Despite its wide use over six decades, warfarin therapy remains challenging due its
narrow therapeutic index. The multitude of factors interacting with warfarin makes
it difficult to maintain anticoagulation within the target International Normalized
Ratio (INR) range (Ageno et al., 2012). Even within this range the dose requirements
vary as much as 20-fold between patients.
Deviations in INR control with frequent over and under-anticoagulation are common
(Chiquette et al., 1998; Chamberlain et al., 2001; Ansell et al., 2007), are associated
with poor outcomes with under-anticoagulation (increasing the risk of thrombosis)
and over-anticoagulation (increasing the risk of serious or fatal hemorrhage), demanding
that anticoagulation control be tightly regulated (Hylek and Singer, 1994; Hylek et
al., 1996, 2000; Hylek, 2003; Wittkowsky, 2004; Wittkowsky and Devine, 2004; Hylek
and Rose, 2009). These adverse outcomes have relegated warfarin to the “top 10 drugs”
for adverse drug-related hospitalizations in the US (Budnitz et al., 2007, 2011).
Between 2007 and 2009 warfarin accounted for 33% of drug-related hospitalizations
for adverse events in the US (Budnitz et al., 2011). The risk for hemorrhage is particularly
elevated when the INR exceeds four, as well as during the initial months of therapy.
Therefore it is critical to achieve a safe and effective level of anticoagulation
for patients starting warfarin.
Current guidelines for initiation of therapy provided by the American College of Chest
Physicians (ACCP) allow flexibility in selecting a starting dose of warfarin, suggesting
5–10 mg. Although the ACCP guidelines recommend lower (2.5–5 mg) doses recognizing
the influence of age, comorbidities, nutritional status, and drug interactions, these
recommendations do not tailor dosing to individual patients (Ageno et al., 2012).
Candidate Gene Studies
The recognition of genetic regulation of warfarin response has stimulated efforts
aimed at quantifying this influence. The bulk of the evidence supports the influence
of single nucleotide polymorphisms (SNPS) in two genes; Cytochrome P450 2C9 (CYP2C9;
codes for the main enzyme involved in warfarin metabolism) and Vitamin K epoxide reductase
complex1 (VKORC1; encodes the vitamin K–epoxide reductase protein, the target enzyme
of warfarin). The influence of SNPs in CYP2C9 and VKORC1 on warfarin dose has been
extensively assessed and reviewed (Wadelius et al., 2007, 2009; Limdi and Veenstra,
2008; Cavallari and Limdi, 2009; Klein et al., 2009). This evidence provided the basis
for the recent warfarin package insert update by the United States Food and Drug Administration
(FDA).
Moreover clinical algorithms that can enable dose prediction incorporating patient-specific
genetic and clinical information have been developed and are freely available. Gage
et al. (2008) have developed a dosing algorithm based on clinical and demographic
factors (body surface area, age, target INR, amiodarone use, smoker status, race,
current thrombosis) along with CYP2C9 (*2, *3, *5, and *6), VKORC1 (-1639/3673G>A),
GGCX (rs11676382), and CYP4F2 (V433M) polymorphisms. The algorithm is freely available
at www.warfarindosing.org and allows calculation of warfarin dose based on clinical
and demographic factors alone (if genotype is not available). Incorporation of novel
and potentially important genetic variants (such as CYP2C9*8) can further improve
dosing prediction in African American patients (Cavallari et al., 2010; Cavallari
and Perera, 2012).
As demonstrated by multiple studies, including the work of the International Warfarin
pharmacogenetics Consortium (IWPC), dosing based on clinical/demographic factors alone
improves prediction of stable therapeutic dose of warfarin (compared to the one-size-fits-all
5 mg/day dose), specifically in patients that need =7 mg/day or =3 mg/day. Furthermore
inclusion of CYP2C9 and VKORC1 provide a substantial gain in improvement of dose prediction
in 46% of patients (Klein et al., 2009). The www.warfarindosing.org also allows the
user to compute the estimated dose requirements based on the IWPC algorithm.
