Vitamin K antagonists (VKAs: warfarin, acenocoumarol, phenprocoumon, phenindione) are prescribed commonly for treatment of venous thromboembolism (VTE), prevention of thromboembolic complications in atrial fibrillation (AF), and after mechanical heart valve (MHV) replacement. VKAs are highly effective but they have a narrow therapeutic index and they require routine monitoring of the International Normalized Ratio (INR). Over recent years, the oral anticoagulant (OAC) therapy scenario has changed as a result of the introduction of direct oral anticoagulants (DOAcs)[2,3] and improvements in OAC management with VKAs, including pharmacogenetic (PGx) studies, focused on identifying genetic determinants affecting VKA dose requirements.[4–6] The best-characterized genes involved in VKA PGx are vitamin K epoxide reductase complex subunit 1 (VKORC1) and cytochrome P450 2C9 (CYP2C9). The human VKORC1 gene (16p11.2) comprises 3 exons encoding the catalytic subunit of the vitamin K epoxide reductase complex, which is the key enzyme in the Vitamin K cycle.[7–10] A single nucleotide polymorphism (SNP) in the VKORC1 promoter (−1639G > A, rs9923231) results in a decreased transcription of the gene and has been strongly associated with warfarin dose requirements.[11,12] The CYP2C9 gene (10q24) encompasses 9 exons and it is highly polymorphic, as more than 60 variant alleles have been identified (http://www.cypalleles.ki.se, last access February 2016). CYP2C9 is one of the most abundant cytochrome P450 in the liver and it metabolizes approximately 15% of clinical drugs.[13,14] Allelic variants are missense, nonsense or frame shift variations, causing a reduced or a null enzyme activity. The most frequent variant alleles in Caucasian population, CYP2C9*2 (rs1799853) and CYP2C*3 (rs1057910), in the homozygous condition, reduce enzyme activity to 12% and 5%, respectively, compared to the wild-type genotype CYP2C9*1*1.[15–16] The U.S. Food and Drug Administration (FDA) updated the warfarin label, providing dose recommendations based on different combinations of VKORC1 and CYP2C9 genotypes, in 2007 and 2010. Guidelines for clinicians and genetic-based algorithms have been implemented by the International PGx Warfarin Consortium. Two recent randomized clinical trials (RTCs)[19,20] aimed to assess the effect of the PGx-guided initial drug dosing on improvement of Time in Therapeutic Range (TTR). The RTCs showed contradictory results because of the differences in the study design, and stimulated a considerable debate on this matter.[21–30] Although there is robust evidence of the association of genetic variants on dose requirement, the role of these genes on the clinical outcome (bleeding and thrombosis) during OAC therapy is controversial, as outlined in contradictory results reported in recent meta-analyses.[31,32] The aim of the present study is to evaluate potential associations between genotype VKORC1 and CYP2C9 and adverse events (hemorrhagic and/or thrombotic) during initiation and long-term VKA treatment, in a Caucasian population. Patient monitoring occurred in 2 specialized anticoagulation clinics. Furthermore, we aimed to determine if the concomitant prescription of other selected drugs (amiodarone, HMGCo-A reductase inhibitors [simvastatin], antiplatelet medication) affected the association between genotype and adverse events.
2 Materials and methods
2.1 Design overview
We performed a retrospective, matched case-control study to examine associations among CYP2C9*2, CYP2C9*3, VKORC1: c. −1639G
A polymorphisms, drug intake, and any hemorrhagic and/or thrombotic event, in oral anticoagulated patients. Cases and controls were enrolled and monitored in 2 Italian anticoagulation clinics (Anticoagulation Centre, Brescia and Haemostasis and Thrombosis Centre, Cremona) between 2009 and 2014. Both centers are affiliated with the Italian Federation of Anticoagulation Clinics (FCSA) and are placed in hospitals in the main city.
2.2 Patients: eligibility criteria
In order to achieve a cohort representative, as far as possible, of real life conditions, no explicit exclusion criteria were defined, except for age and Caucasian ethnicity.
