Experiments suggest that during treatment with vitamin K antagonists (VKA) the activity of coagulation factors (F) II and X better reflect anticoagulation than does FVII. Based on this a new prothrombin time based monitoring test (Fiix-PT) has been invented which is only sensitive to FII and FX. The Fiix-PT can be converted to INR ("Fiix-INR").

The investigators hypothesize that the Fiix-PT may reflect anticoagulation and the antithrombotic effect of VKA as accurately or better than the current PT based tests do (INR based on PT or P&P). The protocol describes a prospective randomized double-blind trial that will be conducted at the Landspitali Anticoagulation Management Center (AMC).

The objective of the protocol is to evaluate the efficacy and safety of Fiix-INR as a monitoring test compared to the current PT based assays (INR) used to monitor patients treated with VKA to prevent thromboembolism. The investigators will randomize 1200 clients of the AMC into two identically sized monitoring groups, Fiix-INR (test group) and INR (control group).

The clinical endpoints to be studied include efficacy (arterial and venous thromboembolic event rate) and safety (bleeding events). Additionally, surrogate convenience endpoints will be studied such as test frequency and time within target range.

Clinical outcome during warfarin anticoagulation to prevent thromboembolism monitored with "Fiix-INR" compared to standard monitoring with INR.

A prospective randomized double-blinded trial. Study protocol Abbreviated Version Jan 25 2012 Corresponding author: Pall T. Onundarson, M.D., Department of Laboratory Hematology and Hemostasis Center, K-building, Landspitali Hringbraut, 101 Reykjavik, Iceland, Phone: +354 543 1000/5010, Fax +354 543 5539, email: pallt@landspitali.is.

1. STUDY SYNOPSIS Experiments suggest that during treatment with vitamin K antagonists (VKA) the activity of coagulation factors (F) II and X better reflect anticoagulation than does FVII. Based on this a new monitoring prothrombin time base test (Fiix-PT) has been invented which is only sensitive to FII and FX. The Fiix-PT can be converted to INR ("Fiix-INR").We hypothesize that the Fiix-PT may reflect anticoagulation and the antithrombotic effect of VKA as accurately or better than the current PT based tests do (INR based on PT or P&P). The protocol describes a prospective randomized double-blind trial that will be conducted at the Landspitali Anticoagulation Management Center (AMC). The objective of the protocol is to evaluate the efficacy and safety of Fiix-INR as a monitoring test compared to the current PT based assays (INR) used to monitor patients treated with VKA in order to prevent thromboembolism. We will randomize 1200 clients of the AMC into two identically sized monitoring groups, Fiix-INR (test group) and INR (control group). Each individual will be assessed monthly for up to 24 months (an expected average of 18 months) until a total of 600 patients years of observation has been reached in each group. The clinical endpoints to be studied include efficacy (arterial and venous thromboembolic event rate) and safety (bleeding events). Also, surrogate convenience endpoints will be studied such as test frequency and time within target range.

2. INTRODUCTION AND RATIONALE 2.1 Background Vitamin K antagonists (VKA, coumarins) are orally active anticoagulants that prevent normal gamma carboxylation of the vitamin K dependent coagulation factors (FII, FVII, FIX, FX). The consequence is a reduction in the clotability of the blood, ie an anticoagulant effect that can be used therapeutically for the purpose of preventing thromboembolism. In order to assure efficacy and safety, the dose of VKA needs to be adjusted to maintain the normalized prothrombin time ratio (international normalized ratio, INR) within a safe therapeutic range (most often with INR target 2.5, range 2.0-3.0). The clotting times obtained with PT based tests have in practice for over 50 years been presumed to directly reflect the antithrombotic effect of VKA in patients (Loeliger 1984). The PT (or P&P-test) are used all over the world for the purpose of monitoring anticoagulation and to make dose-adjustments in patients taking VKA. Their results have been standardized by calculating the International Normalized Ratio (INR) which takes into account the sensitivity (strength) of the thromboplastin used by different reagents.

