Genotype distributions of CYP2C9 and VKORC1 in southern Thais and their association with warfarin maintenance dose in patients with cardiac surgery

After ethical approval for the study was obtained from the Ethics Committee of the Faculty of Medicine, Prince of Songkla University (reference No. EC.57- 345-29-8) and written informed consent from adults and the parents or guardians of children, blood specimens were collected from 210 healthy volunteers aged more than 15 years who lived in Songkhla and nearby provinces during 2011–2013. Genotyping in this set of volunteers was aimed to determine genotype distribution of the 4 polymorphisms in our population. Another set of samples were collected from patients aged 15 years or older who underwent heart valve replacement in the Cardiothoracic Surgery Unit, Department of Surgery, Songklanagarind Hospital, Prince of Songkla University, and received warfarin therapy postoperatively and attended the Warfarin Clinic during the years 2014–2015. Sample size calculation was calculated for primary outcome comparison (maintenance warfarin dose difference among genotypes), using a two-tailed independent sample model. Mean and standard deviation used for the sample size calculation were based on those reported in a previous study in Thais [6]. According to these calculations, a sample size of 90 cases would give a power of 0.80 to detect significant differences at 5 mg/week of maintenance warfarin dose. Data regarding warfarin dose, together with factors that may influence dose adjustment were also collected from the hospital information system and those recorded by the Warfarin Clinic at Songklanagarind Hospital.

Genomic DNA was isolated from peripheral blood leukocyte specimens using a QIAamp DNA Mini Kit (Qiagen, Hilden, Germany), following the manufacturer’s protocol. Two single nucleotide polymorphisms (SNPs) in CYP2C9 and 2 SNPs in VKORC1 were studied (Table 1). SNPs were selected according to previous reports that they have variations in Asians (http://www.1000genomes.org/Homo_sapiens), and other reports indicating an association between the polymorphisms and warfarin metabolism [7, 8, 9, 10].

Table 1

Studied single nucleotide polymorphisms (SNPs) of CYP2C9 and VKORC1

Genotyping was performed using TaqMan genotyping assays on an ABI Prism 7500 Fast Realtime PCR, ABI GeneAmp PCR system 7500, using an Applied Biosystems reaction system (ABI; Foster City, CA). The assay mixes (including unlabeled PCR primers, FAM, and VIC dye-labeled TaqMan MGB probes) of Assays-by-Design were designed and supported by ABI. The basic reaction contained 50 ng of genomic DNA, 10 ml of 2× TaqMan Genotyping Master Mix, 0.5 ml of 40× Assay Mix adjusted with Milli-Q water in a total volume of 20 ml. The primers and probes used in this study followed previous studies with some modifications [7]. The PCR conditions consisted of an initial step at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 60 s in a 96-well plate that included negative (no DNA template) and positive controls to ensure genotyping accuracy. The genotyping results were analyzed by ABI 7500 software, version 2.0.5, and random samples were selected for confirmation by direct sequencing. Quality control was set at a call rate more than 95% and an accuracy rate more than 99%.

Patients receiving warfarin in our institute are managed by a team working in the ‘Warfarin Clinic’ which is a cooperative association between clinical pharmacists and surgical staff. Practice guidelines and dose recommendations are provided to all team members. In our clinic, the target INR for patients with cardiac valve implants is 1.8–3.5. Warfarin therapy is started at 5 mg/day or 35 mg/week and adjusted every 1–2 weeks until the target INR is reached. The maintenance warfarin dose in this study was defined as a stable dose (<15% dose adjustment at the most recent dosage evaluations) that maintained INR within the target range for at least 3 consecutive visits. Time to achieve stable warfarin dose for the study was defined as time from the date of surgery to the first date of stable INR (3 consecutive INRs within desired range).

Body surface area was calculated according to the Du Bois formula [13]. Drugs that may influence warfarin pharmacokinetics were divided into 2 lists; list-1 included drugs that can potentially increase warfarin metabolism (thus leading to a decreased maintenance dose), and list-2 included drugs that can potentially decrease warfarin metabolism (thus requiring an increased dose) [14].

Statistical analysis of the agreement of genotype frequencies with Hardy–Weinberg equilibrium for each SNP were performed using chi-square test. Comparisons of warfarin maintenance doses among genotypes were performed using a Student t test. P < 0.05 was considered significant. All statistical calculations were performed with Stata (version 13.0; College Station, TX, USA).

