Thrombosis Research 130 (2012) 674–681

Contents lists available at SciVerse ScienceDirect

Thrombosis Research

j ourna l homepage: www.e lsev ie r .com/ locate / thromres

Regular Article

Critical role of factors II and X during coumarin anticoagulation and their combinedmeasurement with a new Fiix-prothrombin time☆

Brynja R. Gudmundsdottir a, Charles W. Francis b, Alexia M. Bjornsdottir a,Moa Nellbring a, Pall T. Onundarson a,⁎a Department of Laboratory Hematology and Hemostasis Center, Landspitali University Hospital and University of Iceland School of Medicine 101 Reykjavik, Icelandb Department of Hematology and Oncology, Strong Memorial Hospital/University of Rochester School of Medicine and Dentistry, Rochester, N.Y., USA

☆ Supported by grants from the Landspitali Science⁎ Corresponding author at: Department of Laboratory

Disorders, K-building, Landspitali Hringbraut, 101 Reyk1000/5010; fax: +354 543 5539.

E-mail address: pallt@landspitali.is (P.T. Onundarson

0049-3848/$ – see front matter © 2011 Elsevier Ltd. Alldoi:10.1016/j.thromres.2011.12.013

a b s t r a c t

a r t i c l e i n f o

Article history:Received 26 October 2011Received in revised form 1 December 2011Accepted 6 December 2011Available online 4 January 2012

Vitamin K antagonists (VKA) are monitored with prothrombin time (PT) based assays that are equally sen-sitive to reductions in factors II, VII or X. We compared the effect of vitamin K dependent (VKD) coagulationfactors on PT and also on rotational thromboelastometric (ROTEM) parameters. The PT was equally sensi-tive to reductions in factors II, VII or X but ROTEM parameters correlated poorly with the PT in anticoagu-lated patients´ plasmas. ROTEM parameters were more affected by mild and moderate reductions in FII orFX than by FVII or FIX which had little influence except at very low coagulant activity. We developed a mod-ified PT that was sensitive only to reductions in factors II and X. The Fiix-PT (Fiix-INR) correlated well withPT (INR) but the Fiix-INR fluctuated less than the INR in an anticoagulated patient reflecting its insensitivityto FVII. The ROTEM results suggest that mild to moderate reductions in factors II or X are more importantduring clot formation than factors VII or IX. Reductions in FII and X may better reflect anticoagulationwith VKA than FVII or IX. The new Fiix-PT may more accurately reflect the degree of therapeutic anticoagu-lation in patients treated with VKA than the current PT which is subject to a confounding variation causedby FVII.

© 2011 Elsevier Ltd. All rights reserved.

Introduction

During initiation of treatment with VKA, the coagulant activity(activity) of each VKD coagulation factor declines at different ratesreflecting their half-lives which vary from 3.5 hours for FVII to 72hours for FII [1,2]. The degree of anticoagulation is monitored withprothrombin time (PT)-based tests such as the Quick PT [3] or theOwren PT (prothrombin-proconvertin time, P&P-test) [4], the lattertest being mainly used for monitoring anticoagulation in the Nordiccountries, Holland and Japan. The international normalized ratio(INR) based on either PT variant will lead to practically identical re-sults [5]. Both PT assays use undiluted thromboplastin and recalcifica-tion to activate coagulation in citrated platelet poor plasma (PPP). TheOwren PT differs in that it is done on diluted test plasma whichmakesit less sensitive to heparin and lupus anticoagulant than the PT [6].Since absorbed bovine plasma is added, it corrects for all factor defi-ciencies other than those of vitamin K dependent factors makingthe test insensitive to environmental temperature which can affectFV. Both PT variants are equally sensitive to reduced activity of factors

Fund.Hematology and Coagulationjavik, Iceland. Tel.: +354 543

).

rights reserved.

(F) II, VII and X but are not sensitive to reductions in activity of FIX.Additionally, the Quick PT is sensitive to reduced concentrations offactors V and fibrinogen that are not affected by VKA but may be asource of variation for other reasons.

The PT is presumed to accurately reflect the antithrombotic effectof VKA. However, prior studies have indicated that the antithromboticeffect of FVII may be of minor importance compared to the activity ofFX and, in particular prothrombin [7,8]. Also, factor VII does not startto prevent thrombin generation until levels are well below 5% [7,9].Since FVII has the shortest half-life, fluctuations in the PT may simplyreflect FVII activity changes rather than a true change in antithrombo-tic effect. This may have therapeutic implications, as studies have sug-gested that measuring prothrombin alone as native prothrombinantigen by ELISA more accurately reflects the antithrombotic effectof VKA than does the PT [8,10–12].

