Study to Develop a Reliable Nomogram That Incorporates Clinical and Genetic Information

This is Phase 1 of a 3-phase plan in which we will ultimately test whether rapid turnaround genetic testing can improve the safety and efficacy of warfarin anticoagulation in warfarin naïve patients who are being newly induced and maintained on warfarin. Each of the phases will address a specific and progressively more ambitious question.

Phase 1 will be a 500-patient cohort study to determine whether we can develop a reliable nomogram that incorporates clinical and genetic information to maintain patients within the target therapeutic range more often than 70% of the time, the conventional historical benchmark for excellence in warfarin dosing.

BACKGROUND:

Warfarin was patented in 1948 and was introduced commercially in 1954. In 2004, more than 24 million prescriptions were written for warfarin in the United States alone. Warfarin constitutes the 20th most frequently prescribed drug in the United States. This is remarkable, because few drugs have had a life cycle as long as warfarin, and even fewer can claim an increase in use more than 50 years following introduction.

Warfarin is a Vitamin K antagonist. Gamma-carboxylation of vitamin K is crucial for coagulation factors II, VII, IX, and X to function properly, as well as antithrombotic endogenous Protein C and Protein S. Gamma-carboxylation of vitamin K allows these clotting proteins to bind calcium at phospholipid surfaces upon which coagulation occurs. Warfarin and other vitamin K antagonists interfere with vitamin K and cause the liver to synthesize nonfunctional coagulation factors.

Warfarin is almost completely absorbed by the gastrointestinal mucosa. It is metabolized in the liver by the Cytochrome P450 system. The kidney eliminates largely inactive metabolites.

Warfarin anticoagulation is prescribed to prevent stroke and venous thromboembolism. Optimal dosing requires achieving and maintaining a target range International Normalized Ratio (INR). The INR itself is a prothrombin time that is standardized according to the type of thromboplastin reagent used by the coagulation laboratory. Each thromboplastin reagent has a designated International Sensitivity Index. The formula used to calculate the INR is shown below:

The target INR for most indications, including atrial fibrillation, DVT, and pulmonary embolism, is usually between 2.0 and 3.0.

Excessive dosing is characterized by an elevated INR and precipitates hemorrhage, including stroke due to intracranial bleeding. To minimize the risk of intracranial hemorrhage, it is important to maintain a maximum INR level of 3.0. When the INR exceeds 3.0, the risk of intracranial bleeding increases exponentially.

Inadequate dosing, associated with a subtherapeutic INR, predisposes to stroke due to thromboembolism and to DVT or pulmonary embolism. For example, to prevent a thrombotic stroke in patients with atrial fibrillation, it is important to maintain an INR of at least 2.0.

The most vulnerable period for thrombosis and hemorrhage due to warfarin is during the initiation phase of anticoagulation, when optimal dosing is least certain. Adverse event rates are highest during this vulnerable period, which is generally considered to persist for at least 3 months. Patients who have not been previously exposed to warfarin are most vulnerable because there is no individual history to guide the clinician on an optimal initiation dose for those particular patients.

Dosing nomograms work poorly. Trial and error predominates as the method of dosing warfarin. Warfarin is virtually the only contemporary drug prescribed using trial and error methodology.

Various warfarin initiation regimens have been attempted, but none are in common use. In 1984, Fennerty proposed an every 16-hour dosing initiation nomogram for patients with venous thromboembolism whose average age was 52 years. This nomogram never gained wide support because of the awkward dosing interval and because it had been used in a patient population much younger than the average population of patients who take warfarin. In 1998, the "modified Fennerty" nomogram was introduced and dosed patients with warfarin every 24 hours. The nomogram enrolled an older population, average age of 78 years. However, 35% of these patients had excessively high INRs that exceeded 4.0 within the first 4 days of warfarin initiation.

Many clinical factors predispose to an excessively high INR: advanced age, abnormal liver function, decreased vitamin K intake because of poor nutrition or poor appetite, diarrhea, antibiotics, certain other concomitant medications, alcohol in binges, and perhaps changes in warfarin preparation (substituting one generic preparation for another or interchanging generic with brand-name Coumadin®). The most common reason for abnormally low INRs is intake of high amounts of vitamin K from green leafy vegetables, certain drug-warfarin interactions, or failure to take warfarin appropriately.

ADVANCES IN GENETICS:

Genotyping patients at the onset of warfarin anticoagulation may allow more precise dosing. This may translate into fewer major bleeding and clotting events as well as fewer out-of-target-range INRs, which serve as a surrogate for bleeding and clotting complications. During the first several weeks following prescription of warfarin, INR laboratory tests are obtained frequently, often twice weekly. Better predictive assessment of the optimal dose will decrease laboratory costs and improve convenience for patients.

Cytochrome P450 2C9 genotyping can identify mutations associated with impaired warfarin metabolism. The CYP2C9 genotype accounts for about 10% of warfarin dose variance. Vitamin K receptor polymorphism testing can identify whether patients require low, intermediate, or high doses of warfarin. Five common vitamin K receptor gene haplotypes account for about an additional 25% of warfarin dose variance.