   However, inherent problems exist with the PT. Due to the VKD factors´ significantly different half-lifes (above), the PT clotting time obtained early after initiation or dose change may mainly reflect the concentration of factor VII since FVII has the notably shortest half-life (about 4-8 hours). If the concentration of factor VII has little influence on the antithrombotic effect of VKA (see below), it is possible that unnecessary and too frequent dose adjustments are made based on the use of the PT. This, could potentially be harmful to the patient.

   Prior studies by others have indicated that the thrombin generation(Xi, Beguin et al. 1989; Brummel, Paradis et al. 2001) and antithrombotic influence of factor VII in rabbits (Zivelin, Rao et al. 1993) may be less important than that of prothrombin or factor X. Based on those observations, it is possible that monitoring VKA and their safety could be improved by measuring either coagulation factor II or coagulation factor X. Measuring and monitoring coagulation factor II alone (native prothrombin antigen) has been shown to be very successful and more accurate than PT in clinical studies (Furie, Liebman et al. 1984; Xi, Beguin et al. 1989; Furie, Diuguid et al. 1990; Kornberg, Francis et al. 1993). The native prothrombin antigen assay, however, is not as convenient for laboratories as the PT or P&P. Monitoring coagulation factor X alone has not been tested in patients to our knowledge.

   We have investigated (Thrombosis Research 2012, in press.) the influence of each of the VKD factors on coagulation induced by dilute (trace amount) thromboplastin using rotational thromboelastometry (ROTEM). The dilute thromboplastin is by some scientists considered to reflect physiological clotting better than the 17,000 fold stronger thromboplastin used in the current PT assays. ROTEM measures coagulation in more detail than the traditional clotting times, i.e. initiation phase (clotting time), propagation phase (bulk clot formation) and stabilization phase. Taken together, these experiments using ROTEM show that the ROTEM clotting time (CT, initiation phase), ROTEM maximum velocity of clot formation (Max Vel, propagation phase) and the ROTEM MCF (stabilization phase) were affected much more by mildly and moderately low concentrations of factors II or X (at activity levels that are present during ideal anticoagulation) than by identical FVII or factor IX activity levels. Factors VII and IX mainly influenced the ROTEM parameters at very low concentrations (< 5%) which only occur during marked over anticoagulation. We conclude from these experiments that when coagulation is initiated with trace amount thromboplastin (physiological coagulation), reduction in concentration of coagulation factors II or X has much more effect on the development of a clot than do the concentrations of factors VII or IX except at extremely low activity levels of the latter two factors. Our experimental results may indicate that in patients anticoagulated with warfarin, the factor VII activity is a confounding source of variation in INR (PT) results. Hence, measuring the influence of factor II or X or both II and X simultaneously on fibrin formation might reduce INR fluctuation. Possibly, this could better predict the antithrombotic effect than the current tests do. Thus, measuring II or X or both simultaneously could possibly improve monitoring and dosing during VKA therapy. This has led to our invention of a new test, "Fiix-PT" (or Stuart-Prothrombin time) , which measures the combined action of factor II and factor X activity on clotting induced by tissue thromboplastin in PPP. We are not aware of any other scientists investigating this method and a local patent has been granted with international patent applications in process (March 2012). The Fiix-PT is not affected by reduced activity of any other coagulation factors. In comparison, the traditional test method (Quick PT or Owren´s PT) measures the combined effect of factors II, VII and X on fibrin formation in patients taking VKA.

   2.2 Hypothesis: Based on our experimental results as well as the results of others(Xi, Beguin et al. 1989; Zivelin, Rao et al. 1993; Brummel, Paradis et al. 2001), we hypothesize that the Fiix-PT may reflect anticoagulation and the antithrombotic effect of VKA at least as accurately as than the current PT based tests do (PT, P&P).

3. STUDY OBJECTIVE The objective of the study is to evaluate the efficacy and safety of Fiix-INR as a monitoring test compared to the current PT based assays used to monitor patients treated with VKA to prevent thromboembolism.