Cases included patients receiving OAC therapy with the following characteristics:
- Age greater than 18 years
- Caucasian origin
- OAC therapy use for any condition (atrial fibrillation, AF; venous thromboembolism, VTE; implanted mechanical heart valves, MHV)
- History of an adverse event (thrombotic and/or ischemic) during therapy with VKAs.
- Adverse events are those indicated in the Italian FCSA guidelines:
- Major hemorrhages (cerebral bleeding; extra-cerebral bleeding in a critical area or organ; a decline in hemoglobin levels by 2 g/dL and/or requiring transfusion)
- Thromboembolic events (stroke; transient ischemic attack, TIA; myocardial acute infarction, IMA; venous thromboembolism, VTE, including deep vein thrombosis, DVT, and pulmonary embolism, PE).
Minor hemorrhagic events were excluded.
The control group consisted of 120 unrelated subjects who did not experienced any adverse event and were matched to cases for age, sex, clinical indication, and duration of anticoagulation.
2.3 Data source and genotyping
Electronic search was performed through software ParmaGTS (Instrumentation Laboratory, Bedford, MA) in Brescia Haemostasis Center and in Cremona Haemostasis and Thrombosis Centre through TAONET (EDP-Project, Bozen, Italy), used for the management, archiving, and referral of inpatients and outpatients to the clinic. In the Brescia Haemostasis Centre, we initially identified patients (N = 458) with a history of any adverse event occurring between 2009 and 2014. We excluded patients experiencing a minor adverse event, those who died from any cause, and patients who were not of Caucasian origin. We identified 196 patients with major adverse events. We then excluded patients who interrupted OAC (N = 92); did not measure INR as prescribed; or did not communicate the INR value, when measured in a different setting, on more than 3 occasions (N = 28). We obtained a total of 74 effective final cases. In order to achieve the predetermined statistical power for the study, a further 46 cases were enrolled from the Cremona Haemostasis and Thrombosis Center. All controls were selected from Brescia Anticoagulation Center. For both cases and controls, only patients with complete clinical information were selected (Fig. 1). Clinical data, including patient age, sex, indication for and duration of OAC therapy, concomitant medication, medical history, and INR values, were collected at the time of enrollment. In this phase of the study, we only collected information regarding the following drugs: amiodarone, HMGCo-A reductase inhibitors (simvastatin), antiplatelet medication. Venous blood samples were collected in tubes containing sodium citrate and genomic DNA was extracted by the automated nucleic acid extraction system MagCore HF 16 (RBC Bioscience, Taiwan) using the Whole Blood Kit and following the manufacturer's instructions. Genotyping was performed using the PGX-Thrombo StripAssay (ViennaLab Diagnostics, Vienna, Austria), with slight modification to the manufacturer's instructions. The process was performed using the automated instrument ProfiBlot T48 (Tecan, Männedorf, Switzerland), which allow simultaneous processing of 48 samples. The study protocol was reviewed and approved by the Ethics Committees of the participating hospitals. Written informed consent was obtained for all patients.
2.4 Statistical analysis
Sample size was estimated assuming a case-control design, with a control-case ratio of 1, a significance level of 5%, and a power of at least 80%. We assumed a mutation frequency of at least 14% and an odds ratio (OR) of at least 2.5. This resulted in a sample size of 118 subjects per group. This figure was rounded up to 120, for a total of 240 cases and controls, overall. Adverse events were not separated for thrombotic or hemorrhagic events. Estimation of OR for event occurrence was performed with conditional logistic models, to account for case/control matching and assuming the additive genetic effect for rs9923231 (VKORC1), rs1799853 (CYP2C9*2), rs1057910 (CYP2C9*3). In this model, each genotype is a distinct group; homozygous patients for major alleles (GG and *1*1) are the reference group. Data are reported as OR estimates and 95% confidence intervals. Similarly, time to event was modeled using marginal Cox models, with sandwich variance estimators to account for matching. The statistical significance level was set at P = 0.05.