The 210 healthy volunteers included 108 male and 102 female individuals with an average age of 45 years. On genotyping of the 210 healthy subjects in our population, minor allele frequencies (MAF) of the rs1799853 (T), rs1058910 (C), rs9923231 (A), and rs9934438 (T) were 0.00, 0.04, 0.73, and 0.73, respectively. We found that rs9923231 had complete linkage disequilibrium (LD) with rs9934438 (r²:1.0; D’1.0). Except for rs1799853, which had no variants, genotype distributions in all SNPs conformed to the Hardy–Weinberg equilibrium. Considering no sequence variants in the rs1799853 and complete LD between rs9923231 and rs9934438, only 2 SNPs, rs1059810 and rs9923231, were chosen to be analyzed with respect to warfarin maintenance dosages. For CYP2C9, the genotype frequency of heterozygous AC (CYP2C9*1*3) in the rs1958910 was 17/210 or 8.1%. The genotype distribution of VKORCI (rs9923231) in the 210 controls was 118:72:20 individuals or 56.2%:34.3%:9.5% for AA:AG:GG.

A total of 166 patients receiving warfarin during the period of study were eligible to participate in the study. Of the 12 excluded cases, 11 were excluded because the target INR could not be stably achieved and the other was excluded because of genotyping failure. Thus, ultimately 154 patients (59 male and 95 female) who had stably achieved the target INR were included in the analysis. The average age of these patients was 45 years (18-75 years), and their average body mass index (BMI) was 23.3 kg/m² (15.8-36.3kg/m²).

The MAFs of rs1059810 and rs9923231 in the patients were 0.03 and 0.31, respectively. Genotype distributions of the 2 SNPs are shown in Table 2. The warfarin doses that gave the target INRs ranged 7-70 mg/week with an average dose of 31 mg/week (±12.2 mg/week). Considering rs9923231, the average dose in patients harboring AA (23.8 mg/week) was significantly lower than that required for patients with GA (35.3 mg/week) or GG genotypes (47.3 mg/week) (P < 0.01 for all comparisons) (Figure 1). By contrast with rs9923231, the minor genotype of rs1057910 (AC) required a smaller dose of warfarin (20.7 mg/week) than the predominant genotype [13] (P < 0. 01). When rs1057910-AC was combined with rs9923231- AA, 4 patients with this combination required an even lower maintenance dose at an average of 15.8 mg/week (Table 3). Dose adjustment duration ranged 30-310 days with a median duration of 80 days. Considering VKORCI genotypes, median durations to achieve stable INR in AA, AG, and GG were 20, 160, and 342 days, respectively. Of our 154 patients analyzed, 76 (49.3%) required more than 90 days to adjust the dose to achieve stable INR, and the proportion of those with adjustment time >90 days significantly increased with the VKORCI genotypes (33.3% in AA, 60.3% in AG and 81.3% in GG, P < 0.01).

Figure 1

Table 2

Association between genetic polymorphisms and other factors and warfarin dose

Table 3

Median (95% confidence interval) of maintenance warfarin dose (mg/week) according to combined genotypes of CYP2C9 and VKORC1

Of 11 patients that were not included in the analysis because of unstable INR, 6 (55%) had an AA genotype of VKORCI. One patient harboring the CYP2C9*3*3 genotype who received less than 8 mg/week of warfarin was not included in this analysis because they had not reached the desired INR by the end of the study period.

When we analyzed the warfarin dose between groups with regard to other parameters, we found no significant association between dose and sex, body mass index, or food consumption (Table 1). A higher body surface area was associated with a significantly higher warfarin dose when dose was significantly lower in patients receiving List-1 medications known to interact with warfarin in the increased metabolism direction. We also observed that generally, patients aged >45 years required significantly smaller doses of warfarin.

Complications that were possibly related to the warfarin therapy were recorded in 38 patients (24.6%), most of which (34/38) were minor bleeding such as skin petechiae/ecchymosis, bleeding per gums, and menorrhagia. Four complications that required readmission or surgical treatment included 2 cases of valve dysfunction, a case of severe oral bleeding, and a case of cerebrovascular thromboembolism. There was no significant difference in complication rate between the 2 genotypes of CYP2C9 (23.5% in CYP2C9*1*1 and 30.0% in CYP2C9*1*3, P = 0.80), and neither was among VKORCI genotypes (25.1% in AA/AG and 12.5% in GG, P = 0.24). Incidence of prolonged prothrombin time (INR > 5.0) at least once during the follow-up period was significantly higher in the VKORCI-AA group (45/76, 59.2%), than in AG (22/64, 34.4%), or GG (3/16, 18.8%) groups (P < 0.01).