subgroups by tumor size (<2 cm vs. >2 cm).
a
Adjusted for tumor size and nodal status.
TAM, tamoxifen; chemo, chemotherapy; AI, aromatase inhibitor; V, variant; wt, wild type (*1 or *2); EM, extensive metabolizer; IM, intermediate metabolizer; PM, poor metabolizer; HR, hazard ratio; OR, odds ratio;
TTR, time to recurrence; RFS, recurrence-free survival; OS, overall survival; DSS, disease-specific survival; NS, not significant; SSRI, selective serotonin reuptake inhibitor; LN, lymph node; PVI, peritumoral vascular invasion;
NR, not reported; CI, confidence interval
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.43
Table A. Association of Genotype With Clinical Outcome in Patients of Asian Ethnicity and CYP2D6*10 Major Variant (cont’d)
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
– TAM Use
–Dose
Study
n
Lash et al.
(2011)
991
Study Design,
Patients
Case-control study
from a Danish
national registry
of women with
stage I (2%), II
(46%), or III (52%)
breast cancer
(ER+, TAM-treated
subgroup only)
who were age
<70 years at the
time of diagnosis;
cases had local
or distant breast
cancer recurrence
or contralateral
breast cancer
occurrence during
follow-up; 94%
postmenopausal
–Duration
–Median
Follow-Up
–Adjuvant
–Tam
dose not
reported
–46% at 1 y,
18% at 2 y,
36% at 5 y
–Median
followup not
reported;
registry
follows
patients for
10 y; 66%
enrolled
>10 y
before data
cutoff
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
Yes,
prescription
for SSRI or
other CYP2D6
inhibitor
recorded in
35%; results
unchanged
when
analyzed
by CYP2D6
phenotype
CYP2D6*4
Formalinfixed, paraffin
embedded
tumor;
substudy in
106 patients
showed
perfect
concordance
between DNA
extracted
from LN and
DNA from
tumor tissue
Comparison, n
Results
PM (n=71)
[homozygous
*4]
Logistic regression adjusted for time to recurrence (cases)
or selection (controls), menopausal status, stage, receipt of
chemotherapy, receipt of radiation therapy, and type of surgery
(breast-conserving or mastectomy).
IM (n=313)
[wt/*4]
EM (n=607)
[wt/wt]
Genotype
PM
IM
EM (Reference)
TTR HR (95% CI)
1.4 (0.8 to 2.3)
1.0 (0.8 to 1.3)
1.0
Technology Evaluation Center
44
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations
– TAM Use
–Dose
Study
n
Schroth
et al. (2009)
1,345
Study Design,
Patients
Consecutively
collected
retrospective
German primary
breast cancer
cohort (stage I, II,
or III) treated with
adjuvant tamoxifen
monotherapy, and
patients from the
tamoxifen-only
arm of the NCCTG
89-30-52 trial (see
Goetz 2005); 95.4%
postmenopausal
–Duration
–Median
Follow-Up
–Adjuvant
– 20 mg/d
– 5 y
– 6.3 y
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
No,
information
on
comedication
not available
CYP2D6
(*3, *4, *5,
*10, *41)
Whole
blood
(n=601),
fresh- frozen
tumor
(n=101), or
formalinfixed, paraffin
embedded
tumor (n=659)
Comparison, n
Results
PM (n=79)
[homozygous
or compound
heterozygous
for *3, *4, or
*5]
Recurrence
All-Cause
Genotype
Rate (%)
Mortality Rate (%)
PM29.022.8
IM20.918.0
EM14.916.7
IM (n=637)
[homozygous
*10 or *41 or
either variant +
a PM allele]
EM (n=629)
Absence of PM
or IM alleles
Genotype
PM
IM
EM (Reference)
a
vs. EM
TTR HR (95% CI)a
2.1 (1.3 to 3.5)
1.5 (1.1 to 2.0)
1.0
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.45
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
– TAM Use
–Dose
Study
n
Regan et al.