Both pharmacogenetic algorithms (Gage et al., 2008; Klein et al., 2009) require detailed
mathematical calculations to predict warfarin dose. A simpler alternative is to refer
to the genotype-stratified warfarin dose table recently added to the warfarin label
by the U.S. FDA. Although pharmacogenetic algorithms are most accurate, the genotype-stratified
warfarin dose table provides a more accurate dose prediction than empiric dosing (Finkelman
et al., 2011).
The Clinical Pharmacogenetics Implementation Consortium (CPIC) of the National Institutes
of Health Pharmacogenomics Research Network has developed guidelines to assist clinicians
in the interpretation and use of CYP2C9 and VKORC1 genotype data for estimating therapeutic
warfarin dose to achieve an INR of 2–3, should genotype results be available to the
clinician. These guidelines are published (Johnson et al., 2011) and periodically
updated based on new developments in the field (http://www.pharmgkb.org/page/cpic).
Genome-Wide Association Studies
To identify other genes/SNPs that can explain variability in warfarin dose two genome-wide
association studies (GWAS) have been conducted. Among patients of European descent
these studies have confirmed the influence of VKORC1, CYP2C9, and identified CYP4F2
as the main genes associated with dose (Cooper et al., 2008; Takeuchi et al., 2009;
Cha et al., 2010). The genome-wide significance of the latter association remains
to be confirmed. These studies suggest that identification of common variants in other
genes exhibiting influence of magnitude similar to that of CYP2C9 and VKORC1 is unlikely,
at least in Whites.
As these known variants in candidate genes account for a smaller percent of the variability
in warfarin dose among Blacks, the IWPC is leading a GWA analysis in Blacks. Preliminary
results of the ongoing GWAs meta-analysis were presented at the 2011 American Heart
Association meeting (Perera et al., 2011). A parallel effort is planned to conduct
a GWAs meta-analysis for multiple population groups (Whites, Asians, African American,
Japanese, and Middle-Eastern).
Despite these efforts a large portion (40% among White and 60% among Blacks) of the
variability in warfarin dose remains unexplained. Perhaps emerging genotyping technologies
such as exome sequencing or whole genome sequencing will reveal important clues that
can explain the missing heritability. In addition to interrogating genetic variation
ongoing and future studies must assess in detail environmental (diet, smoking alcohol,
etc.) and lifestyle (compliance, exercise, etc.) factors with similar rigor. It is
very likely that this along with gene-environment interactions holds the key to explaining
the majority of the variability in response.
One group that remains unrepresented is the Hispanic/Spanish population. This is of
great importance in the US as people of Hispanic descent accounted for more than half
the growth in the population between 2000 and 2010 and account for 16.3% of the US
population (surpassing Blacks who account for 12.3%; U.S. Census Bureau, 2010).
Pressing Challenges
Will pharmacogenetic/genomic interventions have an impact on clinically meaningful
outcomes?
Although extensive research efforts have identified several genetic markers strongly
associated with outcomes of interest and hailed them as promising tools, these proclamations
are based mainly on associations rather than their evaluation as predictors. Moreover
such evaluations must be based on clinically relevant hard-endpoints such as anticoagulation
control, hemorrhage and health-care utilization and costs (Limdi and Veenstra, 2010).
At the crux of this debate are three questions:
a)
Can a genetic risk factor (genetic marker) associated with an adverse (or beneficial)
outcome be a clinically useful predictor of that outcome? (clinical validity)
CYP2C9 and VKORC1 genotypes are clinically useful predictors of warfarin dose in clinical
trials (Anderson et al., 2007, 2012).
b)
Can incorporation of the genetic factor predict risk of the outcome more accurately
than existing clinical models? (Clinical utility)
Incorporation of CYP2C9 and VKORC1 genotypes provided superior warfarin dose prediction
compared with the clinical algorithm (or the fixed 5-mg dose algorithm; Klein et al.,
2009; Anderson et al., 2007, 2012).
The updated FDA package insert provides easy to use genotype-stratified warfarin dose
table. Although pharmacogenetic algorithms are most accurate, the genotype-stratified
warfarin dose tables provide a more accurate dose prediction than empiric dosing (Finkelman
et al., 2011).
c)
Will the outcome predicted for individuals be sufficiently different to warrant a
change in treatment decisions? (degree of clinical utility)
1)
Incorporation of CYP2C9 and VKORC1 genotypes improves prediction of warfarin doses
as confirmed by randomized clinical trials proving excellent insight into the effectiveness
of utilizing pharmacogenetics in a real-world setting (Anderson et al., 2012).