4. STUDY OVERVIEW This is a prospective, single-center, randomized, double-blinded, non-inferiority study(D'Agostino, Massaro et al. 2003; Scott 2009) for the efficacy and safety of Fiix-PT as a monitoring method in VKA treated patients involving 1200 patient years of observation.

   4.1 Participants: All ambulatory patients 18 years or older currently on or starting on warfarin to treat and prevent thromboembolism during the study period at the anticoagulation management center (AMC) at the Landspitali University Hospital in Reykjavik, Iceland with a treatment goal of INR 2-3 will be eligible for participation and randomization irrespective of indication for anticoagulation, age or concurrent medication use. Thus, the study population will have various (or mixed) indications for OA but the majority likely will have atrial fibrillation. Both current patients and new patients will be randomized over a 3-6 month period until adequate number of patients have entered the study. After patients sign informed consent they will be randomized (see METHOD of randomization in chapter 7) to either Fiix-PT (test method, Fiix INR) or PT (PT-INR based on Quick PT method as a control method). We will measure the PT- INR as well as the Fiix - INR in the test group for the sake of safety.

   4.2 Randomization standard (PT-INR) vs. new test (Fiix-INR). The aim is to randomize 1200 clients of the AMC into two identically sized monitoring groups, Fiix-INR (test group) and INR (control group). Each individual will be assessed monthly for up to 24 months (an expected average of 18 months) until a total of 600 patients years of observation has been reached in each group.

   4.3 Time frame: The clinical observation period for each participant will be 18 months on average.

   4.4 Endpoints to be recorded: see Chapter 7 for details A.Clinical endpoints. Rate of all TE or major bleeding. The major efficacy endpoint is symptomatic diagnosed TE events or death from TE and the major safety endpoint, is major bleeding, including death from major bleeding (Palareti 1996; Abdelhafiz and Wheeldon 2004). TE and major bleeding as a combined endpoint. Death and causes of death from natural causes. All bleeding (major and any other reported bleeding) B. Convenience/surrogate endpoints, eg. time within target range, number of tests within target range, number of tests during study period, number of tests needed to achieve stable anticoagulation.

   4.5 Study management: The study will be managed from day to day by an Executive Committee (EXC) . All clinical diagnoses will be reviewed by a blinded Independent Adjudication Committee (IAC) and Data Safety Monitoring Board (DSMB). The IAC and DSMB will be blinded to which test group the patient with event for review belongs. All suspected recurrent TEs, deaths, as well as all episodes of bleeding and other possible vascular events will be evaluated by the IAC. The patent holders (PTÖ, BRG) will not be on the IAC due to potential bias. Only adjudicated results will be the basis for the final analyses. The DSMB will also monitor the study progress and patients' safety during the study and give recommendations to the EXC.

   Every time a patient has a contact (eg INR measurement or dosing instructions), the dosing staff or study nurse will ask questions by exactly following the case review form (CRF, 14.2). If the questions are answered with a "yes", the study nurse will without delay contact that particular patient for more detailed information (i.e. HCRU form see: Event report form / Health Care resource utilization (HCRU)), and, as appropriate, will also collect necessary case record documents for the IAC.

5. STUDY POPULATION, RANDOMIZATION AND BLINDING

5.1 Planned number of patients. See 11.3.1 5.2 Inclusion criteria

1. Current or new patients 18 years or older currently on or starting VKA with INR treatment target of 2-3
2. Willingness to participate as volunteers and to signing an informed consent 5.3 Exclusion criteria

1. Age under 18 years 2. Patients unable to sign informed consent

5.4 Randomization and blinding 5.4.1 Random allocation to experimental testing (Fiix-PT, Fiix-INR) group and standard testing (PT, PT-INR) group will be done by the study nurse or coagulation laboratory staff member in the outpatient phlebotomy ward when patients come for scheduled visits or in hospital inpatient wards;

1. In current stable patients: at time of signing informed consent. Stable patients have had at least two INR´s within the target range.
2. In new patients: no later than on the 4th day of warfarin administration, i.e. prior to the first dose adjustment as above.