Until now, the major drawback in applying screening for CYP2C9 polymorphisms to warfarin dosing and VKORC1 genotyping has been slow turnaround time. However, the HPCGG will be able in the Nomogram Development Trial to offer turnaround within 24-48 hours. This rapid turnaround time will allow initial immediate anticoagulation with once or twice daily injectable agents for several days. Therefore, prior to the initiation of warfarin, there will be sufficient time for the genetic information to be received and implemented by the clinical team.

The human cytochrome P450 (CYP) superfamily comprises 57 genes. These genes code for a myriad of enzymatic reactions. The cytochrome P450 CYP2C9 is responsible for metabolism of the S enantiomer of warfarin. Two allelic variants, CYP2C9*2 and CYP2C9*3, differ from the wild type CYP2C9*1 by a single amino acid substitution. The allelic variants are associated with impaired hydroxylation of S-warfarin.

A retrospective cohort study of 200 patients receiving warfarin dosed by anticoagulation clinics suggested that the CYP2C9*2 and CYP2C9*3 polymorphisms are associated with an increased risk of excessive anticoagulation and of bleeding events.

At the Brigham and Women's Hospital Anticoagulation Service, we identified 73 patients for CYP2C9 genotyping, which we assessed with PCR amplification and restriction enzyme digestion analysis of DNA isolated from circulating leukocytes. The CYP2C9 polymorphisms independently predicted low warfarin requirements after adjusting for body mass index, age, acetaminophen use, and race. At least one polymorphism was present in every patient requiring 1.5 mg or less of daily warfarin. We did find higher rates of excessive anticoagulation but did not observe higher bleeding rates in patients with these polymorphisms.

To date, only one prospective study has been published on warfarin dosing based upon cytochrome P450 2C9 genotyping. It was a negative study that failed to prove the hypothesis that genotyping would lead to better control of warfarin dosing. 48 orthopedic surgery patients were studied with genotyping and screening for nongenetic factors that might affect warfarin dosing and bleeding risk. The warfarin dosing algorithms based upon the genetic screening led to achievement of a stable and therapeutic INR in both the wild type group and the CYP2C9 variant groups with a similar time course. Nevertheless, despite pharmacogenetics-based dosing, patients with variants were at higher risk for excessive anticoagulation, defined as an INR greater than 4.0.

This failed study emphasizes the pivotal importance of a comprehensive nomogram development phase prior to initiating a randomized controlled trial. This trial also demonstrates the need to consider other factors that affect the INR and warfarin dose: advanced age, underlying comorbidities and thrombotic disposition, abnormal liver function, inadequate or excessive vitamin K intake, altered vitamin K metabolism (e.g., diarrhea or antibiotics), alcohol and drug-food interactions, and changes in warfarin preparation (switching among different generics or switching to or from brand name Coumadin®). The most important lesson from this cohort study is that to decrease clinical adverse events due to warfarin, it is first necessary to develop a warfarin nomogram that is superior than the current practice of educated guesses and trial and error.

Rieder and colleagues have identified 10 non-coding single- nucleotide polymorphisms (SNPs) in the vitamin K epoxide reductase complex 1 (VKORC1) gene that fall into five major haplotypes. Two of these haplotypes, A and B, can be used to determine whether patients require low (AA, 2.7 mg/day), intermediate (AB, 4.9 mg/day), or high (BB, 6.2 mg/day) doses of drug. The haplotypes differ according to racial distribution. These haplotypes are responsible for 25% of the variance in dose. Therefore, a combination of cytochrome P450 and VKORC1 genotyping will facilitate optimal warfarin dosing.

OBJECTIVE:

Primary Objective:

We will develop a nomogram for warfarin dosing that uses rapid turnaround genetic testing and monthly nomogram modification (if necessary) to achieve effective and safe warfarin induction and maintenance. More than 70% of the time, we will maintain warfarin naïve patients within the target therapeutic range. The percent of time in the therapeutic range will be analyzed beginning 2 weeks after initiation of warfarin. Analyses will be stratified by the indication for anticoagulation.

We will obtain Informed Consent for exploratory genetic testing of the sampled DNA, in addition to cytochrome P450 2C9 and VKORC1 alleles. This flexible approach will permit us to add additional promising tests that emerge and so that we can modify the nomogram to make it even more effective in achieving the target range for the INR.

PLANS:

Genetic Testing:

1. Detect known alleles of cytochrome P450 and Vitamin K epoxide reductase complex 1 (VKORC1).
2. Collect 5 ml of blood into EDTA (lavender top) tubes.
3. DNA will be processed using the ROCHE MagnaPure automated prep system.

Patient Population:

Warfarin naïve patients undergoing initiation of warfarin anticoagulation at participating Partners anticoagulation clinics, including Brigham and Women's Hospital, Massachusetts General Hospital, North Shore Medical Center, Faulkner Hospital, Spaulding Rehabilitation Hospital, and Newton Wellesley Hospital.

Overview and Timeline:

We will enroll patients over 9 months, follow each patient for 3 months with twice weekly coagulation testing of the prothrombin time standardized to the International Normalized Ratio, and adjust monthly the nomogram (if necessary) to improve the fit with emerging data from the cohort.