Warfarin Is Used As An Oral Anticoagulant Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Pharmacogenetics plays an important role in the safety and effectiveness of oral anticoagulants. Genetic polymorphism in warfarin metabolism affects warfarin dosage requirements. Two common allelic variants in the CYP2C9 gene, CYP2C92 and CYP2C93, encode enzymes that are, about 12% and 5% respectively as efficient as the wild-type enzyme that hydroxylates S warfarin (Haining et al., 1996; Rettie et al., 1994). Carriers of variant alleles are associated with lower warfarin dosage requirements and an increased risk of bleeding, particularly at the initiation of warfarin therapy (Aithal et al., 1999). Patients who are homozygous for CYP2C933 require the lowest doses.

Recently, the gene that encodes vitamin K epoxide reductase (termed VKORC1), the target enzyme for warfarin has been cloned and non-synonymous mutations have been found in warfarin resistant patients (Li et al., 2004; Rost et al., 2004). It was reported that the patients with varying degrees of warfarin resistance were carries mutations at least in one copy of the gene that encodes vitamin K epoxide reductase complex 1 (VKORC1). More recent findings showed that VKORC1 genetic variation have a greater impact than CYP2C9 genetic variation on warfarin dose variance (Aquilante et al., 2006). Haplotype analyses have shown that most of the noncoding single nucleotide polymorphisms are in strong linkage disequilibrium. Based on haplotypes, individual may be divided into two groups, haplotypes, A (H1 and H2) and B (H7, H8 and H9), which are associated with lower and higher warfarin dose requirements, respectively (Sconce et al., 2005).

In Caucasian and Asian populations, genotype predicts 25% of the variability in warfarin dose (Yuan et al., 2005). In India a retrospective study done in Andhra Pradesh population reported that the clinical and genetic factors predict 61% of the variability in warfarin dose requirements (Pavani et al., 2012), In Malaysian Indians it was found that the mean daily dose of warfarin was significantly higher in Indians compared with the Chinese and Malay patients (Gan et al., 2011). These findings raise the possibility that the CYP2C9 and VKORC1 genetic variants may alter the dose requirement in Tamilian population.

To the best of our knowledge the study has been conducted first time in Tamilian population to find out the association of genetic polymorphism on warfarin dose requirement. Our study aims to find out the influence of CYP2C9 genetic polymorphisms on warfarin metabolism, plasma warfarin levels and to find out the effect of CYP2C9 and VKORC1 genetic polymorphism on warfarin dose requirement in patients taking warfarin.

Materials and methods

Study subjects

The study cohort consisted of out- patients receiving warfarin maintenance therapy at the cardiology clinic and cardio thoracic and vascular surgery clinics in the Jawaharlal Institute of Postgraduate Medical Education and Research (JIPMER) hospital, Pondicherry. All patients received anticoagulation treatment with warfarin to achieve an INR in the target range of 2 to 3. Patients of age group 18-65 years and of either gender were recruited for the study. Their nativity as Tamilian was assessed by their status based on family history of three generations in Tamil Nadu and Pondicherry and speaking Tamil as their mother tongue. Written informed consent was obtained from all the patients included in this study. Institutional ethics committee approved this study.

Patients with liver or renal dysfunction or on treatment with drugs that may be CYP2C9 inducers / inhibitors or pregnant and lactating women, smokers and alcoholics were excluded. Patients' demographic details such as age, sex, height, weight, body mass index, duration of warfarin therapy, hypertension and diabetic status were obtained from patients' case records. Patients with three consecutive measurements of INR value between 2 and 3.5, three months after initiation of therapy were included for the study.