3.1 Characteristic of the selected cohort
A total of 120 cases and 120 controls were enrolled in the study. Patient mean age was 76 ± 10 years, and 44% of subjects were female. The majority of the patients were prescribed VKAs for the following conditions: atrial fibrillation (N = 134, 55.8%), mechanic heart valve replacement (N = 56, 23.2%), and venous thromboembolism (N = 50, 20.8%). Mean follow-up time was 8 ± 6 years. There were no statistically significant differences in comorbidity between cases and controls, except for history of hypertension, which was slightly more prevalent among controls than among cases (87.5% vs 78.3%, respectively; P
0.047). There were no significant differences in the use of amiodarone, simvastatin, and antiplatelet agents and no significant differences in the TTR between the 2 groups of patients (Table 1).
3.2 Genetic polymorphisms and adverse events association
In 65% of the selected cases, major bleeding events were observed, and in 35%, a thromboembolic event was noted (Table 2). Out of 120 cases, an adverse event occurred in 114 patients (95%) during the maintenance phase of the therapy (over the first 3 months) and in 6 patients (5%) during the induction phase (data not shown). The frequencies of VKORC1 and CYP2C9 genotypes are shown in Table 3. All 3 VKORC1 genotypes were considered, whereas CYP2C9 genotypes were grouped into 2 categories: wild-type (*1*1) and carriers of at least 1 variant CYP2C9 allele (*1*M or *M*M). Genotypes were in Hardy–Weinberg equilibrium for both cases and controls. The effects of 3 genetic polymorphisms on adverse events were evaluated. The variant VKORC1: −1639 G
A and CYP2C9 variant alleles were significantly associated with any thrombotic and/or hemorrhagic adverse event. VKORC1: −1639 G
A was associated with a 1.5-fold increase in the risk of adverse events (OR 1.60, 95% CI interval 1.06–2.42; P
0.025). Each CYP2C9 variant allele was associated with a 2-fold increase of the same risk (OR 2.28, 95% CI interval 1.34–3.86; P
0.002). The combined VKORC1 and CYP2C9 genotype effect were associated with a 2-fold increase in adverse events (OR 2.37, 95% CI interval 1.37–4.10; P
0.002). To take into account several confounding factors, we applied survival models to assess the association between VKORC1 and CYP2C9 polymorphisms and primary outcomes. Adjusting for concomitant drugs (Amiodarone, Simvastatin, and antiplatelet medication), carriers of one of the analyzed SNPs had a significant increased hazard ratio for adverse events (Table 4). VKORC1: −1639 AA homozygous was associated with a 2-fold increased hazard ratio of an adverse event (HR 1.97, 95% CI interval 1.13–3.43; P
0.02) compared to wild-type (GG), whereas CYP2C9*2/*3 carriers had a 1.6-fold increased risk (HR 1.58, 95% CI interval 1.14–2.35; P
0.008). In CYP2C9*2/*3 carriers (HR 1.6, 95% CI interval 1.00–2.56; P
0.05), the same risk was observed for simultaneous use of antiplatelet medication (Table 4.)
Our findings suggest that patients undergoing long-term oral anticoagulation (OAC) therapy with VKAs have a high risk of thrombotic and/or major bleeding adverse events if they are carriers of 1 or more polymorphisms in the VKORC1 (rs9923231) and CYP2C9 (rs1799853 and rs1057910) genes. To the best of our knowledge, this is the first study on VKA pharmacogenetics showing an association between genetic variants and both bleeding and thrombotic events in a cohort of patients affected by a range of diseases (AF, VTE, MHV replacement) and experiencing adverse events during OAC therapy. Published data on the increased risk of side effects associated with genetic variants among VKA users is controversial, as highlighted by recently published guidelines concerning the appropriateness of genetic testing in warfarin therapy. A recent meta-analysis showed that only the CYP2C9*3 allele, but not the CYP2C9*2 allele, was associated with a significant risk of bleeding. However, another meta-analysis concluded that CYP2C9*2 and CYP2C9*3, but not VKORC1 genotypes, were associated with warfarin hemorrhagic complications. A recent case-control study showed that VKORC1 1173 C
