Introduction: Anticoagulation with vitamin K antagonists (VKA) requires monitoring of their effect, traditionally with the prothrombin time (PT) that is affected by VKA influence on coagulation factors (F) II, VII and X. Rapid fluctuations in factor VII, which has a short half-life, contribute to the PT (INR) variation but not to the antithrombotic effect that depends mainly on reductions of FII and FX. This was lately confirmed by the Fiix-trial that showed that monitoring warfarin with Fiix-PT (affected only by FII and FX) improved anticoagulation stability. Here, we assessed anticoagulation variability in relation to the occurrence of thromboembolism and bleeding in patients monitored with Fiix-PT or PT.

Methods and materials: This is a subgroup analysis of the Fiix-trial, a single-center, double blind, prospective, randomized controlled clinical trial, comparing outcomes in patients in whom warfarin was monitored with either Fiix-PT/Fiix-INR (Fiix-warfarin patients) or PT/INR (PT-warfarin patients). Patients on warfarin, 18 years and older, with target INR range of 2.0 - 3.0, were randomized and assessed for occurrence of clinically relevant vascular events (CRVE), i.e. thromboembolism (TE), major bleedings (MB) and other non-major clinically relevant bleedings. Using an intention-to-monitor method, we assessed test parameters, dosing, time in range (TTR) and the variance growth rate (VGR) of the INR (an INR fluctuation index) in relation to occurrence of CRVE.

Results: The median observation time was 1.4 years in 572 patients managed with Fiix-warfarin and 571 with PT-warfarin. CRVE occurred in 115 Fiix-warfarin patients and 132 PT-warfarin patients (PNI=0.0066). MB and TE occurred in 19 vs. 21 (PNI=0.0142) and 10 vs. 19 (PNI=0.0002) patients, respectively. There were 11,026 monitoring tests in the Fiix-arm and 11,499 in the PT-arm. Patients suffering CRVE had significantly more frequent monitoring tests and shorter intervals between tests than those without. Patients with CRVE also had significantly greater dose changes (p<0.0001 in both arms). The median TTR was lower with PT-warfarin than with Fiix-warfarin and patients with CRVE had lower TTR than those without in both study arms (Fiix-warfarin 79% vs. 82%, p=0.0441 vs. PT-warfarin 75% vs. 80%, p=0.0004). The lowest median TTR was observed in PT-warfarin patients suffering from MB (73%) or TE (62%). There was consistently more INR-fluctuation (higher VGR, here shown by B1 method measuring INR jumps from one test to the next) with PT-warfarin than with Fiix-warfarin. Also, patients with CRVE had VGR of 0.35 vs. 0.21 (p=0.0643) and without CRVE 0.21 vs. 0.17 (p=0.0146), respectively. The fluctuation was particularly high in both PT-warfarin and Fiix-warfarin patients suffering from MB (0.59 and 0.31, n.s.). PT-warfarin patients with TE had VGR 0.50 vs. 0.20 with Fiix-warfarin (p=0.0051). Finally, the median INR observed at the time of major events corresponded to the risk of bleeding and TE in the Fiix arm, and with risk of TE in the PT arm.

Conclusions: Fiix-warfarin is a more stable anticoagulant than PT-warfarin. The significantly lower INR variation in Fiix-warfarin patients with TE is in agreement with the reduced long-term thromboembolism observed in the Fiix-trial. Monitoring warfarin with the Fiix-PT (Fiix-INR) instead of the PT (INR) and paying particular attention to patients demonstrating anticoagulation instability could improve the clinical outcome of patients on warfarin further.


Gudmundsdottir:Fiix Diagnostics Ltd.: Equity Ownership, Patents & Royalties: Patent pending for Fiix prothrombin time. Onundarson:Fiix Diagnostics Ltd: Equity Ownership, Patents & Royalties: Patent pending status for Fiix prothrombin time.

Author notes


Asterisk with author names denotes non-ASH members.