Clinical applications of pharmacogenomics guided warfarin dosing

Transcription

1 Int J Clin Pharm (2011) 33:10 19 DOI /s REVIEW ARTICLE Clinical applications of pharmacogenomics guided warfarin dosing Pramod Mahajan Kristin S. Meyer Geoffrey C. Wall Heidi J. Price Received: 20 August 2010 / Accepted: 11 October 2010 / Published online: 4 February 2011 Ó Springer Science+Business Media B.V Abstract Aim of the Review To assess the state of the literature concerning pharmacogenomic testing in patients requiring vitamin K antagonists, specifically warfarin. Method We conducted a literature search of MEDLINE and International Pharmaceutical Abstracts using the following words: warfarin, pharmacogenetic, and pharmacogenomic. The search results were reviewed by the authors and papers concerning pharmacogenomic testing in warfarin dosing were procured and reviewed. Additionally bibliographies of papers procured were also examined for other studies. The authors focused on clinical trials concerning the use of pharmacogenomic testing in warfarin dosing. Results Although numerous studies have demonstrated that a significant portion of warfarin dosing variability can be explained by genetic polymorphisms, few prospective studies have been conducted that examine the integration of this information in practical dosing situations. Those that have, have shown that using pharmacogenomic information improves initial dosing estimates and decreases the need for frequent clinic visits and laboratory testing. Data showing a reduction in serious bleeding events is sparse. Cost-effectiveness analyses have generally shown a small but positive effect with pharmacogenomic testing in patients receiving warfarin. Conclusion Several studies have shown that pharmacogenomic testing for P. Mahajan K. S. Meyer G. C. Wall (&) H. J. Price College of Pharmacy and Health Sciences, Drake University, 2507 University Ave, Des Moines, IA 50310, USA geoff.wall@drake.edu K. S. Meyer Iowa Veterans Home, Marshalltown, IA 50158, USA G. C. Wall Iowa Methodist Medical Center, Des Moines, IA, USA warfarin dosing is more accurate that other dosing schemes. Pharmacogenomic testing improves time to a therapeutic international normalized ratio while requiring fewer dosing adjustments. Patients who require higher or lower than usual doses seem to benefit the most. The costeffectiveness of pharmacogenomic testing as well as preventing of outcomes such as bleeding or thrombosis are not yet elucidated. Pharmacists, especially those in a community setting can play a role in this new technology by educating prescribers and patients concerning pharmacogenomic testing, and by developing and using dosing protocols that incorporate its use. Keywords Dosing Pharmacogenetics Pharmacogenomics Vitamin K antagonists Warfarin Impact of findings on practice Pharmacogenomic differences in patients is responsible for a significant amount of dosing variability with oral vitamin K antagonists such as warfarin Testing for such differences has become commercially available and data to date suggests that routine testing on patients receiving warfarin may decrease clinic visits, improve time in the therapeutic range (target international normalized ratio), and is likely to decrease the risk of major bleeding events. Pharmacists can play a major role in the implementation of this new technology by proper utilization of such testing in anticoagulation clinics, education of prescribers and patients of the implications of pharmacogenomic testing, and research assessing its costeffectiveness.

2 Int J Clin Pharm (2011) 33: Introduction For more than 40 years, vitamin K antagonists (VKAs) including warfarin, have been the gold standard for prevention and treatment of venous thromboembolism, and for prevention of stroke in atrial fibrillation. Indeed, in many countries VKAs are the only drugs available for these indications. For example, in 2004, 30 million prescriptions were written for warfarin in the United States alone [1]. Unfortunately, dosing of this drug is extremely complex. It has a very narrow therapeutic index, numerous drug and food interactions, and the dosing amount required to achieve a therapeutic international normalized ratio (INR) in individual patients can vary by a factor of 10 or more [2]. This complexity of dosing, coupled with the potentially catastrophic results of over-anticoagulation (e.g. intercranial hemorrhage) have led to the evolution of oral anticoagulation clinics designed to maximize the safety and efficacy of warfarin dosing in individual patients. Such clinics, often run by pharmacists, have been shown to be more effective than the usual care clinics at keeping patients in their INR range, and decreasing bleeding rates [3, 4]. In the last 10 years significant progress has been made toward understanding the genetic basis for warfarin dosing variability. It is now known that by itself, genetic polymorphisms are responsible for about 30% of dosing variability with this agent [5]. This, coupled with the commercial availability of genetic testing for warfarin metabolism polymorphisms, has led researchers to investigate the clinical impact of pharmacogenomic (PGx) testing as an aid to improve dosing accuracy. How will PGx testing impact warfarin dosing in oral anticoagulation clinics? What roles will the pharmacist particularly those in community settings play in the implementation and interpretation of this new technology? Aim of the review Our aim was to review the latest literature on the clinical use of PGx testing for warfarin metabolism on the