T (rs9934438) and CYP2C9*2/*3 polymorphisms are not associated with increases in the risk of major bleeding in long-term warfarin users. Conversely, an extensive genetic analysis of an RTC has clearly showed the influence of VKORC1 and CYP2C9 variants on bleeding events in patients with nonvalvular AF and treated with warfarin. The patient cohort presents increased variability compared to other studies; nevertheless, our results corroborate previous findings of major bleeding and thrombotic risks in patients with CYP2C9 and VKORC1 variants. Tomek et al showed that anticoagulated Caucasian patients carriers of 3 variant alleles of the genes CYP2C9 and VKORC1 exhibited a significantly higher risk of major bleeding during the initiation and maintenance phases of warfarin therapy. Kawai et al showed that the CYP2C9*3 allele could double the risk of major bleeding among patients administered warfarin for 30 or more days. However, the authors did not find any association with VKORC1 variants and the study population largely consisted of Caucasians, although it also included Afro-Americans and others ethnic groups. Recent research identified the VKORC1 gene as a susceptibility factor for ischemic cerebrovascular disease in a Chinese subpopulation, and Wang et al demonstrated the involvement of VKORC1 in arterial vascular disease. Our results partially corroborate recent evidence on drug–drug interactions (DDIs). Santos et al showed that simultaneous use of warfarin and amiodarone was not associated with adverse events in chronic OAC therapy. Also, a previous study found that the combination of single antiplatelet therapy with an anticoagulant was associated with a significantly greater risk of bleeding. Sconce et al showed that simvastatin reduced the mean warfarin dose in patients on chronic OAC therapy, whereas the statin effect on bleeding risk in patients on VKA is controversial.[45,46]
Our study has several limitations. Some adverse events may not have been considered, for example, in patients who discontinued OAC therapy after an adverse event, patients who refused to be enrolled into the study, and patients who died before hospital admission due to any cause. Moreover, we enrolled patients monitored in anticoagulation clinics and these may not be representative of the general VKA patient population. Also, we gained information on selected interacting drugs, which are frequently used in associated comorbidities. However, we reduced potential bias by selecting cases from 2 institutions (Brescia Anticoagulation Centre and Cremona Haemostasis and Thrombosis Centre). Furthermore, controls were matched to cases for all variables (age, sex, clinical indication for therapy, and duration of OAC therapy) except for the occurrence of an adverse event. A strength of the study is that we had a homogenous population of Caucasians, important for genetic analyses, largely because CYP2C9*2 and CYP2C9*3 polymorphisms are primarily found in Caucasian population. Our findings may have some important implications, regarding selection and monitoring of VKA patients. A recent pharmacoeconomic study showed that an indiscriminate use of DOACs in patients with atrial fibrillation is less cost-effective than the preemptive genotyping approach. It follows that preemptive genotyping may be recommended even in warfarin-naive patients, for whom DOACs are not suitable, and this approach could improve patient care, reducing the risk of adverse events and the planned days required to achieve the target daily drug dosage by INR monitoring. Genotype determination could be useful even in long-term VKA patients, to improve the quality of OAC therapy, by increasing the frequency of INR monitoring or identifying interacting drugs with greater accuracy. These considerations are of particular relevance in the elderly and in patients with decreased or impaired renal function, as VKAs remain the mainstay of anticoagulation therapy in these patients. Finally, implementation of pharmacogenetics could improve pharmacovigilance practice, as highlighted in guidelines recently released by the European Medicines Agency. VKAs are one of the most common drugs resulting in emergency department admissions[49,50] and VKORC1 and CYP2C9 genetic variants may represent useful predictors of adverse drug reactions. In conclusion, our results suggest that the information on CYP2C9 and VKORC1 variants may potentially enable clinicians to determine individualized anticoagulant treatment for each patient; improve use, management and quality of VKA anticoagulation control; and monitor drug surveillance as part of pharmacovigilance programs.
The authors would like to acknowledge all patients and laboratory technicians of the participating centers.
. Ageno W, Gallus AS, Wittkowsky A, et al. Oral anticoagulant therapy: antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012;141(2 suppl):e44S–88S.
. Thaler J, Pabinger I, Ay C. Anticoagulant treatment of deep vein thrombosis and pulmonary embolism: the present state of the art. Front Cardiovasc Medicine 2015;2:30.