2.2 Genotyping for CYP2C9 and VKORC1 -1639G>A

Five milliliters of venous blood were collected from the study participants for genotyping. DNA was extracted from the stored cellular fraction by using phenol-chloroform extraction procedure. Genotyping of CYP2C9 and VKORC1 were carried out in real-time thermo cycler (7300 Applied Biosystems; Life Technologies Corporation, Carlsbad, CA, USA) using TaqMan SNP genotyping assays (VKORC1 (rs9923231) assay ID: C__30996661_30, CYP2C92 (rs1799853) (assay by design), CYP2C93 (rs1057910) assay ID: C_27104892_10). The PCR was carried out in duplicate in a 20-µL final volume that contained 10 µL of TaqMan universal PCR master mix (2x), 0.5 µL of 20x working stock of SNP genotyping assay and 4.5 µL of genomic DNA diluted in DNAase free water and 5µL of MilliQ water (Millipore Corporate Headquarters, Billerica, MA, USA). The thermocycler conditions included one cycle at 50°C for 2 min; one cycle at 95°C for 10 min to activate the AmpliTaq Gold polymerase followed by 40 cycles of denaturation at 92°C for 15 sec and annealing/extension at 60°C for 1 min. The allelic discrimination analysis was performed using 7300 SDS software version 1.3.1.

Plasma warfarin determination

The plasma warfarin and 7 hydroxy warfarin were measured by HPLC method with modification from previously published work (Kulkarni et al., 2008).

2.3.1. Blood samples and plasma preparation for:

Venous blood (5 ml) was collected into EDTA tubes from all patients 12 h after the last dose of warfarin. The plasma was separated by centrifugation of blood samples at 3000 rpm for 10 minutes and stored at

70-C until analysis was done.

2.3.2. Materials and reagents fore HPLC assay:

Warfarin pure powder, 7 hydroxy warfarin and internal standard carbamazepine were obtained from Sigma-Aldrich, St. Louis, USA, All organic solvents used were HPLC grade and purchased from Merck specialities Pvt Ltd, Worli, Mumbai, India. Potassium di hydrogen orthophosphate and Di Potassium hydrogen orthophosphate were obtained from S.D.fine- chem Ltd, Mumbai, India. BOND ELUT- C18, 100 mg, 3 ml solid phase extraction cartridges was obtained from Varian, Inc.,

2.3.3. Standard preparation:

Stem solution of warfarin, 7 hydroxy warfarin and internal standard carbamazepine was prepared at 1 mg/ml in methanol. A series of six standard solutions of warfarin and, 7 hydroxy warfarin were prepared in drug free human plasma. The standard plasma solution contained 0.05, 0.1, 0.5, 1.0, 2.5 and 5.0 µg/ml of warfarin and 7 hydroxy warfarin, respectively. Plasma standards and QC samples were aliquoted, stored and treated the same way as patient blood samples. The solutions for the standard curve were freshly prepared before the analysis. For recovery estimation of the six analytes standard water solutions were prepared by diluting the standard stock solution in MilliQ water to serve as 100% control.

2.3.4. Extraction procedure

Solid Phase extraction was used to extract the drug from plasma samples. Ten microliters of 1mg/ml internal standard (Carbamazepine) was added in 1 milliliter standard, QC and patient plasma samples. C18 cartridges were used for extraction, briefly, the cartridges were conditioned before adding plasma by using 2 ml of 1% methanol (pH 2.8 adjusted with orthophosphoric acid). The plasma samples were rigorously vortexed and added in the conditioned columns. The warfarin and 7 hydroxy warfarin retained in the column were eluted with 2 ml of acetonitrile. The organic phase was transferred to fresh glass tubes and evaporated to dryness under nitrogen gas in a water bath at 60° C in sample evaporator. The samples were reconstituted with 200 µl of MilliQ water. Fifty microliter was injected into the HPLC.

2.3.5. Chromatography

The mobile phase consisted of isopropanol and potassium phosphate buffer (di potassium hydrogen phosphate pH 7.0 adjusted with potassium di hydrogen orthophosphate). The HPLC system consisted of a Shimadzu LC-10AD VP solvent delivery module, Shimadzu SPD-10A VP UV-VIS detector, 100 µl injection loop and Hypersil ODS column 4.6 mm, 5 mm particle size. The flow rate was maintained at 1 ml/min. The analytes were detected at 308 nm, with absorbance set at 0.005 Aufs. Separation was performed on a C18 column (Phenomenex, 150-4.6 mm, 5 µm)., warfarin, 7 hydroxy warfarin and carbamazepine had the retention time 3.4 min, 2.8 min and 8.1 min, respectively. The mean recoveries of compounds were consistent, at >88% for 7 hydroxy warfarin and >93% for warfarin. The precision and reproducibility of the assay was estimated and was <5.37% for inter-day and <6.90% for intra-day at the concentration 0.1, 0.5, 1.0, 2.5 and 5.0 µg/ml. All the chromatograms were analyzed by using the software CLASS-VP version 6.14 SP2.

