Vanessa Cristina de Oliveira Almeida Daniel Dias Ribeiro Karina Braga Gomes Ana Lúcia Brunialti Godard

Transcription

1 Mol Diagn Ther DOI /s ORIGINAL RESEARCH ARTICLE Polymorphisms of CYP2C9, VKORC1, MDR1, APOE and UGT1A1 Genes and the Therapeutic Warfarin Dose in Brazilian Patients with Thrombosis: A Prospective Cohort Study Vanessa Cristina de Oliveira Almeida Daniel Dias Ribeiro Karina Braga Gomes Ana Lúcia Brunialti Godard Ó Springer International Publishing Switzerland 2014 Abstract Background There are several pharmacogenetic algorithms to determine the warfarin doses required in patients treated for thromboembolism, but they only explain 60 % of dose variation, suggesting that other genes may influence the dose required. Objectives This study aimed to evaluate the impact of clinical factors and CYP2C9*2, CYP2C9*3, VKORC1-1639G[A, MDR1 3435C[T, APOE* e4, and UGT1A1 (TA) n polymorphisms on the warfarin dose required, especially in those individuals requiring a high warfarin dose. Methods We studied 116 Brazilian patients who received warfarin therapy for thromboembolism. Associations V. C. de Oliveira Almeida (&) K. B. Gomes A. L. B. Godard Department of General Biology, Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais, Av. Antonio Carlos, 6627, Cx. Postal 486, Pampulha, Belo Horizonte, Minas Gerais , Brazil vanessa@rumotecnologia.com.br K. B. Gomes karinabgb@gmail.com A. L. B. Godard brunialt@icb.ufmg.br V. C. de Oliveira Almeida Department of Molecular Genetics, Hermes Pardini, Minas Gerais, Brazil D. D. Ribeiro Clinical Hospital, Universidade Federal de Minas Gerais, Minas Gerais, Brazil ddribeiro@terra.com.br K. B. Gomes Department of Clinical and Toxicological Analysis, Faculty of Pharmacy, Universidade Federal de Minas Gerais, Minas Gerais, Brazil between dose variability and age, body mass index (BMI), gender, use of warfarin antagonists, and genetic polymorphisms were examined. Results CYP2C9*2, CYP2C9*3, VKORC1-1639G[A, and APOE *e4 were associated with lower warfarin doses. Of these subjects, 21 % required a warfarin dose higher than 70 mg/week, which was associated with a BMI greater than 25 kg/m 2, use of warfarin antagonists, and the presence of the MDR1 3435T allele and UGT1A1(TA) 7 polymorphism. These individuals were considered to exhibit warfarin resistance. The individuals with the MDR1 3435TT genotype required a dose 21 % higher than that required by 3435CT and 3435CC individuals. The UGT1A1(TA) 7 allele was positively correlated with the warfarin dose. Conclusion CYP2C9*2, CYP2C9*3, VKORC1-1639G[A, and APOE *e4 were associated with lower warfarin doses, while MDR1 3435C[T andugt1a1(ta) n polymorphisms were associated with a requirement for higher doses. This is the first study to evaluate warfarin resistance, APOE *e4 and UGT1A1(TA) n genotypes in the Brazilian population, and the association of these two genotypes with warfarin dose required. Key Points The CYP2C9, APOE and VKORC1 polymorphisms were associated with warfarin dose. Twenty-four subjects (20.7 %) required doses C70 mg/week (mean ± 37.9 mg/week). MDR1 3435TT genotype and 3435T allele were associated with higher warfarin dose. UGT1A1(TA) n was associated with warfarin dose in a dominant inheritance model. Warfarin resistance is affected by MDR1 C3435T and use of antagonists.