appropriate dosing of this drug. Additionally, we will make recommendations on the place of pharmacists in the utilization of PGx testing. For a general overview of this topic, readers are directed to some comprehensive reviews published recently [5 8]. Space limitations required that primary investigations of this subject remain the focus of our paper. Methods The authors undertook a search of the medical literature using MEDLINE and International Pharmaceutical Abstracts with the following terms: warfarin, pharmacogenomic, pharmacogenetic, and dosing with search dates of January 1995 to June Pertinent articles published in English were retrieved. Bibliographies from articles were also searched for other articles that were also procured. Although the focus of our article is on the clinical application of PGx testing for warfarin, we also felt that a review of the basic science surrounding this issue deserved discussion as well. For simplicity we are using the term pharmacogenomics instead of pharmacogenetics throughout this paper. Results Warfarin metabolism and pharmacogenomics Warfarin (3-a-acetonylbenzyl-4-hydroxycoumarin; also known as Coumadin), a compound synthesized in 1948 and first used as a rodenticide, was approved for use as an anticoagulant in the early 1950s [9]. It exists as a recemic mixture, the S form being about 3 5 times more active in its anticoagulation activity than the R form [9]. Moreover, as summarized in Fig. 1, cellular metabolism of these two isoforms is also markedly different [10]. While CYP2C9 is the main mixed function oxidase responsible for hydroxylation of the more potent S-warfarin, CYP1A2, CYP3A4, and CYP2C19 participate in metabolism of the less active R-isoform. Additionally, a carbonyl reductase (not shown in Fig. 1) is also involved in metabolism of the R-form [9, 10]. Anticoagulant activity of warfarin is primarily attributed to its potent, stereospecific inhibitory effect on an enzymatic step essential for clot formation (Fig. 1). Activation of clotting factors catalyzed by c-glutamyl carboxylase requires the hydroquinone or reduced form of vitamin K as a cofactor. During this reaction, vitamin K is converted to the 2, 3 epoxide, which must be rapidly recycled back to the reduced form. This step is catalyzed by the enzyme vitamin K epoxide reductase complex 1 (VKORC1) and warfarin specifically inhibits this enzymatic step. Depletion of the reduced form of vitamin K and inactivation of the clotting factors is the result, eventually leading to inhibition of clot formation. Although warfarin was approved for use as an anticoagulant as early as 1950s, experimental evidence of effects of genetic polymorphism of CYP2C9 on warfarin maintenance dose requirement in patients undergoing anticoagulation therapy was published only in the 1990s [11 13]. A multitude of retrospective [14 23] and prospective [11, 24 34] studies conducted since, have now clearly established the significant influence of polymorphism in CYP2C9 [9 11, 14, 15, 23, 24, 26, 28, 29, 32] and VKORC1 [9, 10, 23, 26, 28, 32] genes on metabolism, and hence, the

3 12 Int J Clin Pharm (2011) 33:10 19 Inactive Clotting factors γ- Glutamyl Carboxylase (GGCX) Active Clotting factors [28], clotting protein FVII [22], c-glutamyl carboxylase [25], and the vitamin K epoxide reductase complex 1 like protein [19], with sensitivity to warfarin. However, additional studies would be required to conclusively establish any correlation of polymorphism in these genes with warfarin sensitivity. S- enantiomer CYP2C9 Vitamin K (Reduced) 7-Hydroxy Warfarin Vitamin K Epoxide Reductase Complex 1 (VKORC1) Warfarin Fig. 1 Warfarin: mode of action and metabolism Vitamin K (Oxidized) R- enantiomer CYP1A1 CYP1A2 CYP3A4 6-Hydroxy Warfarin 8-Hydroxy Warfarin 10-Hydroxy Warfarin anticoagulant activity of warfarin. Thus, individuals homozygous for the CYP2C9*1 allele require a higher maintenance dose compared to the CYP2C9*2 (R 144 C) and CYP2C9*3 (I 359 L) homozygotes, respectively. As expected, individuals heterozygous for *1, *2 or *3 alleles of CYP2C9 gene usually exhibit an intermediate phenotype [13, 23]. Unlike CYP2C9, initial investigations of the VKORC1 gene implicated association of several different genetic polymorphisms with variable phenotypes in warfarin metabolism. However, additional studies with larger s [20, 33, 34] have allowed researchers to define two haplotyes composed of at least six variations across the entire coding as well as non-coding regions of the VKORC1 gene. Individuals homozygous for haplotype A are characterized by a resistant phenotype, whereas individuals homozygous for haplotype B exhibit a sensitive phenotype, the heterozygous subjects showing an intermediate response. In recent years, advent of the genomic technologies has allowed researchers to evaluate the combined effects of genetic polymorphism in CYP2C9 as well as VKORC1 genes on sensitivity to warfarin [16, 17, 20, 25, 27, 31, 32, 34, 35]. Interestingly, a limited number of reports have also investigated potential association of polymorphism in the genes for CYP4F2 [36, 37], EPHX1 [37], Apolipoprotein E Clinical studies using warfarin pharmacogenomics Several researchers have reported on the differences in genetic polymorphisms and dose response among varied groups (Table 1). While it has been established that CYP2C9*2 and *3 variants can be predictive of anticoagulation response in the Caucasian population, the effect may not be as great in African-American patients [38, 39]. Warfarin doses were found to be 11% lower in Caucasian patients