Does Low Dose Oral Vitamin K Improve International Normalized Ratio (INR) Stability?

What is/are the principal research objective(s) and question(s) to be addressed? Research Objectives: Our broad objective is to improve the quality of anticoagulant care for patients receiving warfarin. Our specific objective is to determine if we can improve the TTR with LDVK, a non-toxic, inexpensive and easily administered medication.

Hypothesis: We hypothesize that patients receiving vitamin K will have a higher TTR than those receiving placebo. In a mechanistic study we proposed that VKORC1 and CYP2C9 genotype will modulate the impact of vitamin K on INR control.

Research questions:

Efficacy (primary): When compared with placebo does the addition of 150 micrograms of daily vitamin K to "usual warfarin therapy" improve anticoagulant control as measured by "time in the therapeutic range"? Safety (Secondary): Does LDVK increase the frequency of adverse clinical events, including thromboembolism? Does LDVK reduce the risk of major and/or all bleeding? Mechanistic (Hypothesis generating): Do VKORC1 and CYP2C9 genotypes modulate the effect vitamin K on INR stability?

Study design:

The proposed pilot study is a multi-centre, placebo controlled, randomized trial with an additional pilot mechanistic study (Figure 1 provides a broad outline of the study).

Study interventions:

Patients will receive a daily dose of 150 micrograms of vitamin K or a matching placebo medication for a total of 7 months - a one month "run in period" and a 6 month period of follow-up. They will continue to receive warfarin under the supervision of their anticoagulant clinic - other than one study mandated INR within 7 days of enrolment no changes in patient care will be made as a result of participation in this study.

All patients will be enrolled in a mechanistic study. Patients will have blood drawn at the time they provide consent to have their VKORC1 and CYP2C9 genotypes determined.

Allocation of patients into study groups:

Randomization will be achieved using a computer generated random number table. Patients will be stratified based on clinical centre. Practically, a random number table will be produced at the coordinating centre for each clinical centre. The study coordinating centre will package a six week supply of drug and placebo into sequentially numbered, identical drug containers based on the randomization sequence for each site. The drug packages will then be shipped to each centre. After consent is obtained, and the coordinating centre is notified, patients will be allocated to treatment by dispensing sequential drug containers in numeric sequence. Additional supplies (for a total of seven months of therapy for each patient) will be dispensed by the anticoagulant clinic in blocks of 6 weeks. This method of allocation ensures research subjects and study staff are masked to allocation as drug and placebo are identical. Masked allocation will be further protected by not providing an unblinding schedule for centres as we can foresee no circumstance in which immediate unblinding would be required.

Frequency and duration of trial follow-up:

Patients will have their INR checked at a minimum of 1 week after initiation of study drug and then will be followed as per their anticoagulant clinic routine. We will not mandate any particular frequency of INR testing as changing clinic routines will induce unwanted changes in INR control is a co-intervention that could influence TTR and that is not attributable to the study intervention. Clinical events will be reviewed, as per current clinic routine, at each INR check. If clinical events are reported source documentation will be sought for the purposes of adjudication. Our research group has extensive prior experience with collection and interpretation of INR values, and of the ascertainment, collection and reporting of clinical events in anticoagulated patients.

Primary and secondary outcome measures:

The primary outcome is a simple comparison of mean TTRs in the LDVK and placebo patients. As noted, calculation of the TTR will begin at the first INR obtained a minimum of 4 weeks after randomization and will end 7 months after enrolment. The TTR will be calculated using a previously validated computer algorithm, based on the method of Rosendaal et al (7). Secondary outcome measures will include (a) all significant INR excursion (INR < 1.5 or > 4.5), (b) number of INR determinations and (c) bleeding and thrombotic events of sufficient severity to require a medical intervention - where possible source documents from hospital admissions will be obtained to allow characterization of these events. Major bleeding will be defined as bleeding consistent with the 8th ACCP guidelines (36). All bleeding and thrombotic events will be independently adjudicated by a panel of experts based on submitted source documentation.

Plan for ascertaining outcome measures at follow-up:

INR values are obtained routinely as part of warfarin care. All INR determinations will be performed in clinical laboratories that participate in provincially mandated external quality assessment exercises to ensure quality. INR values and current warfarin dose will be reported to the coordinating centre using the online data entry and reporting system. Clinical events will be measured as per the "Primary and secondary outcome measures" section.

Will health services research issues be addressed within this trial? As a pilot study we do not feel that a valid health services research question can be posed in the context of this study. We do propose a simple cost effectiveness analysis which will examine the estimated total costs of participation in the study - for this analysis we will compare the number and type of medical interventions (and their costs) in the two arms of the study. Costs will be determined from validated Ontario costing data for INR tests, clinical visits and monitoring costs as well as any estimated costs attributable to clinic events, should these occur. Funding for this analysis will be sought from different sources, however we will ensure that the data required to complete this analysis will be collected within the context of this study.

Trial sample size and justification for the assumptions underlying the power calculations:

As a pilot study of a novel intervention we do not have estimates of the improvement of the TTR we expect to see with LDVK. As such the proposed sample size of 400 patients (200 per arm) is one of convenience however the study is powered to detect probable changes in TTR. Thus, we have an 85% likelihood of detecting a 15% increase in TTR assuming a TTR of 55% in the placebo arm (two sided alpha = 0.05, comparison of proportions, Fleiss correction) (37). A table of conditional powers is presented in Table 2. In subsequent studies (if justified, based on the results of this study) we will power based on the rates of clinical events. One prior paper has examined increases in TTR with vitamin K administration; the increase in TTR in this study was 28%, thus we are reassured by our ability to reliably detect a 15% difference in TTR (17).