We evaluated the roles of each VKD coagulation factor on clottingin vitro using PT assays and also rotational thromboelastography(ROTEM), a method that perhaps more accurately reflects the com-plex in vitro clotting process by measuring as well the rate of the fol-lowing bulk clot formation and the final clot strength [13]. In theROTEM experiments coagulation was activated with highly dilutethromboplastin as described previously [13]. The findings suggestedthat FII and FX play a dominant role in clotting, and we, therefore, de-veloped a new PT variant, Fiix-PT, which measures the combined ef-fects of only FII and FX on fibrin formation.

675B.R. Gudmundsdottir et al. / Thrombosis Research 130 (2012) 674–681

Materials and methods

Plasma

Blood was drawn from patients on warfarin anticoagulation andfrom twenty healthy donors (10 men, 10 women). Using a minimalstasis technique, the blood was drawn through a 21 G needle into 1/10 vol/vol 3.2% (0.109 mol/L) buffered sodium citrate Vacuette®tubes (Greiner, Kremsmunster, Austria). Platelet poor plasma (PPP)was prepared by centrifugation of the blood at 4 °C for 15 minutesat 2,500 g, removal of supernatant and recentrifugation for 15 mi-nutes at 2,500 g. The pooled normal plasma (PNP) was prepared bypooling the recentrifuged normal plasma and stored in 0.5 mL ali-quots at−80 °C. Immunodepleted plasmas were selectively depletedof coagulation factors II, VII, IX or X or of both FII and X (HaematologicTechnologies Inc. (Essex Junction, Vermont, USA), and there was noresidual inhibitory activity in the plasmas. Adsorbed plasma wasmade by incubating 9 volumes of PNP with 1 volume aluminium hy-droxide gel (AL(OH)3 1 g in 4 ml H2O) (Sigma Alderich Chemie GmbHproduct number A8222, Steinheim, Germany).

Frozen-thawed platelets and washed platelets were made fromfreshly donated blood bank platelet concentrate in acid-citrate-dextrose (ACD). For both, the platelet rich plasma (PRP) was centri-fuged for 15 minutes at 150 g and the platelet-rich pellet was thenresuspended 1:1 vol/vol in TBS buffer (6.055 g Trizma base fromSigma, Seetze, Germany, 8.766 g NaCl, pH 7.4, 1000 mL H2O). Afterresuspension the tubes were filled with TBS buffer and re-centrifuged for 10 minutes at 2,500 g. The supernatant was removedand the pellet was resuspended twice. The final suspension wascounted automatically using Sysmex 4000 (Sysmex Corp., Kobe,Japan), and the final solution was adjusted to a platelet count ofabout 1000x109/L. To make platelet phospholipids the final solutionwas frozen at −80 °C, thawed again and then aliquoted and frozenagain. Alternatively, the washed platelets were stored at room tem-perature and used for experiments within two days.

Reagents

Factor deficient plasmas used for assaying coagulation factorswere obtained from Diagnostica Stago (Asnieres, France). Hepes buff-er (Hepes 20 mM, NaCl, 150 mM, with and without CaCl2 200 mM, pH7.4) was obtained from Bie & Berntsen Inc. (Herlev, Denmark). Re-combinant thromboplastin (Innovin®) was obtained from Dade Beh-ring (Marburg, Germany). The Innovin was reconstituted and storedin 100 μl aliquots at −80° centigrades in Eppendorf tubes. One tubeof Innovin was reconstituted for each experiment and 10 μl Innovinwere mixed into 990 μl of Hepes CaCl2 to make a 1:100 dilutionwhich was further diluted by mixing 100 μl of the 1:100 dilutionwith 500 μl of Hepes CaCl2 to make a 1:600 working solution.

ROTEM experiments

Dynamic blood clot formation profiles were recorded by aROTEM® Thromboelastometry Coagulation Analyzer (Pentapharm,Munich, Germany) as described in detail elsewhere [13–15]. Howev-er, instead of whole blood we used citrated PPP. In brief, the reactionmixture contained 280 μl of citrated plasma+20 μl of buffer+20 μldilute thromboplastin and CaCl2. All results shown are based on qua-druplicate experiments. The ROTEM measurements were based onactivation of coagulation with highly diluted thromboplastin (finalreaction mixture dilution of TF 1:17,000) [13]. Assessment of plasmaclot formation was based on standard thromboelastometry parame-ters such as the clotting time (CT), maximum velocity of clot forma-tion (MaxVel) and the maximum clot firmness (MCF). The CTcharacterizes the initiation phase of clot formation, the MaxVel

shows the propagation phase of clotting and the MCF reflects the sta-bilization phase.