(2012)
1,243
Study Design,
Patients
Postmenopausal
patients with ER+
breast cancer
from the Breast
Inter-national
Group (BIG) 1-98
trial who received
tamoxifen singleagent therapy; 98%
white; 57% nodenegative; 77% no
chemotherapy
–Duration
–Median
Follow-Up
–Adjuvant
– 20 mg/d
– 5 y
– 6.3 y
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
No
CYP2D6
(*3, *4, *6,
*7, *41)
DNA
isolated
from freshfrozen
paraffinembedded
tumor tissue
Comparison, n
Results
PM (n=112)
homozygous
or compound
heterozygous for
*3, *4, *6, or *7
Adjusted for LN, tumor size and grade, HER2, Ki-67, PVI, race, local
therapy:
IM (n=354)
homozygous for
*10, *17, or*41
or heterozygous
for reduced and
null function
alleles, or
heterozygous for
1 reduced or null
function allele
Geno-Tamoxifen
[Letrozole]
n
EventsHR (95% CI)
type n EventsHR (95% CI)
No chemotherapy
PM
86 8
.58 (.28 to 1.21) 99 11
.95 (.50 to 1.80)
IM
277 40
.95 (.50 to 1.40) 296 37
1.02 (.69 to 1.53)
EM 61075
1
63972
1
p value 0.35 0.98
Chemotherapy
PM 26 3
IM
77 12
EM 16737
.76 (.23 to 2.48 25
3
.57 (.29 to 1.10) 66 12
1
16923
1.00 (.30 to 3.35)
1.68 (.83 to 3.39)
1
Analysis of *4 genotype alone in tamoxifen users, no chemo:
EM (n=777)
Absence of
null or reduced
function alleles
[Unclear, n=138]
Genotype
*4/*4
*4/wt
wt/wt
n
76
168
609
Events
7
22
60
HR (95% CI)
0.57 (.26 to 1.23)
1.01 (.66 to 1.56)
1
p=0.34
Technology Evaluation Center
46
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
n
Rae et al.
(2012)
588
Study Design,
Patients
Postmenopausal
patients with
breast cancer from
the Arimidex,
Tamoxifen, Alone
or in Combination
(ATAC)-UK trial
who received
tamoxifen singleagent therapy
–Adjuvant
– 20 mg/d
– 5 y
– 10 y
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
Yes; analysis
controlled
for potent
CYP2D6
inhibitors,
saw no effect
in the ~9% of
women taking
drugs known
to be potent
inhibitors
CYP2D6
variants
tested
(assigned
points
per tested
allele): *1
(1), *2 (1), *3
(0), *4 (0), *6
(0), *10 (0.5),
*41 (0.5)
Comparison, n
Results
PM = 0 points
HR for outcome: any recurrence including local recurrence and
contralateral cancer; not adjusted
IM = 0.5-1.5
EM = 2
876
Retrospective
genotyping of
postmenopausal
women with
early-stage breast
cancer in the
Austrian Breast
and Colorectal
Study 8
randomized to
tamoxifen for 5 y
or tamoxifen for 2
y plus anastrazole
for 3 y
–Adjuvant
– 20 mg/d
–5 y or
2 y plus
anastrazole
for 3 y
– 5 y
No
CYP2D6 (*3,
*4, *6, *10,
*41)
Paraffinembedded
whole tissue
TamoxifenAnastrozole
HR (95% CI)
p
HR (95% CI)
p
1.00 (referent)
1.00 (referent)
2.15 (0.85-5.40) .10
1.49 (0.33-6.64) .60
0.94 (0.43-2.08) .88
1.28 (0.43-3.78) .66
0.68 (0.23-1.96) .47
0.88 (0.22-3.53) .86
0.99 (0.48-2.08) .99
1.83 (0.66-5.02) .24
Analysis of *4 genotype alone for distant recurrence in tamoxifen
users:
Paraffinembedded
tissue
samples
Goetz et al.
(2013)
Genotype
PM (0 points)
IM (0.5)
IM (1)
IM (1.5)
EM (2)
*4/*4 0.71 (0.26 to 2.00) p=0.48
*4/wt 1.18 (0.75 to 1.88) p=0.51
wt/wt1
PM (V/V)
IM (V/v, V/wt, v/v
or v/wt)
EM (wt/wt)
V = *3, *4, or *6
V = *10 or *41
OR for outcome: documented local, regional, or distant
recurrence, contralateral breast cancer, second nonbreast primary
cancer, or death from any cause
Tamoxifen for 5 Years
Genotype
OR (95% CI)
2.45 (1.05-5.73)
V/V
V/v or V/wt
1.67 (0.95-2.93)
v/v or v/wt
1.23 (0.58-2.61)
wt/wt
1.0 (referent)
Tamoxifen for 2 Years Plus Anastrozole
Genotype
OR (95% CI)
V/V
0.60 (0.15 to 2.37
V/v or V/wt
0.76 (0.43 to 1.31)
v/v or v/wt
1.02 (0.52 to 2.01)
wt/wt
1.0 (referent)
p Value
0.04
0.07
0.60
for 3 Years
p Value
0.47
0.32
0.96
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.47
Study
– TAM Use
–Dose
–Duration
–Median
Follow-Up
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
Study
n
Wegman et
al. (2007)
677
Study Design,
Patients
Retrospective
analysis of
archived samples,
Postmenopausal
ER+ breast cancer
patients with stage
II or III disease
(238 randomized to
2 vs. 5 y TAM)
– TAM Use
–Dose
–Duration
–Median
Follow-Up
–Adjuvant
–20 or
40 mg/d
– 2 or 5 y
–7.1 (range,
0.04–18) y
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
No
CYP2D6
(*4)
(SSRIs “rarely
used in study
population”)
Fresh-frozen
tumor tissue
Comparison, n
Results
*4/*4 (35)
or
wt/*4 (186)
vs.