2 V. C. Oliveira Almeida et al. 1 Introduction Warfarin is an oral anticoagulant widely used for the prevention and treatment of thromboembolic diseases [1, 2]. Stable doses range from 1 to 20 mg/day [3], and prediction of an effective and safe dose is difficult because of a narrow therapeutic range and large inter-individual variability in dose requirements [1, 2]. This variability depends on clinical and environmental factors, such as age, sex, body mass index (BMI), co-morbidities, co-medications and vitamin K intake; as well as genetic factors [4, 5]. In Caucasians, % of the variability in the warfarin dose can be attributed to polymorphisms of the CYP2C9 and VKORC1 genes, which are responsible for warfarin s metabolic clearance and pharmacodynamic action, respectively. CYP2C9*2 (rs ), CYP2C9*3 (rs ), and VKORC1-1639G[A (rs ) polymorphisms are the major genetic determinants of higher sensitivity to warfarin, and consequently, of over-anticoagulation [3, 6, 7]. Several proposed pharmacogenetic algorithms to determine the warfarin dose required in an individual explain only about 60 % of dose variation [2, 8], suggesting that other genes may affect warfarin dose. The APOE e4 (rs429358) polymorphism has been associated with higher warfarin doses owing to its involvement in hepatic vitamin K uptake [9 11]. Furthermore, we have previously shown that the MDR1 3435C[T (rs ) polymorphism contributes to the requirements for higher doses [12]. Recently, Bratton et al. demonstrated that inactive P450 metabolites of warfarin, the hydroxywarfarins, are glucuronidated by uridine diphosphate glucoronosyltransferases (UGTs) enzymes, especially the UGT1A1 enzyme. Significantly lower R-7-hydroxywarfarin glucoronidation occurs in the homozygous UGT1A1(TA) 7/7, with seven thymine adenine (TA) repeats in the UGT1A1 promoter region. The wild-type allele comprises six repeats and is denoted as UGT1A1(TA) 6/6 or UGT1A1*. The TA repeat length is inversely related to enzyme activity; nevertheless, how this polymorphism affects warfarin response remains unclear [13]. Some anticoagulated individuals require high warfarin doses. These patients, classified as warfarin resistant, are those who require a dose higher than 70 mg/week to achieve an international normalized ratio (INR) in the therapeutic range [14, 15], or who never achieve a therapeutic INR with this dose [16]. The causes of warfarin resistance are not completely understood. No studies on warfarin resistance have been conducted with Brazilian patients. The Brazilian population is highly miscegenated, and data obtained in other populations do not necessarily apply to this population, which has particular inter- and intra-ethnic characteristics. Therefore, we investigated the impact of clinical factors and CYP2C9*2, CYP2C9*3, VKORC1-1639G[A, MDR1 3435C[T, APOE *e4, and UGT1A1(TA) n polymorphisms on the warfarin dose in Brazilian patients taking anticoagulants to prevent venous thromboembolism (VTE). To our knowledge, this is the first study to evaluate these polymorphisms in a population with VTE in Brazil, and to investigate the association between UGT1A1(TA) n and warfarin dose required. 2 Material and Methods 2.1 Subjects This prospective cohort study included 116 subjects, randomly selected, who received warfarin anticoagulation therapy and were monitored at the Hematology Ambulatory in the Clinical Hospital, Universidade Federal de Minas Gerais (UFMG), Belo Horizonte Brazil, in All participants presented with first event or recurrence of deep venous thrombosis (DVT) of the lower and/or higher limbs, cerebral venous sinus, vena cava, portal vein, or mesenteric vein, pulmonary thromboembolism, and/or Budd Chiari syndrome. Exclusion criteria included pregnancy, postpartum interval, estrogen use, solid organ transplant, liver disease, neoplasia, and myeloproliferative disorders. Treatment compliance was analyzed after interviewing all patients who had participated in the program, to assess their awareness of the need for, and importance of, anticoagulant therapy, and the care that must be taken. Those who discontinued the treatment were excluded. This study was approved by the Research Ethics Committee at UFMG, under the protocol ETIC 334/08. Informed written consent was obtained from all participants. 2.2 Data Collection Clinical data regarding age, gender, BMI, therapeutic indications for warfarin, and co-morbidities were obtained from the patients medical records. The data regarding weekly warfarin dose (expressed as mg), INR value, and use of co-medications were collected from each patient during the medical visits. Each patient had a total of five medical visits. A stable therapeutic warfarin dose was defined as a dose that led to therapeutic INR values on three consecutives visits (1). Co-medications were classified as non-influencers, warfarin potentiators, and warfarin antagonists [17]. Only well documented drug drug interactions whose clinical significance was classified as contraindicated, major, or moderate and documentation level was classified as excellent and good were considered [17].