ROTEM experiments were done with factor deficient plasma (de-ficient in FII, FVII, FIX or FX) mixed with varying proportions of PNPto make final VKD coagulation factor concentrations of 1, 2, 4, 8, 16,32 and 50%. The experiments were done without and with addedplatelet phospholipid or washed blood bank platelets. The experi-mental cuvettes were mounted in an aluminium cup holder thatwas thermostatically controlled at 37 °C. A mixture of 270 μl of eachplasma dilution and 30 μL of buffer, washed platelets (final count720x109/L) or platelet phospholipid solution were pipetted into themeasuring cups. The measurement was started by adding 20 μl ofthe recombinant thromboplastin 1:600 dilution and placing the cupon a 6 mm diameter pin attached to the lower end of a vertical axis.

Quick- and Owren-prothrombin time

The Quick PT and the Owren PT used the STA-R coagulation ana-lyzer (Diagnostica Stago, Asnieres, France). The Quick PT was per-formed on undiluted PPP using the STA-Néoplastine CI plus® withan international sensitivity index (ISI) of 1.3. The Owren PT wasdone on diluted plasma (1:7 in buffer) using the STA-SPA 50 reagentwith ISI of 1.01 (both from Diagnostica Stago, Asnieres, France). TheOwren PT reagent is similar to the PT reagent but also containsabsorbed bovine plasma as a source of coagulation factor V and fibrin-ogen and, thus, is only sensitive to the activity of coagulation factorsII, VII and X. The Owren PT coagulant activity percentage was calcu-lated based on a standard curve made from the clotting times of seri-ally diluted pooled plasma in Owren buffer [5]. The INRwas calculatedas previously described [5,16–18]. The Quick PT and Owren PT resultin equivalent INR´s [5].

Factor II and X prothrombin time (Fiix-PT, Stuart-prothrombin time)

PPP immunodepleted of both FII and FX (Fiix-depleted plasma)was specially made by double immunodepletion of PNP (Haematolo-gic Technologies Inc., Essex Junction, Vermont, USA). In different ex-periments, the Fiix-depleted plasma was mixed with test plasma invarious proportions and with different test plasma pre-dilutions inOwren buffer as detailed in Results. Standard PT reagent (thrombo-plastin and calcium, Néoplastine CI plus®, as above) was thenadded and clotting times were measured.

Statistical analysis

The GraphPad Prism 5.0 statistical software (GraphPad Inc, CA,USA) was used for graphing and for statistical calculations includingnon-linear regression analyzis. Experimental data is shown graphical-ly as mean±standard error of the mean (SEM).

Results

Dependence of PT on activity of individual of VKD factors

The relationship between PT and reduced activity of factors II, VII,IX and X in PPP is shown in Fig. 1 using experiments in which PNPwas mixed in varying amounts with plasma completely immunode-pleted of a single VKD factor. Both PT methods were equally sensitiveto the activity of factors II, VII and X throughout the activity range.

ROTEM parameters in relation to Owren´s PT in patients on warfarin

The relationship between Owren PT and ROTEM parameters werecompared using samples obtained from 65 patients on VKA (Fig. 2).There was only a moderate correlation between the coagulant activitypercentage and the ROTEM CT (R2=0.27) with considerable

0.1 1 10 1000

20

40

60

80 FII

FVIIFIX

FX

A

Coagulation factor activity (%)

Qu

ick

Pro

thro

mb

in t

ime

(sec

on

ds)

0.1 1 10 1000

50

100

150

200 FII

FVIIFIX

FX

B

Coagulation factor activity (%)

Ow

ren

Pro

thro

mb

in t

ime

(sec

on

ds)

Fig. 1. Quick prothrombin time (panel A) or Owren´s prothrombin time (panel B) in relation to selective reduction in individual vitamin K dependent factor activity.

676 B.R. Gudmundsdottir et al. / Thrombosis Research 130 (2012) 674–681

variation in CT at a similar coagulant activity. Also, there was poorcorrelation between the coagulant activity and MaxVel (R2=0.08)and MCF (R2=0.11). Identical findings were present with the QuickPT (data not shown). ROTEM parameters were then also measuredin PPP that had been depleted of all VKD factors by absorption withAl(OH)3. Absorbed plasma was gradually repleted with VKD factorsby adding PNP, and the results showed that there was a good correla-tion between the PT coagulant activity and the ROTEM parameters(R2=0.79 for CT, 0.64 for MaxVel and 0.77 for MCF) (curves notshown). This difference in results may indicate that VKD factors areunequally reduced in the anticoagulated patient plasmas.