than African-American patients in one study that focused on CYP2C9 variances [11]. Visser and colleagues studied a large Caucasian population, confirming that patients heterozygous for either CYP2C9*2 or CYP2C9*3, as well as those with two variant alleles, require a significantly lower dose of the anticoagulant [40]. For example, there may be yet unknown differences in white populations as well. Croatian patients have a slightly different distribution pattern of CYP2C9 genotypes than other Caucasians, as demonstrated in one study [41]. In contrast to the Caucasians, CYP2C9*2 or CYP2C9*3 alleles are almost non-existent in the Asian population. Chinese and Malaysian patients were found to more commonly express the CYP2C9*1/*1 genotype (93.2%) than the CYP2C9*1/*3 genotype (6.8%) in one study. Patients with the CYP2C9*1/*1 genotype required a higher dose of warfarin whereas those with the CYP2C9*1/*3 experienced a higher rate of serious and life-threatening bleeding [42]. In one of the largest studies of the VKORC1 polymorphisms, the International Warfarin Pharmacogenetics Consortium enrolled1103 Asians, 670 blacks, and 3113 whites to study the difference in response among these ethnic populations. Although the contribution of VKORC1 genetic variation to warfarin dose prediction is greater in the white population, this was stated to be most likely due to differences in allele frequencies among the populations and that nonwhites would still benefit from PGx dosing of warfarin [23]. Considering VKORC1 polymorphism in addition to CYP2C9 status has indeed improved the accuracy of PGx algorithms. Patient age, body weight and genetic polymorphisms of VKORC1 and CYP2C9 were found to account for up to 55 and 63% of the variability in daily warfarin dose of stable Japanese and Chinese patients [43, 44]. A previously published algorithm that solely

4 Int J Clin Pharm (2011) 33: incorporated CYP2C9 only accounted for 39% of variance in maintenance warfarin dose [11]. A group of researchers in the Netherlands conducted a prospective study of 231 patients in which they assessed the effects of both CYP2C9 and VKORC1 genotypes on over- anticoagulation. They found that the VKORC1 1173CT genotype, in combination with CYP2C9*2 or CYP2C9*3, increases risk of over anticoagulation. They were the first to suggest that the probability of achieving stability correlated more with the CYP2C9 genotype, while the VKORC1 genotype was more predictive of mean daily dosage [27]. These findings were further supported by the Table 1 Recent clinical studies of warfarin pharmacogenomics: a summary Reference Study population Trial design Major findings Comments Gage et al. [11]* Voora et al. [24] Anderson et al. [61] Wen et al. [20] Lenzini et al. [45] Schwarz et al. [62] N = % white, 29% African- American N = % white, 10.4% African- American, 6.3% other Orthopedic patients N = % white in PGx group; 94.9% white in standard group N = 108 Han-Chinese N = 86 in derivation N = 179 in validation Orthopedic patients N = 297, 89.2% white, 9.8% African- American, 1% Hispanic Developed clinical and PGx algorithm (CYP2C9) to predict warfarin maintenance dose Prospective use of PGx algorithm (CYP2C9) to predict maintenance warfarin dose Prospective use of PGx algorithm (CYP2C9 and VKORC1) to predict warfarin initial dose. Evaluated efficacy of PGx algorithm (via INR values) and its safety. Also evaluated # of dose adjustments made, serious adverse clinical events, Vit K use, major bleeds, thromboembolic events. Three month study (or to end of therapy, if shorter) Prospective use of PGx algorithm (CYP2C9 and VKORC1)to predict warfarin initial dose Developed clinical and PGx algorithm (CYP2C9 and VKORC1) to refine warfarin dose after 4 days; subsequently validated in prospective Observational study of influence of CYP2C9 and VKORC1 on the time to the first INR is therapeutic, the time to first INR [4 from the date warfarin was started, the time the INR was above therapeutic range, and the INR response over time Dose was 11% lower in white patients than African-American patients. PGx algorithm found to overdose patients (by [2 mg/day) 6.5% compared to 16% who were prescribed 5 mg/day (P \ 0.001) Time to stable INR was similar between groups (14 vs. 13 days). Patients with CYP2C9 variant had more supratherapeutic INRs than wild-type group (HR = 4.6, P \ 0.01) PGx algorithm better predicted dosing in wild-type pts (requiring higher doses) and multiple variant alleles pts (requiring lower doses). PGx dosing did not improve percent of INRs out-of-range compared to standard dosing. PGx dosing decreased # of dose adjustments, and therefore # of INRs drawn Therapeutic INR reached by 83% of patients in 2 weeks. 10% of patients had INR [4. 