. Di Biase L. Use of direct oral anticoagulants in patients with atrial fibrillation and valvular heart lesions. JAHA 2016;5:
. Johnson JA, Cavallari LH. Warfarin pharmacogenetics. Trends Cardiovas Med 2015;25:33–41.
. Pirmohamed M, Kamali F, Daly AK, et al. Oral anticoagulation
: a critique of recent advances and controversies. Trends Pharmacol Sci 2015;36:153–63.
. Wadelius M, Chen LY, Lindh JD, et al. The largest prospective warfarin-treated cohort supports genetic forecasting. Blood 2009;113:784–92.
. Li T, Chang CY, Jin DY, et al. Identification of the gene for vitamin K epoxide reductase. Nature 2004;427:541–4.
. Rost S, Fregin A, Ivaskevicius V, et al. Mutations in VKORC1
cause warfarin resistance and multiple coagulation factor deficiency type 2. Nature 2004;427:537–41.
. Oldenburg J, Bevans CG, Muller CR, et al. Vitamin K epoxide reductase complex subunit 1 (VKORC1
): the key protein of the vitamin K cycle. Antioxid Redox Signal 2006;8:347–53.
. Tie JK, Stafford DW. Structural and functional insights into enzymes of the vitamin K cycle. JTH 2016;14:236–47.
. Rieder MJ, Reiner AP, Gage BF, et al. Effect of VKORC1
haplotypes on transcriptional regulation and warfarin dose. NEJM 2005;352:2285–93.
. Yuan HY, Chen JJ, Lee MT, et al. A novel functional VKORC1
promoter polymorphism is associated with inter-individual and inter-ethnic differences in warfarin sensitivity. Hum Mol Gen 2005;14:1745–51.
. Williams PA, Cosme J, Ward A, et al. Crystal structure of human cytochrome P450 2C9 with bound warfarin. Nature 2003;424:464–8.
. Van Booven D, Marsh S, McLeod H, et al. Cytochrome P450 2C9-CYP2C9
. Pharmacogenet Genom 2010;20:277–81.
. Lindh JD, Holm L, Andersson ML, et al. Influence of CYP2C9
genotype on warfarin dose requirements—a systematic review and meta-analysis. Euro J Clin Pharmacol 2009;65:365–75.
. Puehringer H, Loreth RM, Klose G, et al. VKORC1
-1639G>A and CYP2C9
*3 are the major genetic predictors of phenprocoumon dose requirement. Euro J Clin Pharmacol 2010;66:591–8.
. Klein TE, Altman RB, Eriksson N, et al. International Warfarin Pharmacogenetics ConsortiumEstimation of the warfarin dose with clinical and pharmacogenetic data. NEJM 2009;360:753–64.
. Kimmel SE, French B, Kasner SE, et al. A pharmacogenetic versus a clinical algorithm for warfarin dosing. NEJM 2013;369:2283–322.
. Pirmohamed M, Burnside G, Eriksson N, et al. A randomized trial of genotype-guided dosing of warfarin. NEJM 2013;369:2294–303.
. Li X, Yang J, Wang X, et al. Clinical benefits of pharmacogenetic algorithm-based warfarin dosing: meta-analysis of randomized controlled trials. Thromb Res 2015;135:621–9.
. Cavallari LH, Kittles RA, Perera MA. Genotype-guided dosing of vitamin K antagonists. NEJM 2014;370:1763.
. Daneshjou R, Klein TE, Altman RB. Genotype-guided dosing of vitamin K antagonists. NEJM 2014;370:1762–3.
. Kimmel SE, French B, Geller NL, et al. Genotype-guided dosing of vitamin K antagonists. NEJM 2014;370:1763–4.
. Koller EA, Roche JC, Rollins JA. Genotype-guided dosing of vitamin K antagonists. NEJM 2014;370:1761.
. Maitland-van der Zee AH, de Boer A, Manolopoulos VG, et al. Genotype-guided dosing of vitamin K antagonists. NEJM 2014;370:1765–6.
. Pereira NL, Rihal CS, Weinshilboum RM. Genotype-guided dosing of vitamin K antagonists. NEJM 2014;370:1762.