Dependence of ROTEM paramenters on individual VKD factor coagulantactivity

We evaluated the influence of each VKD factor on coagulationusing ROTEM with PPP selectively immunodepleted of each VKD fac-tor which was then gradually repleted by mixing with PNP. The CTwas prolonged more by mild and moderate selective reduction in fac-tors II or X than by similar reduction in factors VII or FIX (Fig. 3). Theresults were similar in PPP (panel A), PPP with added frozen-thawedplatelets (panel B) and in PPP with added washed blood bank plate-lets (panel C). The MaxVel (Fig. 4) was also more dependent on fac-tors II and X than on VII or IX. Finally, MCF (Fig. 5) in PPP (panel A)and plasma with frozen thawed platelets (panel B) was more depen-dent on factors II and X than on VII or IX, but results with washedplatelets (panel C) showed a greater dependence on factor II than X.Taken together, the results showed that factors II and X influencedROTEM results similarly and much more than factors VII and IX.

Fiix - prothrombin time

Since factors II and X have the greatest effect on coagulation in theROTEM experiments at VKD factor activity similar to that during ther-apy, we developed a PPP based test that is sensitive only to reducedactivity of factors II and X and applicable to routine coagulation in-struments. This was done by correcting all deficiencies other thanthose of factors II and X by mixing factor II and X-depleted plasmawith the test sample prior to measuring the PT.

Plasma standards with decreasing activity of coagulation factorsfor experimental standard curves were made as follows: NormalPNP was serially diluted in vitro in Owren´s buffer with added 3.2% so-dium citrate solution in a method similar to the preparation testplasmas and standard curve for Owren´s PT. Thus, serial standardswere made representing coagulant activity ranging from 1.5-50

percent. Each standard was diluted further 1:7 with Owren´s buffer.In order to identify test conditions allowing the measurement of co-agulant activity as low as 1.5% different experiments were done. Foreach experimental standard curve the serial dilutions were mixedwith undiluted FII and X - depleted plasma (Fiix depleted plasma)in proportions shown in Fig. 6 in order to correct for all factor defi-ciencies other than factors II and X. The clotting times of the test plas-ma dilutions were subsequently measured by adding undilutedthromboplastin and calcium as with the PT. The results are shownin Fig. 6 which demonstrates the clotting times obtained in relationto the combined coagulant activity of FII and FX. In order to be ableto measure coagulant activity as low as 3.1% the proportion of (undi-luted) test plasma to Fiix depleted plasma had to be over 0.17 and foractivity as low as 1.6% over 0.28.

We decided to use the Fiix-PT composition identified in the exper-iment shownwith a solid bold line in Fig. 6 (ie proportion of test plas-ma to Fiix-depleted plasma 0.31) since this allows the measurementof Fiix-coagulant activity as low as 1.6%. Using this curve we subse-quently tested samples from 38 patients on stable warfarin therapy(Fig. 7). The Fiix-INR correlated well with the traditional INR(R2=0.87, y=0.98x – 0.01). The intra-assay variation of the Fiix-PTwas 0.8 – 1.6 percent on fresh samples with INR between 1.4 and4.6. The same samples were frozen in aliquots and once thawed, theintra-assay variation was 0.4-0.8 and inter-assay variation was 0.4-0.6 percent.

We also compared results using the traditional INR and the Fiix-INRas well as coagulant activity of factors II, VII and X during warfarin initi-ation in patients. This figure shows that the fluctuation of traditionalINR is faster and larger than that of the Fiix-INR. Fig. 8A and 8B showthat the fluctuation of INR with the traditional INR is highly dependenton FVII activity. This is best evident during the first days when INR risesto levels over target whereas the Fiix-INR hits target with a slower re-sponse. In light of the ROTEM experiment this may indicate that theFiix-INR better reflects the patients´ coagulability. Interestingly, asanticoagulation stabilizes the FX levels is lower (10-20%) than the FIIlevel (20-30%).

Discussion

The results show that factors II and X play a more predominantrole in comparison to factors VII and IX on global clot formation as ob-served in vitro with ROTEM. The poor correlation of ROTEM parame-ters with the INR in patient plasmas likely reflects a discrepancybetween activity of factor VII and that of the more stable factors IIand X as has been previously suggested based on thrombin

0 25 50 75 100 125 1500

1000

2000

3000

R2 = 0.27

A

Owren prothrombin time % (PP %)

RO

TE

M C

T (

seco

nd

s)

0 25 50 75 100 125 1500

5

10

15

20

25

R2 = 0.08

B

Owren prothrombin time % (PP %)

RO

TE

M M

axV

el (

mm

x 1

00/s

eco

nd

)

0 25 50 75 100 125 1500

20

40

60

80

R2=0.11

C

Owren prothrombin time % (PP %)

RO

TE

M M

CF

(m

m x

100

)

Fig. 2. ROTEM parameters in relation to Owren´s PT coagulant activity (%) in 65 sam-ples from patients on stable anticoagulation with warfarin. Panel A shows the clottingtime (CT), panel B shows the maximum velocity of clot formation (MaxVel) and panelC shows the maximum clot firmness (MCF). Note the marked variation in ROTEM re-sults in patients with a similar coagulant activity. Non-linear regression is shown.

