Long-term anticoagulant effects of CYP2C9 and VKORC1 genotypes in phenprocoumon users


Anke-Hilse Maitland-van der Zee, Division of Pharmacoepidemiology and Clinical Pharmacology, Faculty of Science, Utrecht University, PO Box 80 082, 3508 TB Utrecht, the Netherlands.
Tel.: +31 622 736 715; fax: +31 30 253 9166.
E-mail: a.h.maitland@uu.nl


Verhoef TI, Redekop WK, Hegazy H, de Boer A, Maitland-van der Zee AH, on behalf of the EU-PACT group. Long-term anticoagulant effects of CYP2C9 and VKORC1 genotypes in phenprocoumon users. J Thromb Haemost 2012; 10: 2610–2.

Anticoagulant treatment with phenprocoumon is challenging because of the narrow therapeutic range and the wide inter- and intra-patient variability in dose response. Frequent monitoring of the international normalized ratio (INR) is therefore required. Polymorphisms in two genes, CYP2C9 and VKORC1, explain approximately one-third of the variation in dose requirements [1–3]. CYP2C9 encodes the main metabolizing enzyme of coumarins, the cytochrome P450 2C9 enzyme (CYP2C9), while VKORC1 encodes the pharmacodynamic target enzyme for coumarins, vitamin K epoxide reductase multiprotein complex 1 (VKORC1).

Earlier this year, we found that in the first month of acenocoumarol therapy, the risk of underdosing is highest in patients with a VKORC1 wild type [4]. This increased risk of a subtherapeutic INR was also seen in months 2 and 3, but not after the third month of coumarin treatment. In addition, the risk of overdosing was highest in patients with a VKORC1 TT genotype in the first 6 months. The effect of CYP2C9 genotype on under- or overdosing of acenocoumarol was smaller than the effect of VKORC1 and this effect was only found in the first month of therapy and not after the initiation period [4]. This has not been investigated for phenprocoumon yet. The aim of this study was therefore to examine the association of CYP2C9 and VKORC1 polymorphisms with the risk of over- and under-anticoagulation after the initiation period of phenprocoumon.

To investigate this, we looked at data from two different studies, the pre-EU-PACT study [5] and the study by Schalekamp et al. [6]. The study protocols of both studies were approved by a Medical Ethics Committee (Leiden University Medical Center, Leiden, for the pre-EU-PACT study, and Utrecht Medical Centre, Utrecht, for the study by Schalekamp) and patients provided informed consent before study inclusion. All procedures were conducted in accordance with the Helsinki Declaration. More details about the design and data collection in both studies can be found elsewhere [4].

We examined the occurrence of at least one INR < 2, > 3.5 or > 6.0 in several time periods up to 1.5 years after treatment initiation and tested for differences among the genotypes with chi-square analysis. The time periods we used were: 0–1 month (days 1–30), 1–3 months (days 31–90), 3–6 months (days 91–180), 6–9 months (days 181–270), 9–12 months (days 271–360), 12–15 months (days 361–450) and 15–18 months (days 451–540) after treatment initiation. We also looked at the time within, below and above the therapeutic range, because this method is more robust when the frequency of INR measurements differs between patients. All analyses were performed using SPSS 18.0 (SPSS Inc., Chicago, IL, USA).

In total, 794 phenprocoumon users from the two studies [5,6] were eligible for analyses in this study. Patient characteristics and genotypes of all 794 patients are shown in Table S1. Data on height and weight were only available in the Pre-EU-PACT study (n = 486). The most frequent indication for phenprocoumon treatment was atrial fibrillation. The average number of INR measurements per time period ranged from 3.4 to 5.5. Only data of patients using phenprocoumon during the entire time period were included in the analysis of that period. The maximum follow-up in the Schalekamp study was 6 months (n = 308).

Significant differences in out-of-range INR values between the genotypes were only found during the first month of phenprocoumon therapy. In the first month, 89% of the patients with a VKORC1 wild type had at least one subtherapeutic INR. This frequency was significantly lower among patients with CT (76%, P < 0.001) and TT (50%, P < 0.001). Supratherapeutic INR values occurred in 33% of the VKORC1 wild-type patients, vs. 48% (P < 0.001) and 66% (P < 0.001) in patients with a CT or TT genotype, respectively. Of the wild-type patients, 3% had at least one INR > 6. This percentage was increased in patients with a TT genotype (17%, P < 0.001), but there was no statistically significant difference for patients with a CT genotype (6%, P = 0.12).

Occurrences of subtherapeutic INR values or INR values > 6 were not significantly different among the CYP2C9 genotypes. However, INR values > 3.5 occurred more often in carriers of a CYP2C9*3 allele (62%, P < 0.001) or a CYP2C9*2 allele (52%, P = 0.01) than in wild-type patients (40%). For both VKORC1 and CYP2C9 genotypes, no significant differences in out-of-range INRs were found after the first month. The risk of out-of-range INRs for the different periods and genotypes is shown in the Supporting Information (Figs S1–S4).

Similar results were obtained in the analyses of time within, below and above therapeutic INR range (Fig. 1). In the first month, time below therapeutic INR range was longest in VKORC1 and CYP2C9 wild-type patients (up to 33%) and time above therapeutic INR range was longest in VKORC1–TT and CYP2C9*3 carriers (up to 37%). The risk of having at least one INR < 2 did not vary significantly among the CYP2C9 genotypes, but the time spent below therapeutic INR range was significantly shorter in *2 carriers (19%) and *3 carriers (14%) than in wild-type patients (26%, P < 0.001). No significant differences were found after month 1 of the treatment.

Figure 1.

 Percentage time in different INR ranges during the first month of phenprocoumon use. (A) VKORC1 genotypes, (B) CYP2C9 genotypes.

Our study demonstrated that in the first month of phenprocoumon therapy, the risk of underdosing is highest in patients with VKORC1 and CYP2C9 wild types. In addition, the risk of overdosing was highest in patients with a VKORC1 TT genotype or carriers of a CYP2C9 variant allele. These results correspond with the results we have seen for acenocoumarol users, as described in a previous article [4]. However, the results beyond the first month of treatment are not similar. Specifically, while there were no differences in the risk of out-of-range INRs between the different genotypes after the first month of phenprocoumon therapy, there were differences in risk between the VKORC1 genotypes up to the sixth month of acenocoumarol treatment.

A limitation of this study is the fact that the Pre-EU-PACT study contained retrospective data [5]. The data of Schalekamp et al. [6], however, were collected prospectively. Data for a specific time period were only used in the analysis if the patient used phenprocoumon for this entire period. Because very unstable patients are expected to stop the therapy early, this patient group might be under-represented in our study.

Information about the patient’s genotype can be used to predict the right dose of phenprocoumon [5]. Carriers of a VKORC1 or CYP2C9 variant allele require a lower dose and have an increased risk of supratherapeutic INR values. If these patients are genotyped before treatment initiation, they could be treated with a lower dose, thereby decreasing the risk of over-anticoagulation. In both this study and our previous study on acenocoumarol, we also found an increased risk of a subtherapeutic INR in VKORC1 and CYP2C9 wild-type patients during the first month. Information about the patient’s genotype could therefore also be used to identify patients who need a higher dose to decrease the risk of complications from underdosing. In this way, genetic information could be used to improve the safety and efficacy of anticoagulation treatment in both wild-type patients and variant carriers. The relevance of pharmacogenetic information for phenprocoumon users, however, seems to be limited to the first month of treatment.

Phenprocoumon has a longer elimination half-life than acenocoumarol (110–130 h vs. 6–8 h) [3,7]. Treatment with phenprocoumon is therefore somewhat more stable and patients on phenprocoumon spend more time within the therapeutic INR range than patients on acenocoumarol [8]. This might be a reason why only acenocoumarol users, and not phenprocoumon users, show differences between the genotypes in the risk of out-of-range INRs after the first month of treatment.

The results of this study suggest that pharmacogenetic information might help to prevent subtherapeutic or supratherapeutic INRs in the first month of phenprocoumon therapy and thereby reduce the risk of adverse events. The value of this information after the first month of phenprocoumon treatment appears to be limited. Currently, clinical trials are underway to investigate the effectiveness and cost-effectiveness of a genotype-guided dosing regimen vs. a standard dosing regimen [9,10].


We would like to thank the Anticoagulation Clinic Leiden, F.J.M. van der Meer, T. Schalekamp, S. le Cessie and R.M.F. van Schie for their support in data collection and analysis.

Disclosure of Conflict of Interests

This project is funded by the European Community’s Seventh Framework Programme under grant agreement HEALTH-F2-2009-223062. The Department of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute of Pharmaceutical Sciences, employing authors Talitha I. Verhoef, Hoda Hegazy, Anthonius de Boer and Anke-Hilse Maitland-van der Zee, has received unrestricted research funding from the Netherlands Organisation for Health Research and Development (ZonMW), the Dutch Health Care Insurance Board (CVZ), the Royal Dutch Pharmacists Association (KNMP), the private-public funded Top Institute Pharma (http://www.tipharma.nl, includes co-funding from universities, government, and industry), the EU Innovative Medicines Initiative (IMI), EU 7th Framework Program (FP7), the Dutch Medicines Evaluation Board, and the Dutch Ministry of Health and Industry (including GlaxoSmithKline, Pfizer, and others). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. No writing assistance was utilized in the production of this manuscript.


Rita Barallon, LGC Limited, Middlesex, UK; Anthonius de Boer, Utrecht University, Utrecht, the Netherlands; Ann Daly, Newcastle University, Newcastle upon Tyne, UK; Elisabeth Haschke-Becher, Elisabethinen Hospital Linz, Linz, Austria; Farhad Kamali, Newcastle University, Newcastle upon Tyne, UK; Anke-Hilse Maitland, Utrecht University, Utrecht, the Netherlands; Ken Redekop, Erasmus University, Rotterdam, the Netherlands; Julia Stingl, Humboldt University of Berlin, Berlin, Germany; Vangelis G. Manolopoulos, Democritus University of Thrace, Alexandroupolis, Greece; Munir Pirmohamed, University of Liverpool, Liverpool, UK; Frits R. Rosendaal, Leiden University Medical Center, Leiden, the Netherlands; Mia Wadelius, Uppsala University, Uppsala, Sweden.