. Pirmohamed M, Wadelius M, Kamali F, et al. Genotype-guided dosing of vitamin K antagonists. NEJM 2014;370:1764–5.
. Schwarz UI, Kim RB, Tirona RG. Genotype-guided dosing of vitamin K antagonists. NEJM 2014;370:1761–2.
. Tang Q, Zou H, Guo C, et al. Outcomes of pharmacogenetics-guided dosing of warfarin: a systematic review and meta-analysis. Int J Cardiol 2014;175:587–91.
. Yang J, Chen Y, Li X, et al. Influence of CYP2C9
genotypes on the risk of hemorrhagic complications in warfarin-treated patients: a systematic review and meta-analysis. Int J Cardiol 2013;168:4234–43.
. Stergiopoulos K, Brown DL. Genotype-guided vs clinical dosing of warfarin and its analogues: meta-analysis of randomized clinical trials. JAMA Int Med 2014;174:1330–8.
. FCSA. Italian Federation of Anticoagulation Clinics, Guida alla Terapia con Anticoagulanti Orali. 2014.
. Shaw K, Amstutz U, Kim RB, et al. Clinical practice recommendations on genetic testing of CYP2C9
variants in warfarin therapy. Ther Drug Monit 2015;37:428–36.
. Jorgensen AL, FitzGerald RJ, Oyee J, et al. Influence of CYP2C9
on patient response to warfarin: a systematic review and meta-analysis. PloS One 2012;7:e44064.
. Roth JA, Boudreau D, Fujii MM, et al. Genetic risk factors for major bleeding in patients treated with warfarin in a community setting. CPT 2014;95:636–43.
. Mega JL, Walker JR, Ruff CT, et al. Genetics and the clinical response to warfarin and edoxaban: findings from the randomized, double-blind ENGAGE AF-TIMI 48 trial. Lancet 2015;385:2280–7.
. Tomek A, Matoska V, Kolarova T, et al. The bleeding risk during warfarin therapy is associated with the number of variant alleles of CYP2C9
genes. Cardiology 2013;125:182–91.
. Kawai VK, Cunningham A, Vear SI, et al. Genotype and risk of major bleeding during warfarin treatment. Pharmacogenomics 2014;15:1973–83.
. Zhang H, Yang L, Feng Q, et al. Association between VKORC1
gene polymorphisms and ischemic cerebrovascular disease in Chinese Han population. J Mol Neurosci 2014;53:166–70.
. Wang Y, Zhang W, Zhang Y, et al. VKORC1
haplotypes are associated with arterial vascular diseases (stroke, coronary heart disease, and aortic dissection). Circulation 2006;113:1615–21.
. Santos PC, Soares RA, Strunz CM, et al. Simultaneous use of amiodarone
influences warfarin maintenance dose but is not associated with adverse events
. JMCP 2014;20:376–81.
. Xu H, Ruff CT, Giugliano RP, et al. Concomitant use of single antiplatelet therapy with edoxaban or warfarin in patients with atrial fibrillation: analysis from the ENGAGE AF-TIMI48 trial. JAMA 2016;5:e002587doi: 10.1161/JAHA.115.002587.
. Sconce EA, Khan TI, Daly AK, et al. The impact of simvastatin on warfarin disposition and dose requirements. JTH 2006;4:1422–4.
. Douketis JD, Melo M, Bell CM, et al. Does statin therapy decrease the risk for bleeding in patients who are receiving warfarin? AMJ 2007;120:369.e369–14.
. Schelleman H, Bilker WB, Brensinger CM, et al. Fibrate/statin initiation in warfarin users and gastrointestinal bleeding risk. AMJ 2010;123:151–7.
. You JH. Universal versus genotype-guided use of direct oral anticoagulants in atrial fibrillation patients: a decision analysis. Pharmacogenomics 2015;16:1089–100.
. Magro L, Moretti U, Leone R. Epidemiology and characteristics of adverse drug reactions caused by drug–drug interactions. EODS 2012;11:83–94.
. Dechanont S, Maphanta S, Butthum B, et al. Hospital admissions/visits associated with drug-drug interactions: a systematic review and meta-analysis. PDS 2014;23:489–97.