3 Genetic Polymorphisms and Warfarin Dose The patients were classified according to a weekly average warfarin dose administered: I. In order to evaluate the influence of the warfarin dose required: B21.0 mg, mg, and C49.0 mg/ week [2]. II. In order to evaluate the warfarin resistance-related variables: dose \70 mg, and dose C70 mg [14, 15]. 2.3 Genotyping Genomic DNA was extracted from blood samples using a Puregene DNA isolation kit (Gentra Systems, Inc., Minneapolis, USA) according to the manufacturer s instructions. Using standardized TaqMan assays (Applied Biosystems, California, USA), subjects were genotyped for the polymorphisms CYP2C9*2 (rs ; C_ ) and CYP2C9*3 (rs ; C_ ) in the CYP2C9 gene, -1639G[A (rs ; C_ ) in the VKORC1 gene and 388T[C (rs429358; C _10) and 526C[T (rs7412; C _10) in the APOE gene by using a 7500 Real-Time PCR System (Applied Biosystems, California, USA). Subjects were also genotyped for 3435C[T (rs ) in the MDR1 gene using polymerase chain reaction (PCR) as previously reported [18] and for UGT1A1(TA) n (rs ) in the UGT1A1 gene by PCR using the primers F: 6-FAM (6-carboxifluorescein)- 5 0 GTCACGTGACACAGTCAAAC3 0 and R: 5 0 TTTGCTC CTGCCAGAGGTT3 0 [19]. UGT1A1(TA) n PCR fragments were analyzed by capillary electrophoresis in an ABI3730 DNA Analyzer (Applied Biosystems, California, USA). The alleles were identified by the difference in size of amplified fragment using GeneMapperÒ Software v4.1 (Applied Biosystems, California, USA). Control samples for each allelic variant and interpretation of the results by two different analysts were performed in all genetic tests. Analysts were blinded to the patients clinical data. 2.4 Statistical Analysis The Kolmogorov Smirnov test was used to test for normality. A Chi-square test was used to evaluate the Hardy Weinberg equilibrium. The allelic and genotypic frequencies in patients grouped according to classifications I and II were evaluated using Fisher s Exact Test or Pearson s Chi- Square. Due to the lack of consensus regarding the inheritance model of the polymorphisms studied, dominant [AA 9 (Aa? aa)] and recessive [(AA? Aa) 9 aa] patterns of inheritance were considered. T tests and one-way analysis of variance (ANOVA), followed by Tukey s test, were used to compare the mean warfarin doses. To evaluate the correlations between dose, age, and BMI, the Pearson correlation was applied. The Spearman correlation was applied to evaluate the correlations between dose and ordinal qualitative variables (genotypes, allelic variant, gender, use of warfarin antagonist and potentiators). To determine which of the studied variables (genotypes, allelic variant of each polymorphism, gender, age, BMI, use of warfarin antagonist and potentiators) are independently related to the warfarin dose required, a univariate logistic regression analysis was first applied, followed by a multivariate logistic regression analysis. The latter model included only the variables associated with warfarin dose with p \ 0.2. Two analyses were performed with cutoffs for doses at 70 and 49 mg/week as dependent variables. The result of the proposed multivariate logistic regression was applied to a goodness-of-fit test from Hosmer and Lemeshow to assess the model s adequacy. SPSS version 13.0 software was used for statistical analysis. Values of p B 0.05 were considered significant. Moreover, the multiple testing was performed for genes independently associated and the p value was adjusted for 0.01 (0.05/5). 3 Results 3.1 Patient Characteristics The clinical and demographic characteristics of the 116 studied subjects are summarized in Table 1. The mean warfarin dose was 49.5 ± 33.6 mg/week and a wide variation in warfarin dose ( mg) was required to achieve/maintain INR within the therapeutic range of 2 3. We identified 109 (94 %) participants who presented a stable therapeutic warfarin dose, considered as the dose that led to therapeutic INR values on three consecutive visits, although 20 (18 %) in this group required a warfarin dose C70 mg/week. 3.2 Correlation between Non-Genetic Factors and Warfarin Dose The warfarin dose decreased with increasing age (r 2 = , p = 0.017). Individuals older than 60 years required lower doses of warfarin (28.8 ± 14.1 mg), compared with those under 60 years (mean 52.5 ± 34.6 mg) (p = 0.009). Furthermore, individuals older than 80 years required a 47 % dose lower (20.1 ± 2.3 mg/week) (p = 0.001) than those younger than 18 years (43.1 ± 4.7 mg/week). Moreover, the warfarin dose was positively correlated with BMI (r 2 = 0.039, p = 0.034). The use of warfarin antagonists showed a positive correlation with dose (r 2 = 0.049, p = 0.003). The warfarin dose of individuals using antagonists (75.5 ± 10.1 mg)

4 V. C. Oliveira Almeida et al. Table 1 Clinical and demographic characteristics of the study population Variable % (n) Gender Female 65.5 % (76) Male 34.5 % (40) Age Mean (years) 42.3 ± 14.5 \60 years 87.1 % (101) [60 years 12.9 % (15) Body mass index (BMI) (kg/m 2 ) 27.3 ± 5.2 \ % (41) 25 B BMI \ % (36) C % (39) Indication for warfarin treatment Deep venous thrombosis (DVT) 81 % (94) of the lower and/or higher limbs Pulmonary thromboembolism 17 % (20) Cerebral venous sinus thrombosis 7 % (8) Others (Budd Chiari syndrome, 5 % (6) thrombosis of portal vein and/or vena cava) Co-medications with warfarin a Only warfarin 31.0 % (36) No interacting drug 44.8 % (52) Antagonists of warfarin (drugs) b 10.3 % (12) Potentiators of warfarin (drugs) c 26.0 % (21) Warfarin dose (mg/week) Mean (mg/week) 49.5? 