wt/wt (475)
In analysis of all patients, *4 homozygotes (p=0.05) and
heterozygotes (p=0.04) were significantly associated with
improved RFS by K-M log-rank but results were not significant in
multivariate analysis (p=0.055).
Multivariate analysis (controlled for tumor stage and size, lymph
node status) of each randomized subgroup was not significant for
a difference in RFS, comparing *4 carriers to homozygous wt.
Note: Only 2 events occurred in 35 *4/*4 patients.
Technology Evaluation Center
48
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
Study
n
Goetz et al.
(2005)
223
Goetz et al.
2007)
(reanalysis
of Goetz
et al. 2005)
171
Study Design,
Patients
– TAM Use
–Dose
–Duration
–Median
Follow-Up
Retrospective
analysis of
archived samples
from NCCTG
RCT (89-30-52),
TAM-only arm
(no chemo);
all patients
postmenopausal
–Adjuvant
– 20 mg/d
– 5 y
–11.4
(range,
5.7–14.1) y
Retrospective
analysis of
archived samples
from NCCTG
RCT (89-30-52),
TAM-only arm
(no chemo);
all patients
postmenopausal
–Adjuvant
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
No
CYP2D6
(*4,*6)
Paraffinembedded
tissue
Comparison, n
Results
*4/*4 (13)
vs.
wt/*4 (40)
or
wt/wt (137)
Cox HR unadjusted size:
RFS: HR = 2.7, 95% CI, 1.2to 6.4 p=0.023
OS: HR = 1.73, 95% CI, 0.8to 3.8 p=0.78
Unadjusted RFS and RFS results were significant.
(no *6 alleles
detected)
– 20 mg/d
– 5 y
–Mean 11.4
(range,
5.7–14.1) y
Yes
171 of original
223 CYP2D6*4
genotyped
patients had
medication
information;
of these,
6% were
coadministered a
CYP2D6
inhibitor for
2–3 y
CYP2D6
(*4)
Paraffinembedded
tissue
PM (n=16)
[*4/*4 or *4/
wt + weak/mod.
inhibitor or wt/
wt + potent
inhibitor]
or
IM (n=40)
[*4/wt or wt/
wt + weak/mod.
inhibitor]
vs.
EM (n=115)
[wt/wt]
Cox HR (adjusted for tumor size and nodal status), PM or
IM vs. EM:
TTR: HR = 1.91; 95% CI, 1.05 to 3.45; p=0.034
RFS: HR = 1.74; 95% CI, 1.10 to 2.74; p=0.017
OS: HR = 1.34; 95% CI, 0.83 to 2.16; p=0.223
Cox unadjusted HR, PM vs. EM:
TTR: HR = 3.20; 95% CI, 1.37 to 7.55; p=0.007
RFS: HR = 2.69; 95% CI, 1.34 to 5.37; p=0.005
OS: HR = 2.00; 95% C,I 0.92 to 4.17; p=0.077
Cox unadjusted HR, IM vs. EM:
TTR: HR = 1.40; 95% CI, 0.68 to 3.05; p=0.337
RFS: HR = 1.63; 95% CI, 0.95 to 2.78; p=0.075
OS: HR = 1.40; 95% CI, 0.80 to 2.43; p=0.240
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.49
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
n
Wegman
et al. (2005)
226
Study Design,
Patients
Retrospective
analysis of
archived samples
from a TAM
RCT + chemo or
radiotherapy
–Adjuvant
– 40 mg/d
– 2 y
–10.7
(range,
0.24–18.6
y)
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
No
CYP2D6 (*4)
Fresh-frozen
tumor tissue
Comparison, n
Results
TAM treated vs.
