ROTEM and vitro reversal of warfarin with APCC

Background: Warfarin-treated patients with a prolonged Prothrombin Time (PT) can have a normal rotational thromboelastometry (ROTEM) clotting time (CT). A previous in vitro study found that activated prothrombin complex concentrates (APCC) could reverse an albumin-induced coagulopathy monitored with ROTEM, but that prothrombin complex concentrates (PCC) could not. The aim of this study was to investigate the ability of ROTEM to monitor the in vitro reversal of warfarin-induced coagulopathy using APCC and to de ﬁ ne an APCC dose response. Method: During the routine control of PT in 27 patients treated with warfarin, one extra 4.5 ml test tube of citrated whole blood was retrieved. Two concentrations of tissue factor ROTEM tissue factor (TF) activating reagents were used: a standard ROTEM ExTEM and a diluted 1:19000 concentration. The effects of two separate doses of APCC added in vitro corresponding to in vivo doses of 50 IE or 100 IE/kg were then studied on the ROTEM. Results: The ROTEM EXTEM CT was prolonged beyond the upper normal range of 68 s in patients with PT >3.0, with a correlation coe ﬃ cient of 0.88 to PT. ROTEM with the ExTEM reagent alongside high and low doses of APCC resulted in a signi ﬁ cant shortening of median CT, both compared to baseline (100 s) and after low (65 s) and high doses (57 s). This was most evident in patients with PT >2.0. ROTEM signals of clot propagation to clot formation time (CFT) and α angle had the reverse pattern. There was no effect on maximal clot strength with APCC. With the diluted TF no CT shortening was found with APCC. Conclusion: A clear dose response of APCC added in vitro to correct the effects of warfarin on ROTEM EXTEM CT was veri ﬁ ed. ROTEM CT should be tested with non-activated PCC for in vivo reversal of warfarin in patients along with veri ﬁ cation of a normalised PT. Further studies are needed to verify if a ROTEM CT in the lower normal range of <57-65 s, as found in our in vitro APCC-spiked warfarin blood, is safe for invasive procedures. The aim of this study is to investigate the ability of ROTEM in monitoring in vitro reversal of a warfarin-induced coagulopathy using APCC. A secondary aim is to examine the effect of different dosages of APCC added in vitro to warfarin-treated blood. A hypothesis was formulated that the reversal of warfarin-induced coagulopathies could be monitored using ROTEM


Introduction
Warfarin is a commonly prescribed anticoagulant which inhibits the enzyme vitamin K reductase and thus prevents the -carboxylation of the vitamin K dependent procoagulant factors FII, FV, FIX and FX as well as the anticoagulant proteins C, S and Z [1]. This suppresses the binding of calcium to these factors, which is necessary for their activation. Patients on warfarin need to be monitored regularly to guide the anticoagulative treatment. This is routinely done with Owren PT-INR (Prothrombin Time-International Normalized Ratio), which represents the formation of fi brin from fi brinogen, catalysed by thrombin [2]. By adding tissue factor (TF) to citrated plasma, the PT-INR measures the extrinsic and common pathways of coagulation. Together with FVII, TF activates the FX/FV complex, which drives the cleaving of prothrombin (FII) to thrombin [3]. The prothrombin time is standardized internationally to eliminate inter-laboratory differences. Owren PT-INR is the standard method in Scandinavia, the Netherlands and Japan, while Quick PT-INR, which also takes concentrations of fi brinogen and FV into account, is used in the rest of the world [2].
Newer methods are increasingly employed alongside PT-INR in surgical and emergency settings to guide coagulative treatment and transfusions. One such method is rotational thromboelastometry (ROTEM), which is a Viscoelastic Haemostatic Assay (VHA). ROTEM measures the resistance against a rotating pin during forced oscillation and provides information regarding clot formation, clot strength and fi brinolysis in whole blood [4][5][6][7]. The ROTEM can give an overview of the different steps in coagulation with the initiation, amplifi cation and propagation phases, but also clot structure, platelet function and fi brinolysis [8]. The corresponding whole blood POC PT devices in the market only refl ect clot initiation, amplifi cation and propagation phases and also give rapid test results as compared to standard laboratory plasma based coagulation analyses such as he PT tests.
The ability of ROTEM to detect warfarin-induced coagulopathies has been proposed, but is an area in need of further study [7,8]. Treatment of bleeding in warfarin-treated patients often requires a reversal of the anticoagulative effect. This is commonly done with vitamin K or, in critical situations, with PCC (Prothrombin Complex Concentrate), which contains high concentrations of factors II, VII, IX and X. Using VHA to guide this reversal has not been thoroughly evaluated but the possibility has been suggested [9]. APCC (Activated Prothrombin Complex Concentrate) such as FEIBA (Factor Eight Inhibitor Bypassing Activity) is commonly used to control bleeding in haemophilic patients with inhibitors to factor eight (FVIII), but has also been proposed as an alternative anticoagulation reversal agent [10,11]. Non-activated PCC contains anticoagulants like protein C/S and heparin (to counteract the procoagulant factors, especially factor II/IIa [prothrombin/thrombin]) [12] and leads to ROTEM artefacts even if protamine or heparinase is used to block the heparin.
APCC not containing any anticoagulants is therefore better for in vitro dose response studies than PCC [12].
Previous studies have showed normal ROTEM graphs/ parameters even at high PT-INRs [7,8] The aim of this study is to investigate the ability of ROTEM in monitoring in vitro reversal of a warfarin-induced coagulopathy using APCC. A secondary aim is to examine the effect of different dosages of APCC added in vitro to warfarintreated blood. A hypothesis was formulated that the reversal of warfarin-induced coagulopathies could be monitored using