33.6 Dose range (mg/week) Dose B21 mg/week 9.5 % (11) Dose mg/week 56.0 % (65) Dose C 49 mg/week 34.5 % (40) Dose \ 70 mg/week 79.3 % (92) Dose C 70 mg/week 20.7 % (24) Concomitant diseases d 61.2 % (75) a An individual can have more than one antagonist or more than one potentiator and more than one indication for warfarin treatment. Five individuals used antagonists and potentiators of warfarin concomitantly b Azathioprine (6.9 %), carbamazepine (3.4 %), mesalamine (0.9 %), and phenobarbital (0.9 %) c Omeprazole (6.9 %), salicylates (6.9 %), simvastatin (6.0 %), amitriptyline (4.3 %), fluoxetine (3.4 %), sertraline (3.4 %), valproic acid (2.6 %), amiodarone (1.7 %), citalopram (1.7 %), and cimetidine, desvenlafaxine, glucosamine, pantoprazole, and ranitidine (0.9 % each) d Hypertension, autoimmune diseases, sickle cell anemia, and psychiatric disorders was 62 % higher than that of non-users (46.5 ± 3.2 mg; p = 0.004), and 85 % higher than that of warfarin potentiator users (40.8 ± 4.2 mg; p = 0.002). We observed that 33 % of subjects with a dose C70 mg used antagonists. However, no correlation was observed between warfarin dose and potentiator use (p [ 0.05). 3.3 Association between Genetic Factors and Warfarin Dose (Classification I) Classification I divided the individuals into three dose ranges: B21 mg, mg, and C49 mg/week. The CYP2C9, VKORC1, MDR1, and APOE gene polymorphisms were in Hardy Weinberg equilibrium in this population, except for the UGT1A1(TA) n polymorphism (p = 0.038). For individuals with classification I (dose: B21 mg, mg, and C49 mg/week), the MDR1 and UGT1A1 polymorphisms were not associated with warfarin dose; however, an association was observed with CYP2C9, APOE and VKORC1 frequencies (Table 2). The frequencies of the alleles CYP2C9*2 and CYP2C9*3 were the highest in the group that used B21 mg warfarin (p \ 0.001) (Table 2). No relationship was observed between genotype and dose in the whole study group. However, when individuals with C70 mg warfarin were excluded, the *1/*1 carriers needed higher doses compared with carriers of one or more copies of alleles *2 and *3 (Table 3). The VKORC1-1639G[A polymorphism showed an association with warfarin dose when evaluated by genotype (p \ 0.001), mainly when a dominant inheritance model was considered [GG X (GA? AA)] (p = 0.004, OR = 3.20, CI = ). The -1639GA and -1639AA genotypes were three times more frequent in subjects with a warfarin dose \49 mg than in subjects with dose C49 mg, and the -1639A allele was more frequent in individuals with dose B21 mg (p = 0.006). The mean warfarin dose differed significantly (p = 0.007) between individuals with the -1639GG (57.8 ± 38.3 mg), -1639GA (42.3 ± 25.2 mg), and -1639AA (17.4 ± 4.7 mg) genotypes (p = 0.007). The doses of -1639AA individuals were 70 % (variation of 40.4 mg) lower than doses of -1639GG subjects (p = 0.045). Therefore, the mean warfarin dose of -1639GA individuals was 27 % (variation of 15.5 mg) lower than that of -1639GG subjects (p = 0.028). The presence of the e4 allele, regardless of whether there were one or two copies, was also associated with warfarin dose [(e4/e4? e4/e2? e4/e3) 9 other genotypes] (p = 0.006, OR = 2.31, CI = ). It was observed that 78 % of e4 carriers required a dose \49 mg. Doses of e4 carriers (41.5 ± 18.5) were 21 % lower than doses of e4 non-carriers (52.5 ± 37.5 mg) (p = 0.038). e2 and e3 alleles were not associated with dose (p = 0.669). For classification I, multivariate logistic regression analysis showed that VKORC1-1639G[A (p = 0.014, OR = 0.30, CI = ), e4 allele presence

5 Genetic Polymorphisms and Warfarin Dose (p = 0.041; OR = 3.11, CI = ), age (p \ 0.001, OR = 0.92, CI = ), and BMI (p = 0.011, OR = 1.14, CI = ) were independently associated with warfarin dose (p \ 0.05), (Hosmer and Lemeshow Index = 0.898). 3.4 Association between Genetic Factors and Warfarin Dose (Classification II) Classification II divided the individuals into two dose ranges: \70 mg and C70 mg warfarin/week. It was observed that 24 subjects (20.7 %) required doses C70 mg/week (mean ± 37.9 mg). For classification II (doses \70 mg and C70 mg/week), the VKORC1, MDR1, and UGT1A1(TA) n polymorphisms were significantly associated with the warfarin dose. The CYP2C9 and APOE polymorphisms are not presented (Table 2). The VKORC1-1639G[A polymorphism showed an association with warfarin dose, either by genotype (p \ 0.001) or by the dominant inheritance model [(AA? AG) 9 GG] (p = 0.041, OR = 2.76, CI = ), Table 2 Allelic and genotypic frequencies of CYP2C9, VKORC1, APOE, MDR1, and UGT1A1 and their respective weekly warfarin dose in the groups in the I and II classifications a Dose values without standard deviation refer to a single individual datum b Allelic frequencies comparison between subjects with and without the e4 allele c Dose comparison between subjects with and without the e4 allele Cutoff Polymorphism Genotype/ allele Classification I cutoff 21 and 49 mg Classification II cutoff 70 mg Frequency v 2 test (p value) Dose (mean) T test or Anova (p value) CYP2C9*2 and *1/* \ ± CYP2C9*3 *1/* ± 37.9 *1/* ± 37.8 *2/* a *2/* a *3/*3 * * * VKORC1 GG \ ± G[A GA ± 25.2 AA ± 4.7 G A 0.25 APOE e2/e ± e2, e3 and e4 e3/e ± 31.7 e4/e ± 10.8 e2/e ± 55.6 e2/e ± 12.4 e3/e ± 20.0 e b c e e MDR 3435C[T CC ± CT ± 28.7 TT ± 38.3 C T UGT1A1(TA)n ? ? ?

6 V. C. Oliveira Almeida et al. Table 3 Comparison of warfarin doses between CYP2C9 genotypes in individuals from the group with warfarin dose \70 mg a One copy of CYP2C9 variant b Two copies of CYP2C9 variant c One and/or two copies of CYP2C9 variant CYP2C9 genotypes % (n) Dose (mg/week) % Dose reduction compared with *1/*1 p Value compared with*1/*1 *1/* (69) 38.4 ± 13.0 *1/* (15) 31.7 ± *1/*3 0.08(6) 24.2 ± *2/* (1) *2/* (1) *3/*3 0 *1/*2? *1/*3 a 0.23 (21) 29.5 ± *2/*2? *2/*3? *3/*3 b 0.02 (2) 19.8 ± *1/*2? *1/*3? *2/*2? *2/*3? *3/*3 c 0.25 (23) 28.7 ± and the allele -1639A was more frequent in the group with \70 mg warfarin dose (p = 0.001). Considering the recessive inheritance model [(CC? CT) 9 TT] for the MDR1 polymorphism, 3435TT genotype (p = 0.038, OR = 2.84, CI = ) and 3435T allele (p = 0.04, OR = 2.22, CI = ) were associated with higher warfarin dose. Using this model, individuals with the genotype 3435TT (55.8 ± 38.3 mg/ week) showed a tendency towards higher doses than 3435CT and 3435CC carriers (44.0 ± 28.1 mg/week) (p = 0.056). It is worth noting that 66 % of individuals with doses C70 mg possessed a 3435TT genotype and only 41 % of the group with doses \70 mg showed this genotype. The polymorphism UGT1A1(TA) n was associated with warfarin dose in a dominant inheritance model [(7/7? 7/ 6) 9 6/6] (p = 0.029, OR = 1.83, CI = ). The 7/7 allele frequency was higher in users with C70 mg and the UGT1A1(TA) 7 allele was positively correlated with the warfarin dose (r 2 = 0.041, p = 0.037). The UGT1A1(TA) 5 allele was not found in this population and the UGT1A1(TA) 8 allele was present in a single individual in heterozygous form. For classification II, MDR1 3435TT genotype (p = 0.026, OR = 3.28, CI = ) and the use of antagonists (p = 0.003, OR = 12.60, CI = ) were independently associated with warfarin dose (Hosmer and Lemeshow = 0.889) in multivariate logistic regression analysis. 4 Discussion This study revealed a wide variation in warfarin dose in patients receiving anticoagulant therapy for the prevention of VTE. The results suggest that the dose variability could be influenced by age, BMI, use of warfarin antagonists, and CYP2C9*2, CYP2C9*3, VKORC1-1639G[A, MDR1 3435C[T, APOE *e4, and UGT1A1(TA) 7 genetic polymorphisms. A particular feature of this study is the large percentage of individuals requiring a warfarin dose C70 mg, which was associated with a higher BMI, the use of warfarin antagonists, and the presence of VKORC1, MDR1, and UGT1A1(TA) polymorphisms. Even though gender is not a risk factor for VTE [20, 21], the larger number of women in the group can be attributed to risk factors unique to this gender, such as contraceptives, hormone replacement therapy, pregnancy, and postpartum susceptibility [21]. The influence of age on decreased warfarin dose could be due to decreasing clearance of R- and S-warfarin that occurs at % per year [22], increasing the risk of bleeding. Furthermore, we observed that the dose increases with increased BMI [3, 8, 23]. Although they are relatively young (42.3C ± C14.5 years), 65 % of the subjects have a BMI [25 kg/m 2, which can lead to problems with anticoagulation in this group [21]. The phenomenon of warfarin resistance [14, 15] (i.e., requiring [70 mg/week to obtain an INR in the therapeutic range) was observed in 20.7 % of the subjects examined in this study, notably in an individual who required 245 mg/ week. This percentage is much higher than expected since, according to Sinxadi and Blockman [14], the frequency of warfarin resistance is \1 %. Although noncompliance was an exclusion factor in the selection of the subjects, the use of antagonists observed in our population could explain the differences between our results and those observed by Sinxadi and Blockman. According to classifications used by Watzka et al. [24], 5.2 % of our population showed complete resistance to warfarin, i.e., they needed a high dose of warfarin without achieving a stable INR. In addition, 15.5 % of our population showed partial resistance to warfarin, and required a high dose to achieve a stable INR. We observed that the allele frequencies for CYP2C9*2 and CYP2C9*3 are similar to data for the Brazilian

7 Genetic Polymorphisms and Warfarin Dose population [8, 25]. The dose reductions of 23 and 48 %, for heterozygous and homozygous alleles of CYP2C9*2 and/or CYP2C9*3, respectively, indicate that the presence of variant alleles is associated with lower warfarin doses. To evaluate the relationship between these genotypes and warfarin dose, we excluded individuals requiring C70 mg. The lack of a difference in that group may be due to six group members with alleles CYP2C9*2 and CYP2C9*3, who also possess other important characteristics that may influence the dose, such as age (32.3 ± 3.8 year), BMI (28.1 ± 5.0), genotype (67 % with VKORC1-1639GG; 50 % with MDR1 3435TT), use of antagonists (50 %), and prevalence of the e4 allele (16 %). This highlights the dependence of inter-individual variation in warfarin dose on numerous genetic and non-genetic factors [6, 26, 27]. As previously reported [3, 6 8, 27], the VKORC1-1639G[A polymorphism was one of the main predictors of reduced warfarin dose. The -1639G allele showed an association with warfarin doses [70 mg/week, and the -1639A allele with doses B21 mg. The low frequency of the -1639AA genotype in this population could have affected the relationship between this polymorphism and warfarin dose. The allelic frequencies of -1639G (0.75) and -1639A (0.25) differed from published data for the Brazilian population [8, 23]. This was expected, because the -1639A allele frequency varies among ethnic groups, with 0.38 in Whites [8, 23, 26] and 0.22 in Blacks [23] and the frequency of these groups varies in each Brazilian geographic region. The decrease in warfarin dose for patients with at least one APOE e4 allele agrees with a study by Sconce et al. [28] in a Caucasian population. APOE e4 carriers also required lower maintenance doses of acenocoumarol [5]. However, Kohnke et al. [29] suggest that e4 carriers have enhanced liver uptake of vitamin K and need higher warfarin doses to compensate for this uptake [29]. The impact of the isoforms of APOE on the warfarin dose is related to its frequency in each population, especially the e4 allele, which ranges from in African Americans to 0.02 in Indians [11]. In this study, the e4 allele frequency was We observed that the MDR1 3435TT genotype is associated with higher warfarin doses ([70 mg), in accordance with data previously presented by our group in a study with a smaller sample size [12]. The mechanism for this finding is unknown. There is evidence that transport of warfarin from the liver into the bile is mediated by MDR1 [5]. Elimination is predominantly renal, but there is also evidence for hepatic transport via MDR1 [30]. Based on this, it can be suggested that the variant 3435T is involved in excretion of warfarin, and thus, the need for higher doses of the drug. A recent study found no statistical difference in the warfarin dose between 196 individuals of different ethnicities with 3435CC, 3435CT, and 3435TT genotypes of the MDR1 gene. However, this result cannot be compared to the results of our work. First, Kim et al. [31] considered warfarin resistance in the patients taking more than warfarin 6 mg/day (42 mg/week) to reach an INR of Based on the literature, individuals are classified as warfarin resistant who require a dose higher than 70 mg/ week to achieve an INR in the therapeutic range or who never achieve a therapeutic INR with this dose [14, 15, 24]. Moreover, Kim et al. reported dosages much lower (maximum of 62.9 ± 19.6 mg/week) compared with those observed in our group resistant to warfarin (mean ± 37.9 mg). However, although there was no difference in the dose of warfarin between the groups, they observed that African American individuals with T3435T genotypes needed the largest dose of warfarin. This result contributes to our statement about the influence of polymorphism 3435TT on the increased warfarin dose; mainly through the influence of African ethnicity on the population of this country [31]. A more recent study in the Egyptian population showed that subjects with the MDR1 3435TT/EPHX1 139RH, 139RR/ PZ-13AA genotype showed significant increase in warfarin dose/week when compared with MDR1 3435CC/ EPHX1 139RH,RR/ PZ-13AA subjects (p = 0.014) [32]. This result contributes to indicate that polymorphisms in genes encoding drug transporters have an influence on the dose of warfarin [32]. This is the first study to evaluate a gene involved in phase II metabolism of warfarin. We observed that the UGT1A1(TA) 7 allele was associated with higher warfarin doses. The allele 7 frequency was similar to that reported for Caucasian European and Brazilian populations, and lower than African, ranging from 0.32 to 0.60 [33, 34]. Recent studies have shown that glucuronidation is associated with the inactivation and clearance of warfarin, which may represent an important step in understanding warfarin metabolism. The UGT1A1(TA) 7 polymorphism may be a genetic determinant of the requirement for high doses of warfarin. It is estimated that the glucuronidation activity of the UGT1A1 enzyme is reduced by 25 and 70 %, in heterozygotes and homozygotes for the (TA)7 allele, respectively [13]. It is important to consider that several drugs can induce or inhibit the activity of UGT1A1 [34], and that 69 % of our population had used one or more drugs associated with warfarin interference. This interference from other medications could be also applicable to MDR1. Evidence of the influence of APOE, MDR1, and UGT1A1 on the response to warfarin have emerged in recent studies, even in small groups or in specific populations. It is known that a combination of genetic CYP2C9*2, CYP2C9*3, VKORC1-1639G[A and environmental factors explains only two thirds of the interindividual variability in warfarin dosage requirement. One

8 V. C. Oliveira Almeida et al. third goes without well established explanations [35]. This understanding will be useful to determine the cause of unusual therapeutic responses to warfarin therapy. When evaluated using a multivariate analysis, we observed that VKORC1 and APOE polymorphisms, besides age and BMI, are the variables independently associated with warfarin dose requirements, and should be further investigated in order to determine the optimal dose needed to maintain a stable INR. Furthermore, warfarin resistance is independently affected by MDR1 polymorphisms and use of antagonists. The frequency of polymorphisms varies among different ethnic groups and their impact depends critically on the genetic makeup of each population [26, 33, 36]. Although our study has a limitation in that some genotype subgroups had small numbers of carriers, this study is important because it has shown that genetic and non-genetic factors should be investigated in miscegenated and diversified populations such as those found in Brazil. Because of the heterogeneity and extensive admixture of the Brazilian population, extrapolation of data derived from other ethnic groups is clearly not applicable to the majority of Brazilians. Besides, specific environmental factors can have very particular effects on anticoagulant therapy. Therefore, a personalized strategy for initiation of warfarin therapy must take into account inter- and intra-ethnic/racial/environmental diversity and its complex interplay [37]. There have been no previous studies on warfarin resistance in the Brazilian population, and the frequency of resistance may be different from that reported in other populations. Characteristics of the subjects in two other Brazilian studies, in terms of age, sex, dose variability, and use of antagonists, differ from the population in the current study [8, 23], and those studies are therefore unsuitable for evaluating high requirements for warfarin and warfarin resistance. Moreover, those two studies did not use cutoffs for warfarin dose, and all individuals included in the study were stabilized. 5 Conclusions In conclusion, this study investigated the impact of clinical factors and CYP2C9*2, CYP2C9*3, VKORC1-1639G[A, MDR1 3435C[T, APOE *e4, and UGT1A1(TA) n polymorphisms on the warfarin dose in Brazilian patients taking anticoagulants to prevent venous thromboembolism. This study is important because the Brazilian population is highly miscegenated, and data obtained in other populations do not necessarily apply to this population. The CYP2C9*2, CYP2C9*3, VKORC1-1639A, and APOE *e4, polymorphisms were associated with lower warfarin doses. MDR1 3435T and UGT1A1(TA) 7 polymorphisms are associated with a requirement for higher doses. To our knowledge, this is the first study to evaluate APOE *e4 and UGT1A1(TA) n genotypes in the Brazilian population, and the association of these genotypes with warfarin dose required. The new information presented here on the importance of the UGT1A1 polymorphism contributes greatly to the investigation of metabolic pathways involved in the detoxification and excretion of warfarin, and the mechanisms of drug interactions. This is also the first study to evaluate warfarin resistance in Brazilian individuals. An important feature of this study is the large percentage of individuals requiring a warfarin dose C70 mg (i.e., with warfarin resistance), which was associated with a higher BMI, the use of warfarin antagonists, and the presence of VKORC1-1639G, MDR1 3435T, and UGT1A1(TA) 7 alleles. The high frequency of warfarin resistance found was surprising and highlights the need to understand the causal factors for this high frequency of resistance in the Brazilian population. Such understanding would allow personalized treatment of Brazilian VTE patients with warfarin that achieves stable INR and avoids adverse reactions. Acknowledgments The current study was supported by the Universidade Federal de Minas Gerais and Instituto Hermes Pardini. Disclosure of Conflicts of interest V.C. Oliveira Almeida, D.D. Ribeiro, K.B. Gomes, and A.L. Brunialti Godard have no conflicts of interest that are directly related to this study. References 1. Jorgensen AL, FitzGerald RJ, Oyee J, Pirmohamed M, Williamson PR. Influence of CYP2C9 and VKORC1 on patient response to warfarin: a systemic review and meta-analysis. PLoS One. 2012;7:e doi: /journal.pone The International Warfarin Pharmacogenetics Consortium. Estimation of the warfarin dose with clinical and pharmacogenetic Data. N Engl J Med. 2009;360: Johnson JA, Gong L, Whirl-Carrillo M, Gage BF, Scott SA, et al. Clinical Pharmacogenetics implementation consortium guidelines for CYP2C9 and VKORC1 genotypes and warfarin dosing. Clin Pharmacol Ther. 2011;90: Kangelaris KN, Stephen B, Nussbaum RL, Garcia DA, Tice JA. Genetic testing before anticoagulation? A systematic review of pharmacogenetic dosing of warfarin. J Gen Intern Med. 2009;24: Wadelius M, Chen LY, Eriksson N, Bumpstead S, Ghori J, et al. Association of warfarin dose with genes involved in its action and metabolism. Hum Genet. 2007;121: Wadelius M, Chen LY, Lindh JC, Eriksson N, Ghori MJ, et al. The largest prospective warfarin-treated cohort supports genetic forecasting. Blood. 2009;113: D Andrea G, D Ambrosio RL, Perna PD, Chetta M, Santacroce R, et al. A polymorphism in the VKORC1 gene is associated with

9 Genetic Polymorphisms and Warfarin Dose an interindividual variability in the dose-anticoagulant effect of warfarin. Blood. 2005;105: Botton MR, Bandinelli E, Rohde LE, Amon LC, Hutz MH. Influence of genetic, biological and pharmacological factors on warfarin dose in a Southern Brazilian population of European ancestry. Br J Clin Pharmacol. 2011;72: Kohnke H, Scordo MG, Pengo V, Padrini R, Wadelius M. Apolipoprotein E (APOE) and warfarin dosing in an Italian population. Eur J Clin Pharmacol. 2005;61: Cavallari LH, Butler C, Langaee TY, Wardak N, Patel SR, et al. Association of apolipoprotein e genotype with duration of time to achieve a stable warfarin dose in african-american patients. Pharmacotherapy. 2011;31: Kimmel SE, Christie J, Kealey C, Chen Z, Price M, et al. Apolipoprotein E genotype and warfarin dosing among Caucasians and African Americans. Pharmacogenomics J. 2008;8: Almeida VCO, Ferreira ACS, Ribeiro DD, Gomes KBB, Fernandes APSM, et al. Association of the C3435T polymorphism of the MDR1 gene and therapeutic doses of warfarin in thrombophilic patients. J Thromb Haemost. 2011;9: Bratton SM, Mosher CM, Khallouki F, Finel M, Court MH, et al. Analysis of R- and S-hydroxywarfarin glucuronidation catalyzed by human liver microsomes and recombinant UDP-glucuronosyltransferases. J Pharmacol Exp Ther. 2012;340: Sinxadi M, Blockman M. Warfarin resistance. Cardiovasc J Afr. 2008;19: Loebstein R, Dvoskin I, Halkin H, Vecsler M, Lubetsky A, et al. A coding VKORC1 Asp36Tyr polymorphism predisposes to warfarin resistance. Blood. 2007;109: Harrington DJ, Gorska R, Wheeler R, Davidson S, Murden S, et al. Pharmacodynamic resistance to warfarin is associated with nucleotide substitutions in VKORC1. J Thromb Haemost. 2008; 6: Micromedex. Drug-Reax System (2010) Internet database. Available at: capes.gov.br/micromedex2/librarian/pfdefaultactionid/evidenc expert.showdruginteractionsresults. Accessed 17 June Sakaeda T, Nakamura T, Horinouchi M, Kakumoto M, Ohmoto N, et al. MDR1 genotype-related pharmacokinetics of digoxin after single oral administration in healthy Japanese subjects. Pharm Res. 2001;18: Beutler E, Gelbart T, Demina A. Racial variability in the UDPglucuronosultransferase 1 (UGT1A1) promoter: a balanced polymorphism for regulation of bilirubin metabolism. Proc Natl Acad Sci. 1998;95: White RH. The epidemiology of venous thromboembolism. Circulation. 2003;107: Cushman M. Epidemiology and risk factors for venous thrombosis. Semin Hematol. 2007;44: Jensen BP, Chin PK, Roberts RL, Begg EJ. Influence of adult age on the total and free clearance and protein binding of (R)- and (S)-warfarin. Br J Clin Pharmacol. 2012;74: Perini JA, Struchiner CJ, Silva-Assunção E, Santana IS, Rangel F, et al. Pharmacogenetics of warfarin: development of a dosing algorithm for brazilian patients. Clin Pharmacol Ther. 2008;84: Watzka M, Geisen C, Bevans CG, Sittinger K, Spohn G, et al. Thirteen novel VKORC1 mutations associated with oral anticoagulant resistance: insights into improved patient diagnosis and treatment. J Thromb Haemost. 2011;9: Vianna JR, Perini JA, Rondinelli E, Suarez-Kurtz G. CYP2C9 genotypes and the pharmacokinetics of tenoxicam in Brazilians. Clin Pharmacol Ther. 2004;76: Limdi NA, Wadelius M, Cavallari L, Eriksson N, Crawford DC, International Warfarin Pharmacogenetics Consortium, et al. Warfarin pharmacogenetics: a single VKORC1 polymorphism is predictive of dose across 3 racial groups. Blood. 2010;115: Daly AK. Pharmacogenomics of anticoagulants: steps toward personal dosage. Genome Med. 2009;1: Sconce EA, Daly AK, Khan TI, Wynne HA, Kamali F. APOE genotype makes a small contribution to warfarin dose requirements. Pharmacogenet Genomics. 2006;16: Kohnke H, Sorlin K, Granath G, Wadelius M. Warfarin dose related to apolipoprotein E (APOE) genotype. Eur J Clin Pharmacol. 2005;61: The Pharmacogenomics Knowledgebase - PharmGKB (homepage na Internet). Stanford: Stanford University; (atualizada em 23 de maio de 2013; acesso em 20 de junho de 2013). Clinical PGx. Disponível em Kim Y, Smith A, Wu AH. C3435T polymorphism of MDR1 gene with warfarin resistance. Clin Chim Acta. 2013;425: Issac MS, El-Nahid MS, Wissa MY. Is there a role for MDR1, EPHX1 and protein Z gene variants in modulation of warfarin dosage? A study on a cohort of the Egyptian population. Mol Diagn Ther. 2014;18(1): Marques SC, Ikediobi ON. The clinical application of UGT1A1 pharmacogenetic testing: Gene environment interactions. Hum Genomics. 2010;4: Fertrin KY, Gonçalves MS, Saad ST, Costa FF. Frequencies of UDP- glucuronosyltransferase 1 (UGT1A1) gene promoter polymorphisms among distinct ethnic groups from Brazil. Am J Med Genet. 2002;108: Stojiljkovic M, Patrinos GP, Pavlovic S. Clinical applicability of sequence variations in genes related to drug metabolism. Curr Drug Metab. 2011;12(5): Santos PCJL, Soares RAG, Santos DBG, Nascimento RM, Coelho GL. CYP2C19 and ABCB1 gene polymorphisms are differently distributed according to ethnicity in the Brazilian general population. BMC Med Genet. 2011;12: Suarez-Kurtz G. Population diversity and the performance of warfarin dosing algorithms. Br J Clin Pharmacol. 2011;72:451 3.