Statistical analysis

GraphPad Instat® version 3.06 (San Diego, USA) and IBM® SPSS® Statistics (SPSS Inc., Chicago, IL, USA) were used for statistical analysis. The genotype frequencies were analyzed for Hardy- Weinberg equilibrium. The average maintenance dose between the genotype groups were compared by Kruskal Wallis test and Mann- Whitney test. Plasma levels of warfarin and 7 hydroxy warfarin between the genotype groups were compared by unpaired t test (Welch corrected) and warfarin metabolic ratio (MR) by the Mann- Whitney test. The genotype- phenotype relationship was evaluated using linear regression analysis. Stepwise multivariate regression analysis was used to find the influence of the independent variables (age, BMI, concomitant medications, comorbid conditions and genetic polymorphisms) on the dependent variable (logarithmic transferred daily maintenance dose). P<0.05 was considered statistically significant.


The demographic details were obtained from patient case records (Table 1). The mean average daily dose of warfarin was calculated to be 4.88± 1.63 mg. Warfarin was prescribed mostly for patients with rheumatic heart disease (mitral stenosis) (65.36%). In our study, the allele frequencies of CYP2C91 (91.6%), CYP2C92 (2.5%), CYP2C93 (5.9%), were consistent with those reported for the study population. The G and A allele frequency of VKORC1 were found to be 92.4% and 7.6%, respectively. The genotype frequencies of these variants were found to be in Hardy- Weinberg equilibrium. The daily maintenance dose in the patients with any variant genotype was significantly lower than the normal genotype (Table 2). The normal genotypes of CYP2C9 and VKORC1 was found to be similar with the maintenance dose of warfarin (5.2 mg/day for CYP2C911 and 5.1 mg/day for VKORC1 GG). Patients having two defective alleles in CYP2C9 gene required lower dose (2.5 mg/day). The homozygote variant in VKORC1 (AA) was found in only one patient and the daily maintenance dose was lower (3 mg/day) than the other genotype group.

The effect of combination of variant genotypes on maintenance dose was compared with the normal genotype combinations (Table 3). The patients carrying both the variant genotypes were required lower dose as compared to the any one variant and normal genotype carriers. Patients with both normal genotypes and carrying both variant genotypes were found to be 75.8% and 6.7% (12 patients), respectively.

Patients with two variant alleles or one variant allele in CYP2C9 (12 and 13) had lower 7 hydroxy warfarin plasma levels, while 11 carriers had the highest (Table 4). There was a significant difference observed in metabolic ratios between the patients with 11 and 12 or 13 (p< 0.05). In univariate and multivariate regression analysis it was observed that age (p<0.05), daily maintenance dose (p<0.05), CYP2C92 and CYP2C93 genotype (p<0.0001) were significantly influenced the warfarin metabolic ratio. This result indicates that the CYP2C9 genetic polymorphism influences warfarin metabolic ratio.

Linear regression analysis revealed a positive correlation of plasma warfarin concentration and daily warfarin dose (r2 = 0.24, p=0.016), also a positive linearity was observed between plasma warfarin and 7 hydroxy warfarin (r2=0.47, p<0.0001). Due to skewed distribution the daily dose was converted into logarithmic transformed dose and was taken into univariate and multivariate analysis, univariate analysis revealed that age significantly influence the daily dose (p<0.05), body weight (p<0.0001), height (p<0.05) and body mass index (p<0.0001), genetic polymorphism in the two genes CYP2C92 (p<0.0001), CYP2C93 (p<0.0001), VKORC1-1639G>A (p<0.05) significantly influenced the daily dose. Multivariate stepwise regression analysis was performed by adding all the significant factors from univariate analysis (table 5). The multivariate analysis revealed that the combined effect of age, weight and genotype contributes 35.1% dose variation. Body weight alone significantly reduces 13.2% of the maintenance dose. In our study 8 (4.4%) patients were reported to have bleeding risk. But there was no significant association found with the genetic polymorphisms and other factors.