Ethics
This study was approved by the Regional Ethical Review  can decrease the variation coeffi cients of ROTEM parameters [13,14]. Citrated samples are stable for up to eight hours after sampling [15].

Prothrombin time
Owren PT-INR (international normalized ratio) was performed using a combined thromboplastin reagent (Owren PT, Medirox, Nyköping, Sweden). Plasma analyses were performed on a Sysmex CA-5100 automated coagulation analyser (Siemens Healthcare AB, Upplands Väsby, Sweden). The PT-INR assay was calibrated using reference plasma samples with certifi ed INR from Equalis (Uppsala, Sweden), and the normal reference range for this PT-INR was set at 0.9-1.2, with a coeffi cient of variation of <5%.

Viscoelastic haemostatic assays
ROTEM analysis (TEM International GmbH, Munich, Germany) was performed according to the manufacturer's instructions with ExTEM (tissue factor extrinsic activation). Blood was added to a cuvette and placed in a pre-heated container (37°C) while a rotating pin immersed in the blood created a forced oscillation. Increased resistance as the blood started to coagulate was detected and used to determine several variables. This study focused on clotting time (CT), clot formation time (CFT), the alpha angle () and maximum clot fi rmness (MCF). The reference intervals for the EXTEM protocol were set according to the manufacturer and Lang et al: CT 38-79 s, CFT 34-159 s,  angle 63-83 o and MCF 50-72 mm [15]. ROTEM EXTEM coeffi cients of variation have been reported to be 3-12% for CT, 3-12% for CFT, 1-5% for  angle and 1-5% for MCF [4,16].
This study used the standard commercial ExTEM reagent, which is a high dose TF reagent (exact concentration not given by the manufacturer). The use of a diluted TF reagent to improve the detection of coagulation defects has previously been suggested [12]. Therefore, a commercial TF reagent (Innovin, Siemens Healthcare AB) was diluted to 1:19000, in line with recommendations from the Coagulation Laboratory, Region Skåne, Malmö, Sweden. Four mL of distilled water was added to the Innovin, then the solution was further diluted by titrating 10 μL of the solution with 112 mL of 0.9% NaCl. This diluted TF (DTF) reagent was then used as a substitute for the standard EXTEM reagent provided by TEM International.

APCC in vitro titration
In this experiment, two different doses of APCC (FEIBA 50 IU/mL, Baxter, Vienna, Austria) were used on 27 samples of warfarin-treated blood. These concentrations are equivalent to clinical doses of 50 IU/kg and 100 IU/kg in a male weighing 70 kg with a blood volume of 5000 mL [14]. Each blood sample was analysed in vitro with a high dose of APCC (1.4 IU/mL), a low dose of APCC (0.7 IU/mL) and a baseline control using the standard ExTEM reagent. The fi rst 10 patients were also analysed with the DTF reagent. This resulted in six ROTEM channels per sample in the fi rst 10 samples and 3 ROTEM channels per sample in the following 17.

Statistical analysis
Statistical analysis was performed using the Graph Pad Prism software. Non-parametric tests were chosen, as all data was not normally distributed. P-values were calculated using the Wilcoxon signed-rank test. Linear and non-linear correlations between PT-INR and EXTEM CT in warfarin blood treated with the standard EXTEM reagent were calculated with Spearman's rank correlation. To minimize the risk of type 1 errors, statistical signifi cance was set to p<0.01.

Study population
Twenty-seven patients were included in the study. Three patients were excluded due to insuffi cient test tube sampling volume. The median age was 69 years (range 46-99 years).