0.1 1 10 1000

2000

4000

6000 FIIFXFVIIFIX

A

Coagulation factor activity (%)

RO

TE

M C

T (

seco

nd

s)

0.1 1 10 1000

1000

2000

3000FIIFXFVIIFIX

B

Coagulation factor activity (%)

RO

TE

M C

T (

seco

nd

s)

0.1 1 10 1000

1000

2000

3000FIIFXFVIIFIX

C

Coagulation factor activity (%)

RO

TE

M C

T (

seco

nd

s)

Fig. 3. Quadruplicate experiments showing ROTEM clotting time (CT) in relation to se-lective reduction of individual vitamin K dependent coagulation factors in PPP without(A) and with added platelet phospholipid (B) or washed platelets (C). Note that the CTis more sensitive to reductions in factors II and X than in VII or IX.

677B.R. Gudmundsdottir et al. / Thrombosis Research 130 (2012) 674–681

generation measurements [8,19]. The results suggest the possibilitythat rapid changes in factor VII activity during initiation and followingdose changes of VKA may exaggerate a fluctuation in prothrombintime (INR) that actually does not influence the antithrombotic effector risk of bleeding. This may instead mainly lead to unnecessarydose changes, too frequent monitoring tests and further fluctuation.This confounding effect of FVII can possibly be circumvented by mon-itoring coumarins with the Fiix-PT which is not influenced by factorVII. A randomized double blinded single center clinical study pow-ered to compare the efficacy (thromboembolic events) and safety

(bleeding) of dosing VKA based on Fiix-PT (Fiix-INR) in comparisonto the PT (INR) is now underway.

ROTEM is a global coagulation assay usually done using wholeblood. Whole blood ROTEM results are a function not only of coagula-tion factors but also of platelet function, proteases, inhibitors and fi-brinolysis [20]. In addition to measuring the time to initiation offibrin polymerization (CT, initiation phase) it also measures the

0.1 1 10 1000

2

4

6

8

10FIIFXFVIIFIX

A

Coagulation factor activity (%)

RO

TE

M M

axV

el (

mm

x 1

00/s

eco

nd

)

0.1 1 10 1000

2

4

6

8

10 FIIFXFVIIFIX

B

Coagulation factor activity (%)

RO

TE

M M

axV

el (

mm

x 10

0/se

con

d)

0.1 1 10 1000

5

10

15FIIFXFVIIFIX

C

Coagulation factor activity (%)

RO

TE

M V

max

(m

m x

100

/sec

on

d)

Fig. 4. Quadruplicate experiments showing ROTEM maximum velocity (MaxVel) in re-lation to selective reduction in VKD coagulation factors in PPP without (A) and withadded platelet phospholipid (B) or washed platelets (C). Note how the MaxVel ismore sensitive to reductions in factors II and X than in VII or IX.

0.1 1 10 1000

10

20

30

40 FIIFXFVIIFIX

A

Coagulation factor activity (%)

RO

TE

M M

CF

(m

m x

100

)

0.1 1 10 1000

10

20

30

40 FIIFXFVIIFIX

B

Coagulation factor activity (%)

RO

TE

M M

CF

(m

m x

100

)

0.1 1 10 10010

15

20

25

30 FIIFXFVIIFIX

C

Coagulation factor activity (%)

RO

TE

M M

CF

(m

m x

100

)

Fig. 5. Quadruplicate experiments showing ROTEM maximum clot firmness (MCF) inrelation to selective reduction in VKD coagulation factors in PPP without (A) andwith added platelet phospholipid (B) or washed platelets (C). Note how the MCF ismore sensitive to reductions in factors II and X than in VII or IX, an effect that is some-what attenuated with added washed platelets.

678 B.R. Gudmundsdottir et al. / Thrombosis Research 130 (2012) 674–681

subsequent aspects of coagulation such as the bulk clot formation orpropagation phase (alpha angle or MaxVel) and the maximum clotfirmness reflecting the stabilization phase [13,21]. Therefore, ROTEMmay better describe in vitro clot formation than traditional clottingtime assays such as PT or PTT. ROTEM is still under investigation butmay have utility in predicting bleeding, e.g. in hepatic and

cardiopulmonary bypass surgery or as a bedside tool to tailor pro-hemostatic therapy in patients with congenital and acquired coagulo-pathies [13,22–25].