no TAM, stratified
by*4/*4 (4)
or
*4/wt (43)
and
wt/wt (109)
RFS of ER-positive and ER-negative women ( who likely did not
receive TAM) did not differ significantly by genotype; authors
concluded that CYP2D6*4 has no prognostic value
wt/wt: *4 carrier: Postsurgery,
postmenopausal
breast cancer
patients with
positive nodes or
tumor >30 mm for
whom tissue was
available
Nowell
et al. (2005)
Retrospective
cohort, pre- and
postmenopausal
162
Primary breast
cancer patients
taking TAM,
+ chemo or
radiotherapy
175
Primary breast
cancer patients
not taking
TAM, ± chemo,
radiotherapy
Recurrence rate ratio (RR) for ER-positive patients, TAM treated vs.
no TAM, adjusted for age, tumor size, and lymph node status:
RR=0.91, 95% CI, 0.53 to 1.57, p=0.75
RR=0.28, 95% C,I 0.11 to 0.74, p=0.009
Authors reported selection bias could not be excluded.
–Adjuvant
–(Dose not
reported)
–(Duration
not
reported)
– 5.4 y
No
CYP2D6
(*3,*4,*6)
Paraffinembedded
archived
tissue
*4/*4 or wt/*4
(+TAM, 49;
-TAM, 46)
vs.
wt/wt
(+TAM, 112;
-TAM, 120)
Survival of patients with at least 1 *4 allele (*3 and *6 alleles were
rare) compared to wt/wt evaluated using Cox modeling, adjusting
for age, race, stage, ER and PR status.
TAM
RFS
OS
No TAM
RFS
OS
HR
0.67
0.77
0.69
0.79
95% CI
0.33 to 1.35
0.32 to 1.81
0.40 to 1.18
0.42 to 1.26
p Value
0.19
0.51
0.19
0.26
Technology Evaluation Center
50
Study
– TAM Use
–Dose
–Duration
–Median
Follow-Up
Study
– TAM Use
–Dose
–Duration
–Median
Follow-Up
Study
Accounted
for
CYP2D6
Inhibitors?
Retrospective
cohort analysis of
archived samples
from pre- and
postmenopausal
patients
–Adjuvant
–(Dose not
reported)
–(Duration
not
reported)
–5.9 (range,
No
206
Primary breast
cancer patients
taking TAM alone
0.7–19 y)
280
Control breast
cancer patients
not taking TAM; +
chemo
n
Schroth
et al. (2007)
Study Design,
Patients
Gene
(Typed
Variants)/
Sample
CYP2D6
(*4,*5,*10,
*41)
Formalinfixed,
paraffinembedded
tumor tissue
Comparison, n
Results
V/V (30)
or
wt/*4 or wt/*5
(49)
vs.
wt/wt or wt/*10
or wt/*41
(118)
Cox adjusted (tumor size and nodal status) HR of TAM-treated
patients with decreased vs. normal CYP2D6 activity:
TTR: HR = 2.24; 95% CI, 1.16 to 4.33; p=0.02
RFS: HR = 1.89; 95% CI, 1.10 to 3.25; p=0.02
There were no significant associations with OS.
There were no significant associations between genotype and
outcomes in patients not treated with TAM (chemo not accounted
for).
The p values of risk estimates were adjusted for multiple
comparisons
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.51
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
Study
n
Newman
et al. (2008)
115
90
Study Design,
Patients
Unrelated white
women with
complete clinical
data from a cohort
of probands
with pathogenic
BRCA1 or BRCA2
mutations at 1 UK
cancer genetics
center
+
Ethnically matched
control population
without cancer
– TAM Use
–Dose
–Duration
–Median
Follow-Up
–Adjuvant
– 20 mg/d
– >4 y
– 10 y
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
Yes
CYP2D6
(*3, *4,
*41, and *5
deletion)
Use of
CYP2D6
inhibitors
accounted for
in assigning
metabolizer
status
Baseline
germ-line
DNA
samples
extracted
from
lymphocytes
Comparison, n
Results
PM (n=12)
[homozygous
or compound
heterozygous
for *3, *4, or
*5; or wt/wt +
potent inhibitor
or wt/*3, *4, or
*5 + moderate
inhibitor]
No significant differences between allele and genotype frequencies
in control and patient populations, i.e., CYP2D6 not associated with
breast cancer in patients with BRCA mutations
IM (n=7)
[*41/*41 or
*41/*3, *4, or
*5 ]
EM (n=96)
[wt/wt/ or wt/*3,
*4, *5, or *41
and no inhibitor
Genotype
PM
EM
MedHRa for
TTR, y
Recurrenceb
HR for OSb
5.2
2.1 (0.84 to 5.4) 2.5 (0.8 to 8.2)
17.3Reference Reference
PM BRCA2
EM BRCA2
4.1
19.3
3.8 (1.0 to 14.5)
Reference
PM BRCA1
NR
1.3 (0.3 to 6.2)
EM BRCA1
NR
Reference
a
Adjusted for nodal status.
b
95% CI.