Effects of APCC on warfarin-treated blood
ROTEM analysis with the standard ExTEM reagent alongside high and low doses of APCC resulted in a signifi cant shortening of CT, both compared to the baseline and between low and high doses (p<0.0001, Figure 1). Additionally, a high dose of APCC signifi cantly prolonged CFT and reduced the  angle (p<0.01) compared to the baseline. No other signifi cant results were found using the standard ExTEM reagent (p>0.01). The DTF reagent yielded no signifi cant results (p>0.01, Figure 2).
The difference between high and low dose APCC was limited to samples with higher control CT values when separating the patients after PT-INR into groups of ranges 2-2.9 and ≥ 3 ( Figure 3).

EXTEM CT correlation to Owren PT-INR
Statistical analysis with Spearman's rank correlation test showed a signifi cant linear correlation between elevated PT-INR and increased EXTEM CT (rho 0.88; p <0.0001, Figure 4).

Discussion
In this study, tissue factor-activated ROTEM monitoring of in vitro reversal of warfarin-induced coagulopathies with APCC was studied. We hypothesized that ROTEM could monitor this reversal and detect a dose response effect with both low and high doses of APCC. Only with the standard high-concentration tissue factor ExTEM reagent was the time until initial clot formation (CT) signifi cantly reduced by in vitro administration of APCC to warfarin blood. There was an additional effect of high-dose APCC compared to low-dose APCC, further signifi cantly shortening CT. The reduction in CT was greater with higher control CTs linked to PT-INRs > 2. Therefore, the hypotheses are supported.

ROTEM EXTEM and PT
Schmidt et al. found a strong correlation (rho=0.87) between ROTEM EXTEM CT and PT-INR in a recent study on warfarin-treated patients [7]. Nilsson el al. only found a moderate 0.66 correlation between EXTEM CT and PT-INR, but a better 0.76 between EXTEM CT and Quick-PT [8].

INTEM and FIBTEM CT had lower correlations in both studies.
Our study demonstrated a similar high correlation as in the Schmidt study between PT-INR and ROTEM EXTEM CT, as seen in Figure 4 (rho=0.88). This is likely because both tests were activated by high concentration tissue factor reagents [7], stimulating the extrinsic pathway. This correlation indicates that a reduction in EXTEM CT by administering APCC may equate to a corresponding reduction in PT-INR, but this subject needs to be studied further (see below).  Schmidt et al. so also described a signifi cant but small increase in ROTEM EXTEM MCF in warfarin-treated patients compared to healthy controls; there were no differences in ROTEM EXTEM  angles. ROTEM EXTEM CFT,  angle and MCF ranges in our study are similar to the ranges described by Schmidt et al [7]. We did not measure fi brinogen plasma levels. Increased fi brinogen measured with the Clauss method has been found to correlate positively with ROTEM MCF levels (rho=0.70) [17]. Many patients on warfarin have atherosclerosis, which can be linked to an infl ammatory process with higher fi brinogen plasma levels [18]. Haematocrit and platelet count and function can also affect MCF [19]. In a recent study on stent implantation in superfi cial femoral arteries, only a short ROTEM CT predicted late in-stent restenosis, whereas calibrated automated thrombography, activated partial thromboplastin time, platelet aggregation, platelet adhesion, fi brinogen, and microparticles' procoagulant activity failed [20].  [21]. PT-INR has a high sensitivity to changes in FVII plasma levels and activity and FVII has a shorter half-life than the other vitamin K dependent coagulation factors, such as FII and FX. ROTEM did not show this correlation with FVII but instead correlated with FX and less so with FII, indicating that these factors play a more predominant role than factor VII on global coagulation as seen in vitro using ROTEM [21].
The use of ROTEM for analysing warfarin coagulopathies has the benefi t of being a point-of-care method and can be performed using whole blood. The standard plasma-based PT-INR is not a point-of-care method. However, there are several point-of-care whole blood PT methods [22], and some of these can monitor both high and low PTs in warfarin-treated patients [23].

Diluted Innovin
It has been suggested that the standard commercial ROTEM EXTEM reagent causes too strong activation of the extrinsic pathway due to high concentrations of TF, resulting in the possibility that mild coagulopathies are overridden by the massive activation [12,21]. In contrast to this, our study produced no signifi cant results with the heavily diluted (1:19000) TF reagent (Innovin). As seen in Figure 2, the variation of the different ROTEM parameters with the DTF reagent was considerably higher than those of the corresponding standard EXTEM reagent, especially before adding APCC. The CTs obtained with the DTF reagent were markedly higher than their standard EXTEM reagent CT counterparts; however, no major differences were seen in CFT,  or MCF. Comparing our DTF CT values to a previous study, Elvstam et al. [12], using the NATEM protocol (no TF, only recalcifi cation of citrated blood), our CT was lower while still within NATEM protocol range (300-1000s), which indicates that the dilution can be compared to studies using no TF [12]. Additionally, with the DTF reagent, CT values after adding APCC were higher than those of the previously mentioned study on albumin-diluted blood with the same doses of APCC [12]. Based on our fi ndings and previous results, we argue that the recommended 1:19000 TF dilution is too extensive, while the standard EXTEM TF reagent might have a too high concentration. This could produce unreliable results in both instances. Therefore, a TF concentration somewhere between the standard ExTEM reagent and the diluted Innovin used in this study should be considered.   correlations between ROTEM CT/the alternative system thromboelastography (TEG) reaction time (TEG R) are low [24,25]. Current guidelines do not address whether a low ROTEM CT combined with a slightly prolonged PT-INR >1.2-1.5 or higher (still below 2.0) allows the anaesthesiologist to perform a regional anaesthesia in patients with warfarin medication.

APCC for in vitro reversal of warfarin
A previous study has found that APCC can correct (and in the case of CT, overcorrect) an albumin-induced coagulopathy, but PCC cannot [12]. This failure of PCC is probably an effect of the heparin content or protein C/S added to 4-factor PCC to counteract its strong pro-coagulative effects [12,26]. APCC has no heparin; also, its content of activated VII (fVIIa) can exaggerate the reversal effect of APCC in vitro and in vivo. Still, we did not fi nd ROTEM CT in the very low range (around 38 s) [12,16] after APCC correction of the warfarin blood. Perhaps an even higher APCC in vitro dose would have accomplished this.
APCC has previously been shown to revert the effects of rivaroxaban in vitro [27] and experimentally in pigs [28]. APCC also reversed the effects of the in vitro albumin-induced coagulopathy [12], on clot amplifi cation and clot propagation parameters CFT,  and MCF. With the warfarin-induced coagulopathy, we did not detect any effect of APCC on these parameters with the ExTEM reagent. However, with the DTF reagent the boxplot confi dence intervals markedly decreased ( Figure 2).
Counteractively, our results showed that CFT was prolonged and the  angle was decreased with the high-dose APCC as compared to the baseline EXTEM CFT. This implies impaired clot propagation. Similar results were recently reported by Peng et al. when in vitro-administered TF caused a reduction in CT and an increase in CFT [29]. The authors speculated that a decrease in clot formation kinetics may be the result of excessive activation of the extrinsic tenase pathway (FVII-TF). This mechanism could also explain the decrease in  angle by the high-dose APCC in this study.
APCC should not be used to reverse warfarin in patients. It is an antibody-bypassing agent in haemophilia patients used to treat bleeding. TEG has been used to titrate the optimal dose of APCC as an antibody-bypassing agent in haemophilia patients [30], but also used for monitoring the reversal of NOAC [31], with the TEG parameter reaction time (R) corresponding to ROTEM CT. No comparison of ROTEM and TEG with TF activating reagents has been performed, but monitoring warfarin reversal with PCC has been suggested in an in vitro study using ROTEM EXTEM CT [9] and in patients with intracranial bleeding using TEG R and TEG MA (maximal amplitude corresponding to ROTEM MCF) [32]. Schmidt el al found ROTEM EXTEM CT to correlate more than TEG R with PT-INR in warfarin-treated patients [7].

Limitations
This study was limited by the absence of healthy controls.
In addition, we did not analyse variables other than PT-INR at baseline, so variations in haematocrit, fi brinogen and platelet count, which may interfere with coagulation and ROTEM, are unaccounted for. Also the study period was limited, including more patients at the different PT-INR levels defi ned in Figure   3 would have strenghtened our results. ROTEM has high coeffi cients of variation (CV) for the different test parameters [8] -so repetetive testing of the same sample before and after reversal would detect a change of CVs. In an ongoing study we have found lower CVs for ROTEM test parameters in the lower normal range when testing at 8 simultaneous ROTEM channels.
There are also indications that with the fully automated ROTEM-Sigma®, CVs are lower than with the previous ROTEM models using digital pipetting of activating reagents as in our study.

Conclusion
This study demonstrates that a tissue factor-activated ROTEM analysis can be used to monitor the reversal of warfarin-induced coagulopathies with APCC in vitro. However, this effect was limited to patients with PT-INR >2 when looking at individuals. Further studies are needed to study in vivo warfarin reversal monitored with ROTEM and to fi nd the optimal individual PCC and/or vitamin K in vivo dosages for elective and emergency patients.