In our ROTEM experiments, we used highly dilute thromboplastin(1:17,000) to more closely reflect the physiological concentration oftissue factor than does the high concentration used in PT assays[13,19]. We were unable to use whole blood for the ROTEM ex-periments since we had to work with blood deficient in one or

1 2 4 8 16 32 64 12816

32

64

128

256

512

Factor II and X concentration U/dL

Fiix

PT

(se

con

ds)

1 2 4 8 16 32 64 1280.5

1

2

4

8

16

32

Factor II and X concentration U/dL

Fiix

PT

rat

io

Fig. 6. Fiix-prothrombin times obtained on plasmawith increasing coagulant activity of co-agulation factors FII and FX only. The graph shows the clotting times and Fiix-PT ratio ofrepresentative experiments using the following final proportions of test plasma to totalplasma (ie final volume of test plasma+volume of Fiix plasma): 0.09 (1:11, 1part+10 parts); 0.17 (1:6, 1 part+5 parts); 0.25 (1:4, 1 part+3 parts);0.28 (1:3.5, 1 part+2.5 part); 0.31 (1:3.2, 1 part+2.2 parts); 0.40 (1:2.5, 1part+1.5 part).

679B.R. Gudmundsdottir et al. / Thrombosis Research 130 (2012) 674–681

more of the four VKD coagulation factors. For this purpose only frozencitrated plasma could be obtained by us. However, despite the use ofplasma all three ROTEM coagulation phases were readily identified

0 1 2 3 40

1

2

3

4

R2 = 0.87y = 0.98x - 0.01

INR

Fiix

INR

Fig. 7. Correlation of Fiix-INR with INR in patients on stable warfarin therapy. For bothtypes of INR the laboratory calculated an instrument specific ISI based on an externallycalibrated INR standard.

and identical to the pattern observed in whole blood. In order to par-tially mimic clotting in whole blood we also repeated the ROTEM ex-periments by enriching PPP with platelet phospholipid or with addedwashed blood bank platelets. Since results were similar we considerthe experimental conditions to be adequate.

We are not aware of prior studies using ROTEM to evaluate the re-lationship of VKD coagulation factors on clot formation. Our experi-mental ROTEM results are based on highly diluted thromboplastinas opposed to the undiluted thromboplastin in the PT and this mayexplain how the ROTEM test differs from the PT. However, our resultsand conclusions are supported by others that have come to the simi-lar conclusions as we have using different in vitro methods and ani-mal models. Firstly, a study using anticoagulated plasma frompatients on warfarin demonstrated that a poor relationship existedbetween the INR and thrombin generation in response to dilute tissuefactor [19]. Also, Xi et.al. showed that prothrombinase activity inanticoagulated samples from patients on VKA was dependent onboth factors II and X but less on factor VII concentration which dem-onstrated a threshold relationship first evident at concentrationsbelow 5% [8]. Third, Zivelin et.al. found that the antithrombotic effectof VKD factor deficiency in rabbits depended on the activity of factorsII and X, but not on the activity of factors VII, and these authors con-cluded that monitoring of VKA therapy might be more accurate usingassays that principallymeasure factors II or X [7]. Fourth, it was shownin clinical studies, that monitoring VKA using the native prothrombinantigen by ELISA better predicted adverse events than monitoringwith the PT [10–12]. Taken together, all these studies, using differentmethods, suggest that monitoring that focuses on the activity of fac-tors II and X and ignoring the influence of factors VII and IX couldbe an improvement that might remove confounding fluctuations inPT assays caused by changes in factors VII.

VKA therapy even including monitoring cost is relatively cheapcompared to the new oral anticoagulant agents such as dabigatranand rivaroxaban. The coumarin effect is well studied and can imme-diately be reversed using prothrombin complex concentrates andmore slowly by administering plasma and vitamin K. On the otherhand the dose effect varies markedly due to genetically determinedmetabolic variations [26]. Also, the anticoagulant effect measured bytraditional INR fluctuates considerably in many patients, sometimesdue to known drug and food interactions but often for unknownreasons [27]. This leads to a need for frequent monitoring in orderto ensure efficacy and safety. It is hoped that new oral anticoagulantagents can be administered safely without monitoring [28–30]. Thisremains to be confirmed in practice, especially in the elderlywhich are over half of all anticoagulated patients [31]. Many of theelderly may have decreased renal function which may limit theuse of renally excreted anticoagulants such as dabigatran andrivaroxaban.