9.7 (2.3 to 41.0)
Reference
[No deaths]
Technology Evaluation Center
52
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
Study
n
Bijl et al.
(2009)
85
Ramon y
Cajal et al.
(2010)
91
Study Design,
Patients
– TAM Use
–Dose
–Duration
–Median
Follow-Up
All women in a
population-based
cohort study who
received a first
prescription of
tamoxifen from
1991–2005
–Adjuvant
– 20 mg/day
– >4 years
– 10 years
Pre- and
postmenopausal
patients diagnosed
with primary
invasive breast
cancer 1996–1998
and returned
to hospital for
follow-up in
2007; patients
treated with either
tamoxifen alone or
plus chemo
–Adjuvant
– 20 mg/d
– >4 y
– 10 y
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
Comparison, n
Results
Yes
CYP2D6
(*4)
PM (n=4)
[*4/*4]
Isolated
DNA from
baseline
sample
IM ( n= 29)
[*1/*4]
GenotypeHR a for Breast
Cancer Mortality
PM
4.1 (1.1 to 15.9)
IM
1.9 (0.9 to 3.9)
EM (reference)
1.0
CYP2D6
(33
alleles by
AmpliChipc
CYP2D6
array)
1) PM/PM, PM/
IM (n=10)
No significant correlation found between genotypes and clinical
prognostic variables
2) IM/IM, EM/
PM, EM/IM
(n=50)
Genotype
Whole blood
at follow-up
3) EM/EM, UM/
EM, UM/IM
(n=31)
CYP2D6
inhibitors
administered
in 11 patients;
analyzed
as potential
confounders
No
Information
on CYP2D6
inhibitor
medication
was
incomplete
HR a for AllCause Mortality
1.9 (0.6 to 4.6)
1.5 (0.8 to 2.8)
1.0
Adjusted for age, tamoxifen dose and duration, and calendar
time. 95% CI.
a
EM (n=52)
[*1/*1]
See study
for genotype
classifications
Mean
BetweenDFS, mo
Group p Values
1 98
2 114
31180.41
95
*4/*4, *4/*41
*1/*5, *2/*5
All others
119
0.016
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.53
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
Study
n
Lammers
et al. (2010)
102
Study Design,
Patients
All patients from a
consecutive series
treated with TAM
between 1986 and
2008 who had ER+
metastatic breast
cancer
– TAM Use
–Dose
–Duration
–Median
Follow-Up
–Nonadjuvant
– 40 mg/d
–2.8 y (range,
1.6 mo to
17 y)
– Not reported
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
Yes, charts
reviewed
for use of
CYP2D6
inhibitors,
incorporated
into
phenotype
CYP2D6
(*3, *4, *5,
*6, *10, *41)
Whole blood
collected at
follow-up
Comparison, n
Results
PM: 2 null
alleles (*3, *4,
*5, or *6) or
1 null allele
+ moderate
inhibitor or wt +
potent inhibitor
Phenotypen
PM
13
IM
38
EM
48
IM: homozygous
for reduced
function allele
(*10 or *41) or
heterozygous
for reduced/
null function
alleles or for wt/
null alleles + no
inhibitor or wt/
wt + weak-mod
inhibitor
EM: wt/wt + no
inhibitor
a
HRa for TTR
1.69 (.9 to 3.19)
0.99 (.64 to 1.55) 1 Reference
n
12
22
33
HRa for OS
2.09 (1.06 to 4.12)
0.87 (.50 to 1.50)
1 Reference
Adjusted for age. 95% CI.
Phenotype
PM
EM + IM
a
95% CI.
n
13
48
K-M Survival (95% CI), y
5.0 (4.1 to 5.9)
7.9 (6.2 to 9.5), p=0.012
Technology Evaluation Center
54
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
Study
n
Goetz et al.
(2011)
913
Study Design,
Patients
Case-control study
of 93% of women
enrolled in the
tamoxifen arms
of the National
Surgical Adjuvant
Breast and Bowel
Project (NSABP) P1
and P2 prevention
trials who were
age ≥50 y at trial
entry.