The single nucleotide polymorphisms in these genes widely varied between the populations (Margaglione et al., 2000). There was significant difference observed in the different ethnic population receiving coumarin anti-coagulants. It is well documented that the two alleles CYP2C92 (rs1799852) and CYP2C93 (rs1057910) were significantly associated with decreased warfarin dose requirements (warfarin sensitivity) and higher susceptibility to overdose(Margaglione et al., 2000). A direct association of CYP2C9 genotype anticoagulation status and bleeding was first reported by Higashi et al (Higashi et al., 2002). Also the allele frequencies of CYP2C92 and CYP2C93 diverge considerably among different ethnic groups (Stubbins et al., 1996). The frequencies of CYP2C91, 2 and 3 in the south Indian population were 0.88 (95% CI 0.85- 0.91), 0.04 (95% CI 0.02-0.06) and 0.08 (95% CI 0.06-0.11), respectively (Jose et al., 2005). In the present study the allele and genotype frequencies of CYP2C9 was in agreement with the previous study.

To the best of our knowledge, this study investigates the genotype- phenotype correlation of CYP2C9 and warfarin for the first time in Tamilian population. The study also investigates the influence of genetic polymorphism in two genes CYP2C9 and VKORC1 on warfarin maintenance dose. Further, we have investigated the influence of CYP2C9 genetic polymorphisms on warfarin metabolic ratio. Higher 7 hydroxy warfarin level and lower metabolic ratio were seen in patients with normative alleles. Whereas, lower 7 hydroxy warfarin and higher metabolic ratio were observed in patients with any one defective allele.

There was no significant difference observed between CYP2C912 and CYP2C913 carriers. The range of mean plasma concentration in the previous studies (Bentley et al., 1986; Routledge et al., 1998) was reported as 0.8 mg to 2.4 mg/litter. In our study the mean plasma concentration was found to be 2.81 mg/liter, and slightly higher than the normal range. A pilot study was conducted previously in 25 patients in the North Indians to measure mean total plasma warfarin levels and it was found to be 3.01±2.48 (SD) µg/ml. The mean 7 hydroxy warfarin level was found to be 0.20± 0.13 (SD) µg/ml. In our study the mean plasma warfarin was slightly lower (2.81±1.77 at (SD) µg/ml) and the mean plasma 7 hydroxy warfarin was higher (0.68± 0.69 (SD) µg/ml) than the previous study (Kulkarni et al., 2008). The higher levels of mean 7- hydroxy warfarin was observed possibly due to the lower frequency of variants in CYP2C9 gene in our population.

Previous studies reported that the contribution of CYP2C92 and CYP2C93 on warfarin dose determination were found to be 6.9% in Malaysians and 7.9% in Honkong Chinese population (Lal et al., 2008; Sandanaraj et al., 2009). According to the CYP2C9 genotype it was observed that the higher maintenance dose was administered for carriers of CYP2C911 wild type and carriers of variant genotype group received lower maintenance dose. In our study the CYP2C92 and CYP2C93 genotypes contributes 16.4% of dose variability. Possible reason is that the frequencies of CYP2C9 genotypes were more as compared to Chinese and Malaysian population. According to the VKORC1 genotype, the GG genotype group received higher dose than the carriers of variant genotype. But in our study only one patient was carrier of AA genotype and received 3.0 mg/ day warfarin dose.

In a recent study in Italians the carriers of VKORC1 1639 A allele had lower oral anticoagulant requirements (4.7, 3.7, 2.2 mg/day for GG, GA, AA genotype respectively; P < 0.0001), higher mean INR (2.7, 2.8, 2.9; P = 0.05) and a higher number of patients with a therapeutic range than GG genotype carriers (17 % vs. 0 % in GG genotype, P = 0.036) (Giansante et al., 2012). In a recent study conducted in Malaysia, the mean dose of warfarin for all patients was 3.7 mg, and the mean daily dose of warfarin was significantly higher in Indians compared with the Chinese and Malay patients, 4.9 versus 3.5 and 3.3 mg, respectively (Gan et al., 2011). The reason for the higher dose requirement in Indians was possibly by other factors. Our study was in agreement with these findings, the mean dose required in our population was 4.88 mg/day and our population falls under the intermediate dose group. The reason for this may be that, in our population the more influential VKORC1 -1639 G>A genetic variants are less frequent.