69% of patients maintenance doses at 12 weeks matched prediction R 2 of PGx algorithm was 82% in derivation ; 70% in prospective. R 2 of clinical algorithm was 57% and 48%, respectively. Adverse events were reduced in PGx (HR 0.54; 95% CI: ) Pts with VKORC1 alleles had decreased time to first therapeutic INR (P = 0.02) and to INR [4 (P = 0.003). CYP2C9 genotype did not affect time to first INR in therapeutic range (P = 0.57) but was significant predictor of time to first INR [4 (P = 0.03) PGx algorithm only accounted for 39% of variance in maintenance warfarin dose, suggesting other factors are at play Despite use of a PGx algorithm, patients with CYP2C9 variances are still at higher risk for excessive anticoagulation and thus may need closer monitoring INR \1.8 or [3.2 constituted an out-of-range INR. Standard group used 10 mg initiation algorithm (10 mg on days 1 and 2, followed by 5 mg daily). Day 5 INR determined dose adjustments on days 5 7. INRs on day 8 and onward determined dose adjustments based on healthcare system protocol Study did not include a control group. Low number of adverse events. Results may not be applicable to other ethnic groups, due to differences in clinical factors such as BSA First published study of refinement algorithm. VKORC1 was not as significant a predictor of maintenance dose as CYP2C9 variance. Researchers hypothesize this is because VKORC1 is more related to warfarin sensitivity and therefore affects initial dose to a greater degree, whereas CYP2C9 is involved in S-warfarin metabolism Genetic variation in VKORC1, not CYP2C9, affects response to warfarin during early therapy. VKORC1 and CYP2C9 both affect the warfarin dose after the first 2 weeks

5 14 Int J Clin Pharm (2011) 33:10 19 Table 1 continued Reference Study population Trial design Major findings Comments Gage et al. [30] N = 1015 in derivation N = 292 in validation Wadelius et al. [46] N = 1496 Mostly Swedish patients, data not gathered Huang et al. [33] N = 156 Chinese patients undergoing heart valve replacement Limdi et al. [57] N = 250 African American N = 271 European American International Warfarin Pharmacogenetics Consortium [48] N = % white, 30.4% Asian, 8.7% black, 5.6% other N = 1009 in validation Li et al. [63] N = % European- American 8% Black Developed clinical and PGx algorithm (CYP2C9 and VKORC1) to predict maintenance warfarin dose; subsequently validated in prospective Prospective application of a clinical and PGx algorithm (CYP2C9 and VKORC1) to predict initial warfarin dose Observational study Prospective use of a clinical and PGx algorithm (CYP2C9 and VKORC1) to predict initial warfarin dose Observational study of influence of CYP2C9 and VKORC1 during the first 30 days of warfarin therapy Developed clinical and PGx algorithm (CYP2C9 and VKORC1) to predict initial warfarin dose; subsequently validated in prospective Prospective, observational study. Assessed contribution of genotypes versus early INRs (days 4 6) to early phases of anticoagulation. Primary outcomes were time to INR C lower limit of therapeutic range, time to INR [4.0, and first stable warfarin dose Validation Cohort R 2 of PGx algorithm 54% compared to 17% with clinical algorithm (P \ ) Median absolute prediction error, (mg/day) 1.0 with PGx algorithm compared to 1.5 with clinical (P \ 0.001) Homozygosity for CYP2C9 and VKORC1 variances increased risk of overanticoagulation, HR = (95% CI 9.46; 50.42) and 4.56 (95% CI 2.85; 7.30), respectively. Patients with CYP2C9*3/*3 experienced more bleeding during the first month of therapy (12.5% vs %, P =.066) Study group reached stable warfarin dose at 24 days compared to 35 days in control group. R 2 of PGx algorithm was 0.454, 95% CI (P \ 0.001) Patients with variant VKORC1 ± CYP2C9 reached target INR faster. (P \ ) European Americans (not African Americans) with VKORC1 ± CYP2C9 had higher risk of overanticoagulation. Risk of minor hemorrhage was not influenced by PGx variance. CYP2CP and VKORC1 explained 6.3% of variance in dose change over first 30 days PGx algorithm more accurate than clinical algorithm in patients requiring smaller or larger doses (35% vs. 24%, P \ 0.001, in patients requiring \21 mg/week and 32.8 vs. 13.3%, P \ 0.001, in patients requiring [42 mg/ wk) Early INRs were associated with all outcomes (all P \ 0.001) and were more informative than genotypes By considering VKORC1 status as well as CYP2C9, this PGx equation is able to explain 53 54% of the variability in the warfarin dose, compared to 17 22% with a clinical equation. (P \ 0.001) Despite prediction of dose via algorithm, induction strategies varied greatly by study site, which may have affected results PGx algorithm accounted for 54.1% of variance in warfarin dose Results suggest PGx dosing algorithms may be beneficial beyond predicting initial dose Largest population studied to date; multi-ethnic Greatest benefits of the PGx algorithm were seen in those that required lower or higher weekly doses of warfarin, which amounted to 46.2% of the study population Authors suggest that the most benefit of PGx warfarin dosing will be realized in settings with non-expert warfarin dosing, where the majority of patients are followed

6 Int J Clin Pharm (2011) 33: Table 1 continued Reference Study population Trial design Major findings Comments Lenzini [60] N = 969 in derivation N = 309 in internal validation N = 955 in external validation Eleven sites on 3 continents; patients had varying indications Developed clinical and PGx algorithm (CYP2C9 and VKORC1) to refine warfarin dose on day 4 or 5 of therapy; subsequently validated in prospective s 30-day follow-up period External validation INR on day 4: R 2 of the PGx algorithm was 40% compared to 28% for the clinical refinement algorithm. INR on day 5: R 2 of the PGx algorithm was 42% compared to 26% for the clinical refinement algorithm. Mean time in therapeutic range was 68% for PGx and 61% for clinical (P \ 0.05). Bleeding 1% with PGx and 4% with clinical When INR is known, PGx information can still improve warfarin dose refinement * Numbers in parentheses indicate reference cited research of Lenzini et al. [45], who found that VKORC1 contributed only 1% of the total R 2 of their warfarin dose refinement algorithms. In a study of orthopedic patients, CYP2C9 variances increased risk for excessive anticoagulation [24]. In addition, patients with CYP2C9*3/*3 were found to experience