Note that the first dose change/first dose adjustment decision of warfarin must be based on post-randomization Fiix-PT or PT result.

5.4.2 Randomization After patients sign informed consent they will be randomized to either Fiix-PT (test method, Fiix INR) or PT (PT-INR based on Quick PT method as a control method). Randomization will be done manually at time of signing the informed consent, ie the participants will draw a labeled color card with an assigned study number (CRF number) which will randomize them to either the test group or the control group. The coagulation laboratory (supervised by BRG) will maintain a list of study participants during the study. They will be responsible for randomization (color assignment) following the signing of informed consent. Following randomization, each patient will be assigned a study participant number. The laboratory will based on this participant number and color code as appropriate direct each patient´s sample to test A (Fiix-INR) or test B (PT-INR) which will be measured in the Landspitali coagulation laboratory using the STA-R instrument (Stago).

5.4.3 Blinding. The study will be double-blinded, i.e. all assessors of clinical effect and dosing staff as well as patients will be blinded to the monitoring test-method. Blinded test results (INR) will be administered by the coagulation laboratory staff which cannot be blinded since they do the laboratory testing. The coagulation laboratory will measure INR (PT) on one STA-R instrument and Fiix-INR (Fiix-PT) on another STA-R instrument. All results, however will be reported out as "INR". The patients will not be informed on which test is used to monitor their anticoagulation and only a value ("INR") will be reported to patients. The test results will also be reported electronically as "INR" to the dosing staff (nurses, biomedical scientists, physicians) and study nurse who will not have knowledge on which test is used to monitor each patient. The data will be managed by the study nurse who will not have access to monitoring test allocation. All formal assessors of clinical events and secondary endpoints will also be blinded to test allocation. The dosing staff will dose patients based on INR, i.e. using the DAWN anticoagulation management software (Onundarson, Einarsdottir et al. 2008).

5.5 Prior and concomitant anticoagulant medication 5.5.1 Anticoagulant medication prior to randomization Therapeutic dosages of (LMW)heparin/fondaparinux are allowed concomitantly with warfarin in new patients or during bridging therapy in association with procedures.

6 STUDY TESTS 6.1 INR and Fiix-INR Current patients on stable warfarin therapy (see 5.4.1) can be included. In new patients, no later than after the first dose of warfarin should the dosing be managed by the AMC staff at Landspitali. The VKA dosage will be adjusted to maintain the Fiix-INR or the INR (PT-INR) with the therapeutic target of 2.5 (range 2.0-3.0).

6.1.1 PT-INR The INR is calculated by the following formula: INR= (Patient PT/mean normal PT)ISI .

6.1.2 Fiix-INR Fiix-PT will be measured using Neoplastin CI plus® and Fiix depleted plasma as described in patent application P8151IS00, March, 8th 2011. Fiix-INR will be calculated in the same manner as the PT-INR is calculated. ISI is calculated from DEKS calibrators as in 6.1.1.

6.2 Packaging, labeling, and storage The test reagents will be stored according to their package inserts and the Fiix deficient plasma will be stored in aliquots at -70°C.

6.3 Accountability and study compliance All testing, dosing and follow-up must be done in accordance with the protocol and group assignment. During emergency testing both tests should be measured on all patients. When the laboratory is not sure both tests should be measured. All samples drawn during weekend and holidays should be kept until coagulation laboratory staff has decided that they can be discarded.

6.3.1 VKA treatment monitoring The INR and Fiix-INR should initially be measured every 2 to 4 days and when stable (at least two consecutive tests within target range) no less that every 8 weeks.

7 STUDY PROCEDURES 7.1 Study period This is the period between randomization (first testing with assigned test = Day 1) and final assessment of each individual. Data will be retrieved continuously as described above. After 3, 6, 9 and 12 months "interim" (safety, efficacy) analyzes will be performed (see Chapter 11). We will randomize 1200 clients of the AMC into two identically sized monitoring groups, Fiix-INR (test group) and INR (control group). Each individual will be assessed monthly for up to 24 months (an expected average of 18 months) until a total of 600 patients years of observation has been reached in each group. The study may also be stopped by the recommendation of the DSMB if a clinically significant difference in outcome becomes evident at an earlier time-point indicating inferiority of the new method.