We conclude that at VKD factor activity relevant to therapeuticVKA therapy, factors II and X have a predominant role compared tofactors IX and VII. The results suggest further that rapid changes infactors VII activity during initiation and following dose changes ofVKA may exaggerate fluctuations in the INR that do not translateinto an antithrombotic affect, but lead to unnecessary dose changesand too frequent monitoring. This confounding effect of factor VIIcould be circumvented by monitoring VKA therapy with a prothrom-bin time variant such as the Fiix-PT, which is sensitive to reductionsonly in factors II and X and not factor VII as the Fiix-PT may more ac-curately reflect the anticoagulant effect of VKA than do the currentlyapplied INR tests.

Conflict of Interest Statement

Brynja R. Gudmundsdottir and Pall T. Onundarson have inventedthe Fiix-PT and have applied for a patent on the invention.

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 500

20

40

60

80

100

0

1

2

3

4

5

FIIFXFVII

INRFiix-INR

A

Days from warfarin initiation

Co

agu

lati

on

fac

tor

acti

vity

per

cen

tag

e

INR

value

14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 500

20

40

60

80

100

0

1

2

3

4

5

6

7

8

9

10INR

FIIFXFVII

Fiix-INR

B

Days of treatment

Co

agu

lati

on

fac

tor

acti

vity

per

cen

tag

e

INR

value

0 5 10 15 20 25 30 35 40 45 50 55 600

1

2

3

4

5

0

1

2

3

4

5

INRFiix-INR

C

Days of treatment

INR

val

ue

Fiix-IN

R valu

e

0 5 10 15 20 25 30 35 40 45 50 55 600

1

2

3

4

5

0

1

2

3

4

5

INRFiix-INR

D

Days of treatment

INR

val

ue

Fiix-IN

R valu

e

Fig. 8. Four examples of sequential INR and Fiix-INR measurements in patients dosed based on INR. In two patients (panels A and B) VKD factor coagulant activity is also shownillustrating the marked changes in INR and not Fiix-INR related to fluctuations in FVII.

680 B.R. Gudmundsdottir et al. / Thrombosis Research 130 (2012) 674–681

Acknowledgement

P.T. Onundarson and B.R. Gudmundsdottir designed the experi-ments which were conducted by B.R. Gudmundsdottir, A.M. Bjorns-dottir and M. Karlsson. The manuscript was written by P.T.Onundarson. C.W. Francis contributed to analysis and manuscript.

Disclosure:

Pall T. Onundarson and Brynja R. Gudmundsdottir have filed a patentapplication IS 050010 for the Fiix method.

References

[1] Roberts HR, Escobar MA. Other coagulation factor deficiencies. In: Loscalzo JSAI,editor. Thrombosis and Hemorrhage. 3rd edn. Philadelphia: Lippincott Williams& Wilkins; 2003. p. 575–98.

[2] van Dam-Mieras MC, Hemker HC. Half-life time and control frequency of vitaminK-dependent Coagulation factors. Theoretical considerations on the place of fac-tor VII in the control of oral anticoagulation therapy. Haemostasis 1983;13:201–8.

[3] Quick A. The prothrombin time in haemophilia and in obstructive jaundice. J BioChem 1935;109:73–4.

[4] Owren PA, Aas K. The control of dicumarol therapy and the quantitative determi-nation of prothrombin and proconvertin. Scand J Clin Lab Invest 1951;3:201–8.

[5] Haraldsson HM, Onundarson PT, Einarsdottir KA, Gudmundsdottir BR, PeturssonMK, Palsson K, et al. Performance of prothrombin-proconvertin time as a monitor-ing test of oral anticoagulation therapy. Am J Clin Pathol 1997;107:672–80.

[6] Le DT, Weibert RT, Sevilla BK, Donnelly KJ, Rapaport SI. The international normal-ized ratio (INR) for monitoring warfarin therapy: reliability and relation to othermonitoring methods. Ann Intern Med 1994;120:552–8.

[7] Zivelin A, Rao LV, Rapaport SI. Mechanism of the anticoagulant effect of warfarinas evaluated in rabbits by selective depression of individual procoagulant vitaminK-dependent clotting factors. J Clin Invest 1993;92:2131–40.

[8] Xi M, Beguin S, Hemker HC. The relative importance of the factors II, VII, IX and Xfor the prothrombinase activity in plasma of orally anticoagulated patients.Thromb Haemost 1989;62:788–91.

[9] Mariani G, Dolce A, Marchetti G, Bernardi F. Clinical picture and management ofcongenital factor VII deficiency. Haemophilia 2004;10:s180–3.

[10] Furie B, Liebman HA, Blanchard RA, Coleman MS, Kruger SF, Furie BC. Comparisonof the native prothrombin antigen and the prothrombin time for monitoring oralanticoagulant therapy. Blood 1984;64:445–51.