Previous studies have shown that the VKORC1 genetic polymorphisms were a better predictor than CYP2C9 genotype. In our study we have observed that the influence of CYP2C92, CYP2C93 (16.4%, P<0.0001) was higher than VKORC1 (3.3%, p<0.05). A possible reason for this may be the lower frequency of VKORC1 variants in the study population. The previous studies have shown that the daily warfarin dose of the study population does not show a Gaussian distribution and square root transformed dose were used (Sconce et al., 2005; Tham et al., 2006). However in another previous study it was found that the logarithmic transformation of the dose yielded a highly significant test for Gaussian distribution (Zhu et al., 2007). Similarly, in our study we have used the logarithmic dose for regression analysis.

Following the study by Sconce et al (Sconce et al., 2005) many other studies have conducted in different ethnic populations using multiple linear regression analysis technique to predict the warfarin maintenance dose (Anderson et al., 2007; Gage et al., 2008; Klein et al., 2009; Pavani et al., 2012; Perini et al., 2008; Schelleman et al., 2008; Wadelius et al., 2009). Those studies included covariates such as age, height, weight, VKORC1 -1639 G>A and CYP2C9 genotypes. The inclusion of height as a variable had a greater impact in the study by Sconce et al (Sconce et al., 2005). Tham et al (Tham et al., 2006) found that the age and weight were highly influential with the genetic factors. Weight based dosing of warfarin is more conventional approach, although, in our study we have found that weight alone shows 13.2% variability.

Previous studies (Puehringer et al., 2010; Schalekamp et al., 2004; Sconce et al., 2005) have explained that age was the second most important predictor and shows a greater variability in dosage of coumarin anticoagulants. In agreement with the previous studies (Gage et al., 2008; Kamali et al., 2004) we found that the dose requirement reduce with age and shows 3.6% variability and associated with significant reduction in daily maintenance dose.

Many studies have proposed algorithms for calculating the maintenance dose and the initial dose of oral anticoagulants using the multivariate statistical techniques (Anderson et al., 2007; Gage et al., 2008; Klein et al., 2009; Sconce et al., 2005; Tham et al., 2006; Zhu et al., 2007). In our study we were able to assign only 35.1% of observed inter-individual variability in warfarin dose with respect to the variable factors. Other than these specific factors additional genetic factors, and clinical factors that modify dose requirements and therapy. The limitation of our study was the strict exclusion criteria for not including the patients with interacting co-medication, comorbidities (liver and renal dysfunction), alcoholic and smokers. It was already known that these factors significantly influence the warfarin dose requirement but the relative contribution of genetic factors was the prime aim in our study.

A study in North Indian patients on treatment with acenocoumarol, explains 41.4% variability in dose requirement (Rathore et al., 2012). They included CYP4F2 and GGCX genetic polymorphisms in their algorithm other than CYP2C9 and VKORC1, In Andhra Pradesh population Pavani et al explained 61% variability in dose prediction based on their algorithm and the study included CYP4F2 genetic polymorphism as additional genetic predictor (Pavani et al., 2012).

Several other studies have been recently conducted to identify the unknown genetic factors influencing warfarin (Cooper et al., 2008; Takeuchi et al., 2009; Wadelius et al., 2009). It was found that only VKORC1 and CYP2C9 genetic variations were highly associated with the warfarin dose. Additionally only one SNP in the CYP4F2 gene was associated with 1-2% dose variability. Furthermore additional genetic, clinical and environmental factors may significantly contribute to improve our dosing model. The present study provided only the underlying clinical and genetic variation and their fraction of influence on warfarin daily maintenance dose in Tamilian population.


The genetic variation of CYP2C9 and VKORC1 genes were predictive factors of warfarin dose requirements in Tamilian population and the influence of CYP2C9 genetic variation on warfarin metabolism has been explained. The present study provided the basic information for developing a pharmacogenetic algorithm for predicting the initial dose of warfarin in Tamilian patients.


This research project was funded by Indian Council of Medical Research (ICMR), New Delhi, India and UMR-775, Bases Moléculaires de la réponse aux xénobiotiques, INSERM, Université Paris Descartes, Paris, France. (Ref. No. INDO/FRC/646/2010-IHD, Dated 10.01.2012).


Mrs. G.Saraswathi, Ms.S. Kalaivani, Mrs. Revathy, technical assistants are gratefully acknowledged.