7.1.1 Patient written information Prior to enrollment potential study participants will receive an introductory information letter as well as informed consent form detailing the study and symptoms suggestive of new or recurrent thromboembolism or major bleeding.

7.1.2 Patient contacts Data will be retrieved continuously and each patient will be contacted on a monthly basis. As described in detail in the protocol, in the case of suspected bleeding or thrombotic events or death data will be retrieved from case records, hospital summaries, autopsy reports or other records from physicians. When suspected serious events occur, the patients will be immediately referred to appropriate level care in the hospital.

7.1.2.1 Initial CRF form (see: Initial Case Report Form (Initial CRF) form) When the patient has signed an informed consent, an initial CRF form (ICRF) will be filled in.

7.1.2.2 Follow-up CRF form (see: Follow-up Contact Case report (review) form (Follow-up CRF)) Every time a patient has a contact with the AMC (e.g. INR measurement or dosing instructions), the dosing staff or study nurse will ask questions by exactly following the case review form (CRF, 14.2). If the questions are answered with a "yes", the study nurse will without delay contact that particular patient for more detailed information (ie HCRU form, see: Event report form / Health Care resource utilization (HCRU)). The study nurse will then, as appropriate, collect necessary documents for the independent adjudication committee (IAC), see chapter 9.3.3. During each patient event, the following will be systematically checked by interviewing patients and reviewing the hospital chart for thromboembolism, bleeding events, concomitant medication, in particular antiplatelet agents such as aspirin or clopidogrel and NSAID use and compliance with monitoring 7.1.2.3 Observational period All patients will have a 5 day observational period after discontinuing from study.

7.2 Efficacy and safety outcome criteria for the IAC 7.2.1 Assessment of clinical outcomes All suspected clinical outcomes i.e. recurrent TE, bleeding, death and vascular events, will be notified expedited to the IAC in a manner blinded to which test was applied. Blinded review of the event by the IAC should be completed within 2-4 weeks after occurrence of the event. The primary efficacy outcome is symptomatic or fatal recurrent arterial or venous TE event during the study period. The primary safety outcome is bleeding. Prespecified criteria for these events are excluded from this abbreviated protocol version.

7.3 Surrogate endpoints/Convenience endpoints

1. During initiation of VKA therapy: i. Number of tests needed to reach stable INR, i.e. until the first of two consecutive INR values are obtained within the target range. ii. Time until stable INR within target range
2. During stable phase: i. Total number of tests per treatment year ii. Dose adjustment frequency (interval or number of dose adjustments during observation period) iii. Percent tests within treatment range over observation period iv. Percent time within target range (TTR) over observation period

7.4 Measures to minimize bias A series of measures will be implemented to minimize the potential for bias in this study as detailed in the full protocol.

7.5 Approach to the bleeding patient If a patient has a serious bleed during study, the following routine measures should be considered.

Refer the patient immediately to the emergency department Delay or discontinue the next anticoagulant dose and consider vitamin K administration and usual treatment for bleeding, including blood transfusion, prothrombin complex concentrate administration, recombinant factor VIIa (NovoSeven and fresh frozen plasma

8 PREMATURE DISCONTINUATION OR TEMPORARY DISRUPTION

Subjects may be withdrawn from the study for the following reasons:

At their own request or at the request of their legally acceptable guardian. If, in the investigator's opinion, continuation on anticoagulation would be detrimental to the subject's well-being.

This is further detailed in the full protocol 9 DATA QUALITY ASSURANCE 9.1 Data quality The monitoring and auditing procedures defined below will be followed.

9.2 Documentation Entries made in the CRF must be either verifiable against source documents, or have been directly entered into the CRF, in which case the entry in the CRF will be considered as the source data. The source data parameter to be verified and the identification of the source document must be documented. The study file and all source data will be retained for 5 years following final publication of study results when source data can be destroyed. No published result will be traceable to particular persons.