severe bleeding more often than other patients during the first month of therapy [46]. Research by Limdi and colleagues showed a higher risk of overanticoagulation in European Americans, but not in African Americans with VKORC1 ± CYP2C9 variances [57]. In recent years, several prospective studies have been published using algorithms that incorporate CYP2C9 and VKORC1 genotype, and various clinical factors to predict or modify warfarin dosage. Huang and colleagues conducted a prospective, randomized study to compare a PGxbased dosing model to traditional dosing in Chinese patients undergoing heart valve replacement. The primary endpoint was time to reach stable dose. PGx-based warfarin dosing significantly shortened the time to reach stable dose from 35 to 24 days (P = 0.001) [33]. Lenzini et al. [60] reported on a clinical refinement algorithm that could be used to modify warfarin dose when INR was available on day 4 or 5 of warfarin therapy. This prospective, randomized study included 955 subjects in the external validation and was conducted at eleven sites on three continents. For subjects with INR on day 4, the R 2 of the PGx algorithm was 40% compared to 28% for the clinical refinement algorithm. For subjects with INR on day 5, the R 2 of the PGx algorithm was 42% compared to 26% for the clinical algorithm. Patients were followed for one month and those that were dosed with the PGx algorithm spent more time in therapeutic range (68 vs. 61%) and had less bleeding (1 vs. 4%) compared to those dosed using clinical algorithm (P \ 0.05). The International Warfarin Pharmacogenetics Consortium also reported the development and validation of a PGx warfarin dosing algorithm. Their research involved 4,043 patients at a total of 21 sites in nine countries and is the largest clinical study of PGx-based warfarin dosing to date. The primary outcome measure was the stable warfarin dose and the PGx algorithm was tested against a clinical dosing method as well a fixed-dose approach of 5 mg daily. Patients were stratified in three groups: B21 mg per week, C49 mg per week and an intermediate group. As a measure of dose prediction, the researchers reported the number of patients for whom one algorithm estimated a dose within 20% of the actual dose. For the patients in the 21 mg or less, the PGx algorithm provided a significantly better prediction of dose than the clinical algorithm (35 vs. 24%; P \ 0.001). In the patients requiring 49 mg or more per week, the PGx algorithm also outperformed the clinical approach (32.8 vs. 13.3%); P \ 0.001) [48]. Although additional data is needed to address implications of PGx testing vis a vis many important clinical endpoints, these studies have supported the use of PGxbased warfarin dosing in high-risk populations, to reach INR goals more quickly and more effectively, even when INR information is available to the clinician. There are limitations of the application of these tools across ethnic populations, however, much is still unknown, as most algorithms can still only account for a little over half of the variation in warfarin dose [12, 47]. Further research encompassing larger, multi-ethnic populations and exploring hitherto unidentified genetic variables will allow user to predict warfarin therapy with more certainty, and give us more evidence as to the potential benefit of PGxbased warfarin dosing on meaningful clinical outcomes. An internet based program using an algorithm to facilitate

7 16 Int J Clin Pharm (2011) 33:10 19 pharmacogenomic guided warfarin dosing is available ( Prospective information concerning online, widespread utilization of this algorithm is currently not available. Discussion Summary of findings Applicability to community pharmacists Although pharmacists are involved with anticoagulation programs both in inpatient and outpatient settings, data looking at oral anticoagulation services in the traditional community pharmacy is relatively sparse [49, 50]. The most robust data to date comes from the U.S. and the Eckerd PatientCARE Network using the retail pharmacists employed by the Eckard s chain [51]. Specifically trained community pharmacists performed warfarin management on 50 patients with 435 patient visits over a 6 month period under a protocol approved by the patient s physician. Four hundred and thirty-five INR determinations were made with 243 (56%) being in the therapeutic range determined by the physician. Minor bruising was reported by ten patients, four patients reported blood in their stool and one hemorrhage occurred that required referral to the physician. No reports of thromboembolism occurred during the study. INR checks were done using the point-of-care CoaguChek System (Roche Diagnostics, Indianapolis, IN, USA). This small study suggests that trained community pharmacists, given time and resources, can perform oral anticoagulation management. Currently there is no point-of-care testing for warfarin metabolism polymorphisms, nor are there published data examining community pharmacists acquisition and interpretation of this information. Nevertheless, pharmacists have a potentially critical role to play in PGx testing and interpretation. For example, the ethical, legal, and social issues surrounding routine PGx testing are evolving rapidly [52]. Economic implications of PGx testing of warfarin metabolism have not been studied in detail. One pharmacoeconomic model used data from the COUMAGEN trial to estimate cost-effectiveness of PGx testing [35]. This model found only a small improvement in Quality Adjusted Life Years (QALY) (0.003) with