– TAM Use
–Dose
–Duration
–Median
Follow-Up
–Prophylaxis
– 20 mg/d
– 5 y
– 5 y
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
Yes, charts
reviewed for
prescription
of CYP2D6
inhibitors,
incorporated
into
phenotype
CYP2D6 (*1,
*2, *3, *4,
*5, *6, *10,
*17, *41)
Blood
samples
Comparison, n
Results
PM: 2 null
alleles (V/V, *3,
*4, *5, or *6)
or 1 null allele
+ 1 reduced
function allele
(*10, *17, *41)
Phenotype
PM
IM – wt/v
IM – v/v
IM – wt/V
IM: homozygous
for reduced
function
allele (v/v) or
heterozygous
for reduced/
null function
alleles or for wt/
null alleles +
no inhibitor or
wt/wt + weak
inhibitor
EM: wt/wt + no
inhibitor
IM – V/v
EM
n
48
132
12
OR (95% CI) for Breast Cancer Incidence
0.9 (0.5 to 1.7)
1.0 (0.7 to 1.5)
1.1 (0.3 to 3.9)
292
52
372
0.9 (0.7 to 1.3)
1.5 (0.8 to 2.7)
1 Reference
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.55
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.
Study
n
Sestak et al.
(2012)
265
Study Design,
Patients
Case-control
study of healthy,
tamoxifen-treated
women in the
International
Breast Cancer
Intervention-I
Study (IBIS-I) who
were age 35–70
y with ≥2-fold
relative risk of
breast cancer
– TAM Use
–Dose
–Duration
–Median
Follow-Up
–Prophylaxis
– 20 mg/d
– 5 y
– 8 y
Study
Accounted
for CYP2D6
Inhibitors?
Gene
(Typed
Variants)/
Sample
Yes, 25
women (12%;
5 cases and
20 controls)
who had an
extensive
metabolizer
phenotype
used a strong
CYP2D6
inhibitor
either at entry
or during
follow-up.
Results were
unchanged
when these
women were
excluded.
CYP2D6
(33 alleles by
AmpliChipc
CYP2D6
array)
Whole blood
samples
Comparison, n
Results
PM = 2 null
alleles
PhenotypeCases
(n=54)a PM
4 (7)
IM
5 (9)
PM+IM
9 (16) EM
45 (83) IM ≥ 1 reduced
function allele
but no wt allele
EM ≥ 1 wt allele
Controls
(n=211)a
14 (7)
24 (11)
ORb (95% CI)
38 (18)
173 (82) 1.1 (0.5 to 2.3)
Reference
1.0 (0.3 to 3.3)
0.8 (0.3 to 2.2)
Values are n (%).
OR for incidence of estrogen receptor positive invasive breast
cancer.
a
b
Abbreviations: TAM, tamoxifen; chemo, chemotherapy; AI, aromatase inhibitor; V, variant; wt, wild type (*1 or *2); EM, extensive metabolizer; IM, intermediate metabolizer; PM, poor metabolizer; HR, hazard ratio;
OR, odds ratio; TTR, time to recurrence; RFS, recurrence-free survival; OS, overall survival; DSS, disease-specific survival; NS, not significant; SSRI, selective serotonin reuptake inhibitor; NCCTG, North Central Cancer
Treatment Group;
LN, lymph node; PVI, peritumoral vascular invasion; RCT, randomized controlled trial; ER, estrogen receptor; PR, progesterone receptor; Ki-67, Ki-67 labeling index.
c
AmpliChip detects CYP2D6 alleles *1 to *10AB, *11, *14A, *14b, *15, *17, *19, *20, *25, *26, *29 to *31, *35, *36, *40, *41, *1xN, *2xN *4xN, *10xN, *17xN, *35xN, and *41xN.
Technology Evaluation Center
56
Table B. Association of Genotype With Clinical Outcome in Predominantly White Populations (cont’d)
CYP2D6 Pharmacogenomics of Tamoxifen Treatment
©2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.57
Technology Evaluation Center
Technology
Evaluation
Center
Blue Cross and
Blue Shield Association
225 North Michigan Avenue
Chicago, Illinois 60601-7680
www.bcbs.com/tec 
® Registered marks of the Blue Cross and Blue Shield Association, an Association of Independent Blue Cross and Blue Shield Plans
®’Registered trademark of Kaiser Permanente
© 2014 Blue Cross and Blue Shield Association. Reproduction without prior authorization is prohibited.