2)
Whether the benefits of pharmacogenetic guidance of warfarin dosing would translate
into improved anticoagulation control is being evaluated.
Anderson et al. showed that pharmacogenetic-dosing resulted in a higher percent time
in target range (PTTR; 69 and 71% at 1 and 3 months compared to the control group
(58 and 59% at 1 and 3 months; Anderson et al., 2012).
The improvement in PTTR achieved by pharmacogenetic-dosing (vs. standard dosing) is
greater than that achieved by specialty anticoagulation clinics (vs. usual medical
care). Importantly, these benefits in the pharmacogenetic cohort accrued in a setting
where warfarin-treated patients were typically managed by standard protocol by an
anticoagulation service/clinic.
3)
Observational cohort studies have demonstrated that possession of CYP2C9 variant allele
increases the risk of hemorrhage (Aithal et al., 1999; Margaglione et al., 2000; Higashi
et al., 2002; Limdi et al., 2008). Whether the increased risk can be mitigated and
whether the benefits of pharmacogenetic guidance of warfarin dosing would translate
into a decrease in risk of hemorrhage remains to be determined.
Not unlike randomized clinical trials conducted to prove drug efficacy and attain
FDA approval, pharmacogenetic tests/interventions are being held to higher standards.
To gain acceptance (in practice with reimbursement) the test/intervention must demonstrate
improvement in intermediate endpoints (e.g., PTTR for warfarin) and preferably hard-endpoints
(e.g., hemorrhage risk reduction) in randomized clinical trials.
When will this evidence be available? Are we there yet?
A number of clinical trials studies have assessed whether dose prediction and anticoagulation
control are superior in patients receiving pharmacogenetically guided dosing vs. standard
dosing or on clinical factors excluding genotype. However trials to date have been
limited by study design issues (not blinded, use of historic control group) and limited
sample size. Nonetheless these studies have provided valuable information on effect
sizes and genotype-specific expectations which was valuable in the design of large
ongoing randomized trials.
The Clarification of Optimal Anticoagulation through Genetics (COAG; NCT00839657)
trial is a multicenter double-blind, randomized trial aims to determine whether the
use of genetic and clinical information for selecting the dose of warfarin in 1022
patients during the initial dosing period will lead to improvement in stability of
anticoagulation at 4 weeks relative to a strategy that incorporates only clinical
information (without genetics).
The EUropean Pharmacogenetics of AntiCoagulant Therapy (EU-PACT; NCT01119300) trials
are single-blinded randomized controlled trials aiming to assess the safety and clinical
utility of genotype-guided dosing of the three main coumarins used in Europe: acenocoumarol,
phenprocoumon, and warfarin. The warfarin arm will recruit 900 (UK and Sweden) warfarin
patients to determine if genotype-guided dosing vs. standard dosing improves PTTR
at 3 months.
The Genetic InFormatics Trial (GIFT; NCT01006733) of Prevent DVT is a randomized controlled
trial aiming to determine the benefit of genotype-guided dosing vs. clinically guided
dosing among 1600 participants (age 65 years or older) in reducing the composite endpoint:
non-fatal VTE, non-fatal major hemorrhage, death, and an INR > 4.0 at 1 month.
Pharmacogenetic-dosing of warfarin: a controlled randomized trial is led by the Taiwan
Warfarin Consortium which aims to determine whether genotype-guided dosing can improve
safety (as measured by time to target INR and PTTR) of warfarin treatment in 600 participants
compared to clinically guided dosing.
Warfarin Adverse Events Reduction for Adults Receiving Genetic Testing at Therapy
Initiation (WARFARIN; NCT01305148) is a randomized blinded interventional trial where
4300 patients (age > 65 years) are to be randomized to warfarin dosing based on the
GenoSTAT test plus clinical factors, or clinical factors alone, using the warfarindosing.org
website. The primary aim is to determine if genotype-guided therapy reduces the incidence
of warfarin-related clinical events, including major hemorrhage and thromboembolic
events at 30 days and in fewer hospitalizations and/or deaths compared to clinically
guided therapy at 90 days compared to clinically guided therapy.