an absolute increase in costs of US$162. An earlier decision analysis found that PGx testing would be cost-effective in elderly atrial fibrillation patients only if it significantly reduced bleeding events [53]. A more recent paper by Patrick and colleagues used a baseline cost of PGx test at about approximately US$500 (2009 dollars) to perform a similar cost effectiveness analysis on atrial fibrillation patients receiving warfarin [54]. Using a Markov model to elucidate what clinical circumstances PGx testing would be cost effective, these investigators found that such testing would be cost effective if it improved the amount of time patients were in therapeutic range by up to 10%. Thus, the largest impact of PGx testing may be on patients with a significant amount of INR variability. These findings were more optimistic than those of Eckman and associates who also used a Markov model to assess cost effectiveness of PGx testing in warfarin patients with atrial fibrillation [55]. They found that such testing would only have a favorable cost effectiveness profile if it prevented 32% of major bleeding events, was available within 24 h, and cost less than US$200 (2009 dollars). Clearly the information concerning cost effectiveness of warfarin PGx testing is variable and more data on the widespread use of this modality is needed a potential role for pharmacists with expertise in pharmacoeconomics. Another function of community pharmacists relates to education of both patients and prescribers. As PGx is a relatively new and rapidly expanding field, many patients and prescribers may not be armed with the latest knowledge concerning its role in clinical medicine. That pharmacists may be able to aid in educating patients about PGx testing was noted by the UK s Royal Pharmaceutical Society [55]. Pharmacists themselves, especially those out of training for a significant period of time, may need to participate in an educational program in this area [56]. In the U.S. the American Pharmacists Association has developed the PharmGenEd program designed to increase awareness and knowledge about PGx testing [57]. This continuing education program utilizes both online ( pharmacogenomics.ucsd.edu) and live formats through collaboration with health care organizations, colleges of pharmacy, and experts in PGx testing. Such programs will be necessary to assure an appropriate level of training for pharmacists in this area. Finally one could speculate that community pharmacies may eventually become a point of sale when self-use PGx testing becomes widely available and affordable. Interpretation of such tests by the lay public may be difficult without the aid of health care professionals such as pharmacists [58]. Pharmacists may incorporate PGx testing of many drugs, including warfarin, as they design dosing regimens for drugs. A survey of oral anticoagulation clinics published in 2009 found that, at that time, only about 12% of these sites had used PGx testing for warfarin dosing [59]. Should future pharmacoeconomic and clinical studies show a consistent benefit to such testing, this number is very likely to increase. As noted by Amruso community pharmacy implementation of oral anticoagulation services is feasible, but many pharmacies will require significant changes in staff training, resources, and physician

8 Int J Clin Pharm (2011) 33: acceptance to implement them [51]. These should occur prior to or concomitant with the implementation of PGx testing for warfarin. Conclusion Pharmacogenomics, a science in its relative infancy, has the potential to improve the safety and efficacy of medications in humans. Like many new innovations in drug therapy, initial efforts have been fraught with uncertainty in several areas. To date, genetic polymorphisms in warfarin metabolism have been conclusively shown to explain a significant amount of dosing variability with this drug. Several groups of investigators have incorporated PGx testing into warfarin dosing algorithms and have generally found an improvement in time within the therapeutic range with less laboratory testing. Larger studies are needed to examine hard outcomes of such testing including a possible reduction of major bleeding or thromboembolism. Cost effectiveness studies to date have not shown a robust benefit of PGx testing, although this will only change for the better as facile and less expensive tests/technologies are developed. Pharmacists, both those practicing in anticoagulation clinics as well as community based practitioners, are ideally positioned to aid patients and providers in the proper application of PGx testing. Training in this area should be a priority for both academic centers and employers. Funding None. Conflicts of interest None of the authors have any real or potential conflicts of interest concerning this work. References 1. Wysowski DK, Nourjah P, Swartz L. Bleeding complications with warfarin use: a prevalent adverse effect resulting in regulatory action. Arch Intern Med. 2007;167: Limdi NA, Veenstra DL. Warfarin pharmacogenetics. Pharmacotherapy. 2008;28: Chiquette E, Amato MG, Bussey HI. Comparison of an anticoagulation clinic with usual medical care: anticoagulation control, patient outcomes, and health care costs. Arch Intern Med. 1998;158: Wilt VM, Gums JG, Ahmed OI, Moore LM. Outcome analysis of a pharmacist-managed anticoagulation service. Pharmacotherapy. 1995;15: Sanderson S, Emery J, Higgins J. CYP2C9 gene variants, drug dose, and bleeding risk in warfarin-treated patients: a HuGEnet systematic review and meta-analysis. Genet Med. 2005;7: Lindh JD, Holm L, Andersson ML, Rane A. Influence of CYP2C9 genotype on warfarin dose requirements a systematic review and meta-analysis. Eur J Clin Pharmacol. 