9.3 Study administration/committees 9.3.1 Executive committee (EXC) The EXC has the overall scientific responsibility of the study. Its tasks and responsibilities are to create and approve the final protocol, co-author protocol amendments whenever necessary, ensure a scientifically sound and safe conduct of the study, including by abiding to DSMB recommendations, decide on the DSMB recommendations, review and approve the statistical analysis plan, guarantee the integrity of data collection and analyses and to decide on the publication- and presentation policy of the final results and ancillary studies 9.3.2 Study management and coordination committee (SMCC) The SMCC has the overall clinical responsibility of the study. Its tasks and responsibilities are to review the protocol, review the CRF, support the EXC by advice, monitor progress of study enrollment, address/resolve study management problems and participate in the analysis and presentation of the results.

Member (author group):

The Independent adjudication committee (IAC) has the following roles: A. Blinded review of clinical events (IAC- function), i.e. all suspected recurrent thromboembolic events, bleeding of any kind, any vascular events and all deaths during the study period and the observational period will be evaluated by IAC .The study nurse will provide the IAC with all relevant documentation related to the events but the IAC will be blinded as to the monitoring method of each patient while evaluating events. Adjudication results will be the basis for the final analyses. and B. Data safety and monitoring (DSMB-function), i.e. this committee also has the responsibility to provide the EXC with recommendations related to the protection of the patients' safety, including stopping recruitment and stopping the study. For this purpose, the DSMB will review all incidences of recurrent thromboembolism and bleeding every 3 months, see Chapter 11.

9.3.4 External monitoring of study Adherence to study protocol will be monitored by an external agency. 10 ETHICAL AND LEGAL ASPECTS 10.1 Ethical committees (EC) The study will we conducted in accordance with the Helsinki declaration. Documented approval from appropriate ethical committees (EC) (applications will be sent to National Bioethics Committee (NBC) ("Vísindasiðanefnd"), The Data Protection Authority ("Persónuvernd") and from the Landspitali Medical Director will be obtained prior to the start of the study, according to good clinical practice (GCP), local laws, regulations and organizations. Regulatory authority approvals/authorizations notifications will be in place and fully documented prior to study start.

10.2 Ethical conduct of the study The procedures set out in this protocol, pertaining to the conduct, evaluation, and documentation of this study, are designed to ensure that the investigators abide by GCP Guidelines and under the guiding principles detailed in the declaration of Helsinki. The study will also be carried out in keeping with applicable local law(s) and regulation(s). This may include an inspection by regulatory authority representatives at any time, including allowing direct access of source documents to the regulatory authority.

10.3 Subject information and consent A core information and informed consent form will be provided. Written informed consent must be obtained before any study specific procedure takes place. Participation in the study and date of informed consent given by the subject should be documented appropriately in the subject's files.

10.4 Insurance All subjects participating in the study will be covered by the Landspitali Hospital patient insurance in line with applicable laws and/or regulations.

10.5 Confidentiality All records identifying the subject will be kept confidential and, to the extent permitted by the applicable laws and/or regulations, will not be made publicly available. Only the subject number will be recorded in the CRF. A list with names and subjects numbers will be kept separately by the EXC and its employees. If the results of the study are published, the subject's identity will remain totally confidential. The investigator will, however, maintain a list to enable subjects' records to be identified.

10.6 Archiving of data The investigator will arrange for the retention of the study documentation file for 5 years. All data and documents should be made available if requested 11 STATISTICAL AND ANALYTICAL METHODS 11.1 Statistical analysis plan (SAP) The plan is described in the following sections. However, the SAPs will allow protocol amendments or unexpected issues in study execution or data that affect planned analyses.

11.2 Analysis populations The intention-to-monitor ("ITM") population will consist of all patients who have been randomized, i.e., when the patient number and allocated monitoring method are recorded in study-data base.