Additional information on the trials below and others can be found at www.clinicaltrials.gov.
These trials will provide a robust data for efficacy/effectiveness and cost effectiveness
analysis and will provide the foundation for policy development.
Although detractors claim pharmacogenetics/genomics in general has not yielded information
to justify the investment of effort and funds, progress in the genomics/genetics arena
has been maintained a rapid pace compared to other fields in medicine. For warfarin,
the first report identifying the CYP2C9
*2 polymorphism was published in 1994. With the discovery of VKORC1 in 2004 the field
burgeoned with investigations in multiple populations across the world documenting
the effect of CYP2C9 and VKORC1 on warfarin dose, anticoagulation control, and risk
of hemorrhage. Twenty years from the report identifying the CYP2C9 *2 (10 years following
the identification of VKORC1), the results of the first double-blinded randomized
clinical trial (COAG trial) testing genotype-guided dosing intervention are expected
to be available.
Although the uptake of newly available oral anticoagulants is slow, a frequently raised
question is “Will warfarin or warfarin pharmacogenetics matter in the coming years?”
The introduction of Dabigatran (DBG; October 2010), Rivaroxaban (2011), and Apixaban
(awaiting approval) is changing the landscape of anticoagulation therapy (Melnikova,
2009; Weitz et al., 2012). Although the use of DBG is increasing (0.6 million users
in the US; Johnson, 2012; Pradaxa, 2012), warfarin remains the most widely used oral
anticoagulant (25 million users) (Schirmer et al., 2010; Altman and Vidal, 2011; Cabral
et al., 2011; Wittkowsky, 2011; Tzeis and Andrikopoulos, 2012). The Practice INNovation
And CLinical Excellence (PINNACLE) registry focusing on patients with atrial fibrillation
(AF) reports among patients who received oral anticoagulation, 87.4% were treated
with warfarin while just 12.6% were prescribed one of the two new oral anticoagulants
(Cardiology, 2012).
The oral anticoagulant market is expected to exceed $ 9 billion by 2014 (Melnikova,
2009), driven by demographics of the aging population and increased incidence of cardiovascular
disease, and the uptake of newly approved agents. Although the market share of DBG
(and the newer agents) is expected to increase, its uptake is hindered by lack of
monitoring, reversibility, and expense. Almost 2 years after approval of DBG, warfarin
remains the most widely used oral anticoagulant.
Despite the approval of four CYP2C9/VKORC1 rapid throughput genotyping platforms by
the U.S. FDA over the last decade clinical implementation of genotype-guided dosing
is lagging. Although genotype-guided therapy improves dose prediction is recognized,
evidence that such an intervention will improve anticoagulation control, reduce risk
of adverse events and health-care costs is limited. Results of ongoing clinical trials
are expected to address these issues and will perhaps provide the much needed impetus
to reevaluate reimbursement for genetic testing and for wider implementation.
The challenges unique to pharmacogenomic efforts have created an intangible benefit
for science and humanity. Of note, most genetic/genomic investigations have identified
genes with small effect sizes. To enable these discoveries investigators stepping
outside the conventional paradigm of lab-based investigative efforts formed consortia
to build collaborations across laboratories, departments, institutions, countries,
and continents. Investigators within these consortia sharing a common goal, pooled
unpublished data, working with complete strangers, while maintaining enviable focus
and relentless effort to advance science. Among the consortia, the IWPC has successfully
brought together >100 investigators from >25 institutions across >10 countries providing
valuable contributions (Klein et al., 2009; Limdi et al., 2010; Perera et al., 2011)
to the pharmacogenetic literature and much needed insight to inform the design of
ongoing clinical trials in five short years.
One has to but conduct a PubMed search for GWAS for their favorite disease/phenotype
and scroll through the authors and contributors list to understand the magnitude of
such efforts, the network of collaborators created and progress made. These collaborations
will continue (beyond the single publication) to advance science and enable discoveries
beyond what can be gaged solely by investments; past, current, and future.