2009;65: Rosove MH, Grody WW. Should we be applying warfarin pharmacogenetics to clinical practice? No, not now. Ann Intern Med. 2009;151: Kamali F, Wynne H. Pharmacogenetics of warfarin. Annu Rev Med. 2010;61: Stehle S, Kirchheiner J, Lazar A, Fuhr U. Pharmacogenetics of oral anticoagulants: a basis for dose individualization. Clin Pharmacokinet. 2008;47: Herman D, Locatelli I, Grabnar I, Peternel P, Stegnar M, Mrhar A, et al. Influence of CYP2C9 polymorphisms, demographic factors and concomitant drug therapy on warfarin metabolism and maintenance dose. Pharmacogenomics J. 2005;5: Gage BF, Eby C, Milligan PE, Banet GA, Duncan JR, McLeod HL. Use of pharmacogenetics and clinical factors to predict the maintenance dose of warfarin. Thromb Haemost. 2004;91: Caldwell MD, Berg RL, Zhang KQ, Glurich I, Schmelzer JR, Yale SH, et al. Evaluation of genetic factors for warfarin dose prediction. Clin Med Res. 2007;5: Gulseth MP, Grice GR, Dager WE. Pharmacogenomics of warfarin: uncovering a piece of the warfarin mystery. Am J Health Syst Pharm. 2009;66: Adcock DM, Koftan C, Crisan D, Kiechle FL. Effect of polymorphisms in the cytochrome P450 CYP2C9 gene on warfarin anticoagulation. Arch Pathol Lab Med. 2004;128: Gan GG, Phipps ME, Ku CS, Teh A, Sangkar V. Genetic polymorphism of the CYP2C9 subfamily of 3 different races in warfarin maintenance dose. Int J Hematol. 2004;80: Carlquist JF, Horne BD, Muhlestein JB, Lappé DL, Whiting BM, Kolek MJ, et al. Genotypes of the cytochrome p450 isoform, CYP2C9, and the vitamin K epoxide reductase complex subunit 1 conjointly determine stable warfarin dose: a prospective study. J Thromb Thrombolysis. 2006;22: Tham LS, Goh BC, Nafziger A, Guo JY, Wang LZ, Soong R, et al. A warfarin-dosing model in Asians that uses single-nucleotide polymorphisms in vitamin K epoxide reductase complex and cytochrome P450 2C9. Clin Pharmacol Ther. 2006;80: Meckley LM, Wittkowsky AK, Rieder MJ, Rettie AE, Veenstra DL. An analysis of the relative effects of VKORC1 and CYP2C9 variants on anticoagulation related outcomes in warfarin-treated patients. Thromb Haemost. 2008;100: Yin T, Hanada H, Miyashita K, Kokubo Y, Akaiwa Y, Otsubo R, et al. No association between vitamin K epoxide reductase complex subunit 1-like 1 (VKORC1L1) and the variability of warfarin dose requirement in a Japanese patient population. Thromb Res. 2008;122: Wen MS, Lee M, Chen JJ, Chuang HP, Lu LS, Chen CH, et al. Prospective study of warfarin dosage requirements based on CYP2C9 and VKORC1 genotypes. Clin Pharmacol Ther. 2008; 84: Wu AH, Wang P, Smith A, Haller C, Drake K, Linder M, et al. Dosing algorithm for warfarin using CYP2C9 and VKORC1 genotyping from a multi-ethnic population: comparison with other equations. Pharmacogenomics. 2008;9: Fuchshuber-Moraes M, Perini JA, Rosskopf D, Suarez-Kurtz G. Exploring warfarin pharmacogenomics with the extreme-discordant-phenotype methodology: impact of FVII polymorphisms on stable anticoagulation with warfarin. Eur J Clin Pharmacol. 2009;65: Limdi NA, Wadelius M, Cavallari L, Eriksson N, Crawford DC, Lee MT, et al. Warfarin pharmacogenetics: a single VKORC1 polymorphism is predictive of dose across 3 racial groups. Blood. 2010;115: Voora D, Eby C, Linder MW, Milligan PE, Bukaveckas BL, McLeod HL, et al. Prospective dosing of warfarin based on cytochrome P-450 2C9 genotype. Thromb Haemost. 2005;93:

9 18 Int J Clin Pharm (2011) 33: Kimura R, Miyashita K, Kokubo Y, Akaiwa Y, Otsubo R, Nagatsuka K, et al. Genotypes of vitamin K epoxide reductase, gamma-glutamyl carboxylase, and cytochrome P450 2C9 as determinants of daily warfarin dose in Japanese patients. Thromb Res. 2007;120: Kosaki K, Yamaghishi C, Sato R, Semejima H, Fuijita H, Tamura K, et al. 1173C[T polymorphism in VKORC1 modulates the required warfarin dose. Pediatr Cardiol. 2006;27: Schalekamp T, Brassé BP, Roijers JF, Chahid Y, van Geest- Daalderop JH, de Vries-Goldschmeding H, et al. VKORC1 and CYP2C9 genotypes and acenocoumarol anticoagulation status: interaction between both genotypes affects overanticoagulation. Clin Pharmacol Ther. 2006;80: Lal S, Sandanaraj E, Jada SR, Kong MC, Lee LH, Goh BC, et al. Influence of APOE genotypes and VKORC1 haplotypes on warfarin dose requirements in Asian patients. Br J Clin Pharmacol. 2008;65: Tanira MO, Al-Mukhaini MK, Al-Hinai AT, Al Balushi KA, Ahmed IS. Frequency of CYP2C9 genotypes among Omani patients receiving warfarin and its correlation with warfarin dose. Community Genet. 2007;10: Gage BF, Eby C, Johnson JA, Deych E, Rieder MJ, Ridker PM, et al. Use of pharmacogenetic and clinical factors to predict the therapeutic dose of warfarin. Clin Pharmacol Ther. 2008;84: Limdi NA, Arnett DK, Goldstein JA, Beasley TM, McGwin G, Adler BK, et al. Influence of CYP2C9 and VKORC1 on warfarin dose, anticoagulation attainment and maintenance among European-Americans and African-Americans. Pharmacogenomics. 2008;9: Limdi NA, Beasley TM, Crowley MR, Goldstein JA, Rieder MJ, Flockhart DA, et al. VKORC1 polymorphisms, haplotypes and haplotype groups on warfarin dose among African-Americans and European-Americans. Pharmacogenomics. 2008;9: Huang SW, Chen HS, Wang XQ, Huang L, Xu DL, Hu XJ, et al. Validation of VKORC1 and CYP2C9 genotypes on interindividual warfarin maintenance dose: a prospective study in Chinese patients. Pharmacogenet Genomics. 