[11] Furie B, Diuguid CF, Jacobs M, Diuguid DL, Furie BC. Randomized prospective trialcomparing the native prothrombin antigen with the prothrombin time for moni-toring oral anticoagulant therapy. Blood 1990;75:344–9.

[12] Kornberg A, Francis CW, Pellegrini Jr VD, Gabriel KR, Marder VJ. Comparison of na-tive prothrombin antigen with the prothrombin time for monitoring oral antico-agulant prophylaxis. Circulation 1993;88:454–60.

[13] Sorensen B, Johansen P, Christiansen K, Woelke M, Ingerslev J. Whole blood coag-ulation thrombelastographic profiles employing minimal tissue factor activation. JThromb Haemost 2003;1:551–8.

[14] Ingerslev J, Poulsen LH, Sorensen B. Potential role of the dynamic properties ofwhole blood coagulation in assessment of dosage requirements in haemophilia.Haemophilia 2003;9:348–52.

[15] Fenger-Eriksen C, Anker-Moller E, Heslop J, Ingerslev J, Sorensen B. Thrombelasto-graphic whole blood clot formation after ex vivo addition of plasma substitutes:improvements of the induced coagulopathy with fibrinogen concentrate. Br JAnaesth 2005;94:324–9.

[16] Loeliger EA. ICSH/ICTH recommendations for reporting prothrombin time in oralanticoagulant control. Acta Haematol 1984;72:405–7.

681B.R. Gudmundsdottir et al. / Thrombosis Research 130 (2012) 674–681

[17] Loeliger EA, Poller L, Samama M, Thomson JM, Van den Besselaar AM, Vermylen J,et al. Questions and answers on prothrombin time standardisation in oral antico-agulant control. Thromb Haemost 1985;54:515–7.

[18] Rosendaal FR, Cannegieter SC, van der Meer FJ, Briet E. A method to determine theoptimal intensity of oral anticoagulant therapy. Thromb Haemost 1993;69:236–9.

[19] Brummel KE, Paradis SG, Branda RF, Mann KG. Oral anticoagulation thresholds.Circulation 2001;104:2311–7.

[20] Luddington R, Scales C, Baglin T. Lupus anticoagulant testing with optical endpoint automation. Thromb Res 1999;96:197–203.

[21] Rand MD, Lock JB. van't Veer C, Gaffney DP, Mann KG. Blood clotting in minimallyaltered whole blood. Blood 1996;88:3432–45.

[22] Watson HG, Greaves M. Can we predict bleeding? Semin Thromb Hemost2008;34:97–103.

[23] Ganter MT, Schmuck S, Hamiel CR, Wischmeyer PE, Heule D, Zollinger A, et al.Monitoring recombinant factor VIIa treatment: efficacy depends on high levelsof fibrinogen in a model of severe dilutional coagulopathy. J Cardiothorac VascAnesth 2008;22:675–80.

[24] Rugeri L, Levrat A, David JS, Delecroix E, Floccard B, Gros A, et al. Diagnosis of earlycoagulation abnormalities in trauma patients by rotation thrombelastography. JThromb Haemost 2007;5:289–95.

[25] Doran CM, Woolley T, Midwinter MJ. Feasibility of using rotational thromboelas-tometry to assess coagulation status of combat casualties in a deployed setting. JTrauma 2010;69(Suppl. 1):S40–8.

[26] Ansell J, Hirsh J, Hylek E, Jacobson A, Crowther M, Palareti G. Pharmacology andmanagement of the vitamin K antagonists: American College of Chest PhysiciansEvidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008;133:160S–98S.

[27] Witt DM, Delate T, Clark NP, Martell C, Tran T, Crowther MA, et al. Twelve-monthoutcomes and predictors of very stable INR control in prevalent warfarin users. JThromb Haemost 2010;8:744–9.

[28] Bauersachs R, Berkowitz SD, Brenner B, Buller HR, Decousus H, Gallus AS, et al.Oral rivaroxaban for symptomatic venous thromboembolism. N Engl J Med2010;363:2499–510.

[29] Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al. Dabi-gatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009;361:1139–51.

[30] Diener HC, Connolly SJ, Ezekowitz MD, Wallentin L, Reilly PA, Yang S, et al. Dabi-gatran compared with warfarin in patients with atrial fibrillation and previoustransient ischaemic attack or stroke: a subgroup analysis of the RE-LY trial. LancetNeurol 2010;9:1157–63.

[31] Onundarson PT, Einarsdottir KA, Gudmundsdottir BR. Warfarin anticoagulationintensity in specialist-based and in computer-assisted dosing practice. Int J LabHematol 2008;30:382–9.