Patients will be analyzed by the monitoring method assigned. Patients will contribute to patient years during the time they are in the study For patient year analysis the following patients will be censored at the last day the patient had a complete assessment for study outcomes within the intended treatment duration. Patients who did not have a thromboembolic or bleeding event during the time of the predefined randomized test duration, patients lost to follow-up patients who died because of other reasons than thromboembolic events patients who withdrew informed consent before the end of the predefined treatment duration and who did not have an efficacy or safety outcome. The per-protocol-population (PPP) will consist of all randomized patients without any major deviation from the protocol. The following deviations will lead to exclusion from the PP population: Patients not tested according to allocation or no VKA treatment received at all 11.3 Study population/sample size and power/ demographics and patient characteristics Baseline characteristics: Demographic characteristics (including age, sex, indications for anticoagulation, concomitant drug use) will be summarized by treatment group using descriptive statistics and the groups will be compared with regard to baseline characteristics using traditional statistical methods. As, all major clinical events are likely to be admitted to Landspitali, the only acute care hospital in the Reykjavik area, loss to follow-up is expected to be negligible. Therefore, no adjustment of sample size is considered necessary. Patients excluded from the analysis population in the per protocol analysis will be listed by test method and reason for exclusion. The number and percentage of randomized patients who discontinued from the study prematurely will be tabulated by main reason for discontinuation and treatment group.

11.3.1 Non-inferiority margin estimation This study is powered for analysis of successful anticoagulation treatment and monitoring. For this calculation we have used data from studies on the clinical outcome of patients with mixed indications for anticoagulation, monitored and dosed by anticoagulation management centers (Palareti, Leali et al. 1996; Wilson, Wells et al. 2003; Abdelhafiz and Wheeldon 2004; Menendez-Jandula, Souto et al. 2005). In these studies the combined endpoints of major bleeding and thromboembolic events ranged from 2.7 to 7.3 per 100 patients years and event rate for thromboembolic events was 0.9 to 5.4 per 100 patient years (0.9-5.4%) with the lower margin of the 95% CI ranging form 2.9 to 8.2%. The largest observational study (Palareti, Leali et al. 1996; Wilson, Wells et al. 2003; Abdelhafiz and Wheeldon 2004; Menendez-Jandula, Souto et al. 2005) and the only randomized study (Palareti, Leali et al. 1996; Wilson, Wells et al. 2003; Abdelhafiz and Wheeldon 2004; Menendez-Jandula, Souto et al. 2005) had an TE event rate of 3.5 and 5.4% respectively. We therefore chose to use 3% TE-event rate as our expected TE-event rate (97% success rate) and the non-inferiority of 5.5% (94.5% success rate for efficacy). We consider this 2.5% difference one that is clinically significant, yet within the 95% CI for event rate in some prior studies. Based on these considerations, the appropriate number of participants needed to demonstrate statistical non-inferiority of clinically important events for the test method with an 80% certainty after one year of observation are about 600 (576) patients in each group, i.e. 600 (576) patient years in each group.

11.4 Major event analysis (Efficacy and safety analysis) Major events analyses, primarily the TE event rate (efficacy) but also the secondary end points of the rate of bleeding and major bleeding (safety) as well as the combined endpoint of TE events and major bleeding events will be based on the ITT population. For the primary and secondary endpoints, event free survival will be compared in the two groups using Kaplan-Meier curves. Optionally, the time to the first events will be analyzed using Cox's proportional hazard models.

11.5 Interim analysis 11.5.1 Study monitoring The study progress will be monitored continuously. Interim analysis for safety purposes is planned at 3, 6, 9 and 12 months. Based on this, risk-benefit will be evaluated by the DSMB.

11.6 Surrogate endpoint analysis Median time within target range, time until stable INR as well as other surrogate parameters will be evaluated using the Mann-Whitney test for numeric rows or other appropriate statistical tests.

Inclusion Criteria:

• age 18 or over,
• INR treatment goal of 2-3,
• ability to sign informed consent

Exclusion Criteria:

• nursing home patients