2009;19: Palacio L, Falla D, Tobon I, Mejia F, Lewis JE, Martinez AF, et al. Pharmacogenetic impact of VKORC1 and CYP2C9 allelic variants on warfarin dose requirements in a hispanic population isolate. Clin Appl Thromb Hemost. 2010;16: Meckley LM, Gudgeon JM, Anderson JL, Williams MS, Veenstra DL. A policy model to evaluate the benefits, risks and costs of warfarin pharmacogenomic testing. Pharmacoeconomics. 2010; 28: Takeuchi F, McGinnis R, Bourgeois S, Barnes C, Eriksson N, Soranzo N, et al. A genome-wide association study confirms VKORC1, CYP2C9, and CYP4F2 as principal genetic determinants of warfarin dose. PLoS Genet. 2009;5: Pautas E, Moreau C, Gouin-Thibault I, Golmard JL, Mahé I, Legendre C, et al. Genetic factors (VKORC1, CYP2C9, EPHX1, and CYP4F2) are predictor variables for warfarin response in very elderly, frail inpatients. Clin Pharmacol Ther. 2010;87: Kealey C, Chen Z, Christie J, Thorn CF, Whitehead AS, Price M, et al. Warfarin and cytochrome P450 2C9 genotype: possible ethnic variation in warfarin sensitivity. Pharmacogenomics. 2007; 8: Momary KM, Shapiro NL, Viana MA, Nutescu EA, Helgason CM, Cavallari LH. Factors influencing warfarin dose requirements in African-Americans. Pharmacogenomics. 2007;8: Visser LE, van Vliet M, van Schaik RH, Kasbergen AA, De Smet PA, Vulto AG, et al. The risk of overanticoagulation in patients with cytochrome P450 CYP2C9*2 or CYP2C9*3 alleles on acenocoumarol or phenprocoumon. Pharmacogenetics. 2004;14: Topic E, Stefanovic M, Samardzija M. Association between the CYP2C9 polymorphism and the drug metabolism phenotype. Clin Chem Lab Med. 2004;42: Ngow H, Teh LK, Langmia IM, Lee WL, Harun R, Ismail R, Salleh MZ. Role of pharmacodiagnostic of CYP2C9 variants in the optimization of warfarin therapy in Malaysia: a 6-month follow-up study. Xenobiotica. 2008;38: Ohno M, Yamamoto A, Ono A, Miura G, Funamoto M, Takemoto Y, et al. Influence of clinical and genetic factors on warfarin dose requirements among Japanese patients. Eur J Clin Pharmacol. 2009;65: Miao L, Yang J, Huang C, Shen Z. Contribution of age, body weight, and CYP2C9 and VKORC1 genotype to the anticoagulant response to warfarin: proposal for a new dosing regimen in Chinese patients. Eur J Clin Pharmacol. 2007;63: Lenzini PA, Grice GR, Milligan PE, Dowd MB, Subherwal S, Deych E, et al. Laboratory and clinical outcomes of pharmacogenetic vs. clinical protocols for warfarin initiation in orthopedic patients. J Thromb Haemost. 2008;6: Wadelius M, Chen LY, Lindh JD, Eriksson N, Ghori MJ, Bumpstead S, et al. The largest prospective warfarin-treated supports genetic forecasting. Blood. 2009;113: Sconce EA, Khan TI, Wynne HA, Avery P, Monkhouse L, King BP, et al. The impact of CYP2C9 and VKORC1 genetic polymorphism and patient characteristics upon warfarin dose requirements: proposal for a new dosing regimen. Blood. 2005; 106: International Warfarin Pharmacogenetics Consortium, Klein TE, Altman RB, Eriksson N, Gage BF, Kimmel SE, et al. Estimation of the warfarin dose with clinical and pharmacogenetic data. N Engl J Med. 2009;360: Airee A, Guirguis AB, Mohammad RA. Clinical outcomes and pharmacists acceptance of a community hospital s anticoagulation management service utilizing decentralized clinical staff pharmacists. Ann Pharmacother. 2009;43: Willey ML, Chagan L, Sisca TS, et al. A pharmacist-managed anticoagulation clinic: six-year assessment of patient outcomes. Am J Health Syst Pharm. 2003;60: Amruso NA. Ability of clinical pharmacists in a community pharmacy setting to manage anticoagulation therapy. J Am Pharm Assoc. 2004;44: Lee KC, Ma JD, Kuo GM. Pharmacogenomics: bridging the gap between science and practice. J Am Pharm Assoc. 2010;50: e Leey JA, McCabe S, Koch JA, Miles TP. Cost-effectiveness of genotype-guided warfarin therapy for anticoagulation in elderly patients with atrial fibrillation. Am J Geriatr Pharmacother. 2009;7: Patrick AR, Avorn J, Choudhry NK. Cost-effectiveness of genotype-guided warfarin dosing for patients with atrial fibrillation. Circ Cardiovasc Qual Outcomes. 2009;2: Eckman MH, Rosand J, Greenberg SM, Gage BF. Cost-effectiveness of using pharmacogenetic information in warfarin dosing for patients with nonvalvular atrial fibrillation. Ann Intern Med. 2009;150: Clemerson JP, Payne K, Bissell P, Anderson C. Pharmacogenetics, the next challenge for pharmacy? Pharm World Sci. 2006;28: Limdi NA, Wiener H, Goldstein JA, Acton RT, Beasley TM. Influence of CYP2C9 and VKORC1 on warfarin response during initiation of therapy. Blood Cells Mol Dis. 2009;43: Beckman L. Are genetic self-tests dangerous? Assessing the commercialization of genetic testing in terms of personal autonomy. Theor Med Bioeth. 2004;25:

10 Int J Clin Pharm (2011) 33: Kadafour M, Haugh R, Posin M, Kayser SR, Shin J. Survey on warfarin pharmacogenetic testing among anticoagulation providers. Pharmacogenomics. 2009;10: Lenzini P, Wadelius M, Kimmel S, Anderson JL, Jorgensen AL, Pirmohamed M, et al. Integration of genetic, clinical, and INR data to refine warfarin dosing. Clin Pharmacol Ther. 2010; 87: Anderson JL, Horne BD, Stevens SM, Grove AS, Barton S, Nicholas ZP, et al. Randomized trial of genotype-guided versus standard warfarin dosing in patients initiating oral anticoagulation. Circulation. 2007;116: Schwarz UI, Ritchie MD, Bradford Y, Li C, Dudek SM, Frye- Anderson A, et al. Genetic determinants of response to warfarin during initial anticoagulation. N Engl J Med. 2008;358: Li C, Schwarz UI, Ritchie MD, Roden DM, Stein CM, Kurnik D. Relative contribution of CYP2C9 and VKORC1 genotypes and early INR response to the prediction of warfarin sensitivity during initiation of therapy. Blood. 2009;113: