Archived Comments for:
Efficiency and effectiveness of the use of an acenocoumarol pharmacogenetic dosing algorithm versus usual care in patients with venous thromboembolic disease initiating oral anticoagulation: study protocol for a randomized controlled trial
Efficiency and effectiveness of the use of an acenocoumarol pharmacogenetic dosing algorithm versus usual care in patients with venous thromboembolic disease initiating oral anticoagulation: study protocol for a randomized controlled trial
Rianne van Schie, Utrecht Institute for Pharmaceutical Sciences, Utrecht University
14 January 2015
R.M.F. van Schie1, T.I. Verhoef1, F.J.M van der Meer2, A. de Boer1, A.H. Maitland-van der Zee1, for the EU-PACT study group
1Department of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands
2Department of Thrombosis and Hemostasis, Leiden University Medical Center, Leiden, The Netherlands
To the Editor ¿ In the article of Carcas and coworkers1, the study protocol of a randomized controlled trial investigating the efficiency and effectiveness of the use of an acenocoumarol pharmacogenetic dosing algorithm is described. They conclude that there are no other clinical trials comparing pharmacogenetic acenocoumarol dosing algorithm versus routine clinical practice in venous thromboembolism. However, in the European Pharmacogenetics of Anticoagulant Therapy (EU-PACT) trial the added value of pretreatment genotyping is investigated in patients with atrial fibrillation as well as with venous thromboembolism, although we are not comparing it with standard clinical care2. EU-PACT¿s primary outcome measure is the percentage time patients spend within target INR range based on a dose algorithm including genotype information versus a dose algorithm without genotype information. This is how we can investigate the added value of genotyping. If Carcas and coworkers will find a difference between the two groups, they are not able to conclude whether the difference arises from the use of genotype or the use of a dose algorithm in general.
Carcas et al. use the CYP4F2 and APOE genotype to further optimize the dose algorithm. In the EU-PACT algorithm these genotypes are not included since they only explain a small part of the dose variability (in the situation of Carcas et al. an additional 3.9%) which seems not clinically relevant in estimating the coumarin dose. The question is whether the determination of the extra two genotypes is a cost-effective approach.
As a final remark, we would like to address the sample size calculation. Carcas et al. based their sample size on the outcome measure ¿percentage of patients with an INR between 2 and 3 on day 7 of the anticoagulation therapy¿. As we all know, acenocoumarol has a short half-life of less than 8 hours, which causes rapid INR fluctuations especially during the initiation period in which the personalized dose still needs to be determined. We therefore think that a longer-term perspective, such as percentage time spend within target INR range, is more appropriate.
References
1. Carcas AJ, Borobia AM, Velasco M, Abad-Santos F, Díaz MQ, Fernández-Capitán C, Ruiz-Giménez N, Madridano O, Sillero PL; And the PGX-ACE Spanish Investigators Group. Efficiency and effectiveness of the use of an acenocoumarol pharmacogenetic dosing algorithm versus usual care in patients with venous thromboembolic disease initiating oral anticoagulation: study protocol for a randomized controlled trial. Trials. 2012 Dec 13;13:239. doi: 10.1186/1745-6215-13-239.
2. van Schie RM, Wadelius MI, Kamali F, Daly AK, Manolopoulos VG, de Boer A, et al. Genotype-guided dosing of coumarin derivatives: the European pharmacogenetics of anticoagulant therapy (EU-PACT) trial design. Pharmacogenomics 2009; Oct; 10: 1687- 1695.
Competing interests
None declared
Response to van Schie et al.
ALBERTO M. BOROBIA, La Paz University Hospital. School of Medicine. Universidad Autónoma de Madrid. IdiPAZ
10 April 2015
Antonio J. Carcas*, Alberto M. Borobia*
*Clinical Pharmacology Department, La Paz University Hospital, Pharmacology Department, School of Medicine, Universidad Autónoma de Madrid, IdiPAZ
Dear Editor, we very much appreciate the comments made by van Schie et al. on our protocol published in Trials (1).
We agree that our protocol aims to answer the question on the effectiveness of an algorithm that includes pharmacogentic information, not only the phamacogenetic information, in comparison with usual care. The reason why we planned the trial this way was that in Spain acenocoumarol dose algorithms are not widely used and therefore the comparison with usual care was the logical option for an effectiveness trial. In addition, is our view that pharmacogenetic information must be managed in combination with other factors influencing the dose needed to obtain an INR in the therapeutic range and taking into account the patient’s characteristics.
In relation with the use of CYP4F2 and APOE genotypes, being true that their contribution to explain the variability in the acenocoumarol dose is not impressive, they could help to optimize the dose in some patients (namely those needing higher doses). We must consider that the objective of individualization is precisely to target those patients in the extremes of the dose requirements; for those needing the average dose a pharmacogenetic algorithm (or any other tool) would be of quite low utility. Inasmuch, we can genotype patients for these variants trough an array we have developed (Pharmarray® that include 192 SNPs from 57 genes (2). The unitary cost of this array is 23 euros including DNA extraction (not including instrument or technician cost)
Two other aspects that we see as related deserve explanation. Patients with atrial fibrillation and with thromboembolic disease (VTE) have quite different characteristics: age or concomitant medications and pathologies, just to mention the most obvious. The majority of the algorithms developed for acencocoumarol (and warfarin) have been developed in cohorts formed mainly of patients with atrial fibrillation. In fact in some of these cohorts this aspect is not even described (3). The reason why we choose primary variables exploring an early achievement of INR in the therapeutic range is that in the treatment of VTE reaching an effective anticoagulation as soon as possible is of paramount importance. As we mention in the discussion section of our paper, the risk of progression or recurrence (even death due to PTE) is greater in the first weeks after the diagnosis. Also the risk of bleeding is greater at the beginning of treatment. In any case we also included variables looking for long-term control as the number of INRs within the range during the three months of the study, proportion of time within the therapeutic range and number of unscheduled INR measurements during the study.
References
1. Carcas AJ, Borobia AM, Velasco M, Abad-Santos F, Díaz MQ, Fernández-Capitán C, Ruiz-Giménez N, Madridano O, Sillero PL; And the PGX-ACE Spanish Investigators Group. Efficiency and effectiveness of the use of an acenocoumarol pharmacogenetic dosing algorithm versus usual care in patients with venous thromboembolic disease initiating oral anticoagulation: study protocol for a randomized controlled trial. Trials. 2012 Dec 13;13:239.
3. van Schie RM1, Wessels JA, le Cessie S, de Boer A, Schalekamp T, van der Meer FJ, Verhoef TI, van Meegen E, Rosendaal FR, Maitland-van der Zee AH; EU-PACT Study Group. Loading and maintenance dose algorithms for phenprocoumon and acenocoumarol using patient characteristics and pharmacogenetic data. Eur Heart J. 2011 Aug;32(15):1909-17.
Trials
Efficiency and effectiveness of the use of an acenocoumarol pharmacogenetic dosing algorithm versus usual care in patients with venous thromboembolic disease initiating oral anticoagulation: study protocol for a randomized controlled trial
14 January 2015
R.M.F. van Schie1, T.I. Verhoef1, F.J.M van der Meer2, A. de Boer1, A.H. Maitland-van der Zee1, for the EU-PACT study group
1Department of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands
2Department of Thrombosis and Hemostasis, Leiden University Medical Center, Leiden, The Netherlands
To the Editor ¿ In the article of Carcas and coworkers1, the study protocol of a randomized controlled trial investigating the efficiency and effectiveness of the use of an acenocoumarol pharmacogenetic dosing algorithm is described. They conclude that there are no other clinical trials comparing pharmacogenetic acenocoumarol dosing algorithm versus routine clinical practice in venous thromboembolism. However, in the European Pharmacogenetics of Anticoagulant Therapy (EU-PACT) trial the added value of pretreatment genotyping is investigated in patients with atrial fibrillation as well as with venous thromboembolism, although we are not comparing it with standard clinical care2. EU-PACT¿s primary outcome measure is the percentage time patients spend within target INR range based on a dose algorithm including genotype information versus a dose algorithm without genotype information. This is how we can investigate the added value of genotyping. If Carcas and coworkers will find a difference between the two groups, they are not able to conclude whether the difference arises from the use of genotype or the use of a dose algorithm in general.
Carcas et al. use the CYP4F2 and APOE genotype to further optimize the dose algorithm. In the EU-PACT algorithm these genotypes are not included since they only explain a small part of the dose variability (in the situation of Carcas et al. an additional 3.9%) which seems not clinically relevant in estimating the coumarin dose. The question is whether the determination of the extra two genotypes is a cost-effective approach.
As a final remark, we would like to address the sample size calculation. Carcas et al. based their sample size on the outcome measure ¿percentage of patients with an INR between 2 and 3 on day 7 of the anticoagulation therapy¿. As we all know, acenocoumarol has a short half-life of less than 8 hours, which causes rapid INR fluctuations especially during the initiation period in which the personalized dose still needs to be determined. We therefore think that a longer-term perspective, such as percentage time spend within target INR range, is more appropriate.
References
1. Carcas AJ, Borobia AM, Velasco M, Abad-Santos F, Díaz MQ, Fernández-Capitán C, Ruiz-Giménez N, Madridano O, Sillero PL; And the PGX-ACE Spanish Investigators Group. Efficiency and effectiveness of the use of an acenocoumarol pharmacogenetic dosing algorithm versus usual care in patients with venous thromboembolic disease initiating oral anticoagulation: study protocol for a randomized controlled trial. Trials. 2012 Dec 13;13:239. doi: 10.1186/1745-6215-13-239.
2. van Schie RM, Wadelius MI, Kamali F, Daly AK, Manolopoulos VG, de Boer A, et al. Genotype-guided dosing of coumarin derivatives: the European pharmacogenetics of anticoagulant therapy (EU-PACT) trial design. Pharmacogenomics 2009; Oct; 10: 1687- 1695.
Competing interests
None declared
Response to van Schie et al.
10 April 2015
Antonio J. Carcas*, Alberto M. Borobia*
*Clinical Pharmacology Department, La Paz University Hospital, Pharmacology Department, School of Medicine, Universidad Autónoma de Madrid, IdiPAZ
Dear Editor, we very much appreciate the comments made by van Schie et al. on our protocol published in Trials (1).
We agree that our protocol aims to answer the question on the effectiveness of an algorithm that includes pharmacogentic information, not only the phamacogenetic information, in comparison with usual care. The reason why we planned the trial this way was that in Spain acenocoumarol dose algorithms are not widely used and therefore the comparison with usual care was the logical option for an effectiveness trial. In addition, is our view that pharmacogenetic information must be managed in combination with other factors influencing the dose needed to obtain an INR in the therapeutic range and taking into account the patient’s characteristics.
In relation with the use of CYP4F2 and APOE genotypes, being true that their contribution to explain the variability in the acenocoumarol dose is not impressive, they could help to optimize the dose in some patients (namely those needing higher doses). We must consider that the objective of individualization is precisely to target those patients in the extremes of the dose requirements; for those needing the average dose a pharmacogenetic algorithm (or any other tool) would be of quite low utility. Inasmuch, we can genotype patients for these variants trough an array we have developed (Pharmarray® that include 192 SNPs from 57 genes (2). The unitary cost of this array is 23 euros including DNA extraction (not including instrument or technician cost)
Two other aspects that we see as related deserve explanation. Patients with atrial fibrillation and with thromboembolic disease (VTE) have quite different characteristics: age or concomitant medications and pathologies, just to mention the most obvious. The majority of the algorithms developed for acencocoumarol (and warfarin) have been developed in cohorts formed mainly of patients with atrial fibrillation. In fact in some of these cohorts this aspect is not even described (3). The reason why we choose primary variables exploring an early achievement of INR in the therapeutic range is that in the treatment of VTE reaching an effective anticoagulation as soon as possible is of paramount importance. As we mention in the discussion section of our paper, the risk of progression or recurrence (even death due to PTE) is greater in the first weeks after the diagnosis. Also the risk of bleeding is greater at the beginning of treatment. In any case we also included variables looking for long-term control as the number of INRs within the range during the three months of the study, proportion of time within the therapeutic range and number of unscheduled INR measurements during the study.
References
1. Carcas AJ, Borobia AM, Velasco M, Abad-Santos F, Díaz MQ, Fernández-Capitán C, Ruiz-Giménez N, Madridano O, Sillero PL; And the PGX-ACE Spanish Investigators Group. Efficiency and effectiveness of the use of an acenocoumarol pharmacogenetic dosing algorithm versus usual care in patients with venous thromboembolic disease initiating oral anticoagulation: study protocol for a randomized controlled trial. Trials. 2012 Dec 13;13:239.
2. Borobia A.M. , Lubomirov R. , Tong H.Y. , Arias P. , Tenorio J., Frías J., Lapunzina P., Carcas A.J. implementing pharmacogenetics: Pharmarray, Acodose and pharmacogenetic consultation. Basic & Clinical Pharmacology & Toxicology, 113 (Suppl. 2), 16–37.
3. van Schie RM1, Wessels JA, le Cessie S, de Boer A, Schalekamp T, van der Meer FJ, Verhoef TI, van Meegen E, Rosendaal FR, Maitland-van der Zee AH; EU-PACT Study Group. Loading and maintenance dose algorithms for phenprocoumon and acenocoumarol using patient characteristics and pharmacogenetic data. Eur Heart J. 2011 Aug;32(15):1909-17.
Competing interests
None declared