Keywords anticoagulants - blood coagulation - spatial clot growth rate - spatial thrombin distribution
- Thrombodynamics-4D
Introduction
The formation of a hemostatic plug at the site of vessel injury is a result of complex
interactions among the vessel wall, blood cells, and plasma coagulation proteins.[1 ] Typically, under physiological conditions, the initiation phase starts with the
contact of blood with tissue factor (TF) on the cell surface at the site of endothelium
damage, where thrombus growth begins. When a thrombus covers the site of endothelial
damage (surface with TF), further work of the coagulation cascade continues on the
surface of activated platelets and vesicles carrying phosphatidylserine.[2 ]
[3 ]
[4 ]
[5 ]
[6 ] At this stage, coagulation does not depend on the initial TF signal but depends
mainly on the reaction of the intrinsic pathway (activation of factor XI by thrombin).[6 ]
[7 ]
[8 ]
[9 ] In some pathologies, the risk of thrombosis increases. For the treatment and prevention
of thrombotic complications, coagulation inhibitors (anticoagulants) are used. The
molecular mechanisms of anticoagulant action are well-characterized. The anticoagulants
studied in this paper act either through an antithrombin (AT)-dependent mechanism
(heparin), increasing to varying degrees the inhibition of several coagulation factors
by AT or directly inhibiting one of the coagulation factors (thrombin or factor Xa
[FXa]). It is assumed that the inhibition of individual procoagulant factors can have
a more predictable effect than the inhibition of all serine proteases of the cascade
by AT or antagonists of vitamin K.[10 ]
[11 ] Thrombin and FXa are necessary for clot formation at any of activation mechanisms
(intrinsic or extrinsic). Thrombin is also important for the positive feedback loops
of the cascade, which is responsible for FV, FVIII, and FXI activation, and further
clot growth.[12 ]
[13 ]
[14 ] Thus, direct inhibition of these factors decreases their activity and can affect
the activation and spread of coagulation.
Previously, to study coagulation, we developed the Thrombodynamics method, in which
the coagulation activator is TF, localized on a solid surface and simulating a damaged
vessel wall. This system is principally different from homogeneous systems with complete
mixing of the activator and blood.[15 ]
[16 ] The idea of this experimental approach is the physical separation of the phases
of coagulation initiation and clot growth, as occurs in living organisms. The addition
of a thrombin-specific fluorogenic substrate to this system (Thrombodynamics-4D [TD-4D])
allowed us to monitor both fibrin clot formation and the generation and propagation
of active thrombin in space simultaneously. The propagation of thrombin formation
in this system is independent of TF and is determined by the positive feedback loop
of the intrinsic pathway.[8 ]
The thrombin generation test (TGT) is now widely used for the analysis of coagulation
in various situations.[17 ]
[18 ]
[19 ] However, this test is performed in a homogeneous system with a fully mixed activator,
which makes it less physiological. After the plasma is mixed with TF, the thrombin
activity in the sample begins to increase rapidly. Then, a phase of inhibition follows.
As a result, a peak appears on the calculated kinetic curve of active thrombin. In
this system, fibrin formation occurs when only a small amount of thrombin is produced,[20 ] and most thrombin is generated after clot formation. However, within the framework
of this approach, there is no answer to the question of why this excessive amount
of thrombin is needed for blood clotting.
The spatial distribution of thrombin in normal plasma during clot growth was previously
described.[8 ] Initially, thrombin is produced on the activating surface in direct contact with
TF. Fibrin formation begins approximately 1 minute after activation ([Fig. 1A ]). A fundamental feature of the blood coagulation system is that the further spread
of thrombin occurs not through the diffusion of active thrombin from the site of activation
but rather through the subsequent activation of the entire cascade at each point where
a small amount of thrombin enters. This leads to self-supporting clot growth.[1 ]
[8 ]
[12 ]
[14 ]
Fig. 1 Propagation of thrombin and fibrin generation in space. (A ) Images of growing fibrin clots at sequential time points (typical experiment). The
clot size (in mm) is presented. The vertical scale bar on the right is 4 mm. (B ) Peak thrombin generation (black line) and fibrin formation (blue line) at successive
points reached by thrombin (only fibrin caused by thrombin formed at that particular
time point is represented). C
max and T
max correspond to the maximum thrombin peak near the activator surface and the time to
reach this concentration, respectively. (C ) Two-dimensional representation of thrombin and fibrin propagation in space. A
st is the amplitude of the propagating thrombin peak 60 minutes after the activation
of coagulation.
At each time point, a new thrombin peak is initiated. Thus, sequentially generated
peaks of thrombin form a moving wave of thrombin, followed by fibrin formation ([Fig. 1B, C ]). At each next time point, the thrombin activity curve has the same peak shape but
with an increasing delay time. Fibrin formation at each point occurs at the very early
stages of thrombin generation, which is in accordance with previously published data.
The height of the thrombin peak decreases rapidly with increasing distance from an
activator, as the influence of TF weakens. However, even if the TF is covered with
1 to 2 mm of fibrin, the clot continues growing. In this system, steady-state propagation
of thrombin can be observed,[8 ] which means that thrombin supports its own propagation.[8 ]
[21 ]
[22 ]
[23 ]
[24 ] The goal of this work was to study the detailed effects of different types of anticoagulants
on the spatial distribution of thrombin in the phases of coagulation activation and
its further propagation, as well as the resulting effects of these anticoagulants
on fibrin clot growth. We examined direct reversible inhibitors of FXa (rivaroxaban)
and thrombin (dabigatran or dabigatran etexilate in ex vivo experiments), as well
as low-molecular-weight heparin (LMWH; nadroparin) and unfractionated heparin (UFH;
only in vitro), which irreversibly inhibit several coagulation factors, including
FXa, FIXa, FXIa, and thrombin, in the presence of AT. Reversible coagulation inhibitors
at therapeutic concentrations act on the activation phase of coagulation, increasing
the time to thrombin appearance and the onset of fibrin clot formation, but cannot
completely stop further clot growth. On the other hand, irreversible anticoagulants
(nadroparin and UFH) affect the rates of clot growth and thrombin propagation in space
but do not significantly affect the time of onset of thrombin and fibrin formation.
Materials and Methods
Donors and Patients
The effects of various anticoagulants on plasma coagulation were studied both in vitro
and ex vivo. For in vitro experiments, blood from healthy donors (N = 36) was used (16 males and 20 females aged 22–62 years). Donors did not receive
any medication for 2 weeks prior to the study. Coagulation in vitro was studied in
the presence of different concentrations of LMWH nadroparin, UFH, dabigatran, or rivaroxaban.
The ex vivo part of the study included 75 adult patients who underwent elective replacement
of the hip or knee joints between February 2015 and June 2016 at the N.N. Priorov
National Medical Research Center for Traumatology and Orthopedics, Moscow, Russia
(25 males, 50 females, 24–80 years old, weighing 45–127 kg). After surgery, all patients
were randomly divided into three groups and received various anticoagulant prophylactic
therapies. Thirty-three patients received rivaroxaban, a direct inhibitor of FXa-10 mg
once a day;[25 ] 27 patients received dabigatran etexilate, which turns into a direct thrombin inhibitor
(DTI) dabigatran—220 mg once a day;[26 ] and 15 patients received LMWH nadroparin twice a day at a dosage of 2,850 or 5,700
anti-Xa IU in the morning and 2,850 anti-Xa IU in the evening (depending on the patient's
weight), which corresponds to a high prophylactic dose of nadroparin.[27 ] The male-female ratio, mean age, and mean body weight did not significantly differ
among the three groups (one-way two-tailed analysis of variance [ANOVA], p < 0.05; [Supplementary Table S2 ] [available in the online version]).
Maximum plasma concentrations of all anticoagulants studied (Amax) after administration
of the used prophylactic or maximum therapeutic doses of these drugs were not measured
but were calculated based on literature data[28 ]
[29 ]
[30 ]
[31 ]
[32 ]
[33 ] (see [Supplementary Material, sections 1.2 ] and [1.3 ] [available in the online version]). In particular, the concentration of nadroparin
was calculated based on data from a study,[28 ] which showed that when it was transfused at a dose of 43.1 anti-Xa IU/kg, the Amax
was 0.252 IU/kg. In the case of our patients, Amax was 95 anti-Xa IU/kg, because they
obtained the nadroparin dose 5,700 anti-Xa IU at the weight of 60 kg but less than
90 kg, or 8,550 anti-Xa IU at ≥90 kg. Given that for LMWH, the dose of the drug is
directly proportional to its maximum plasma concentration, we deduced that a dose
of 95 IU/kg should result in Amax = 0.58 IU/mL (∼0.6 IU/mL).
Thrombodynamics-4D Test
According to reference,[22 ] simultaneous measurements of light scattering (at λ = 625 nm) and fluorescence of
7-amine-4-methylcoumarin (AMC), which is a product of thrombin-specific fluorogenic
substrate hydrolysis by thrombin (at λexcitation = 365 nm; λemission = 440 nm) were performed using a Thrombodynamics® Analyzer T2T for 90 minutes at
37 °C. Coagulation activators (plates coated at one end with immobilized TF) were
prepared via a previously described method.[34 ] Measurements were performed in platelet-free plasma ([Supplementary Material ], section 2.2 [available in the online version only]). All rules for the preanalytical
preparation of plasma samples were standard, as for all coagulation assays.[35 ] The results were recorded via a charge-coupled device (camera) every 6 seconds and
processed using an automated calculation algorithm via software specially developed
by the manufacturers. Images from red and UV light were processed similarly. To obtain
light scattering or AMC fluorescence intensity profiles, the light intensity along
the line perpendicular to the activator surface was measured for each frame. The principle
of calculating active thrombin concentrations was similar to that used in the TGT
but took into account the diffusion of AMC in space.[8 ]
[22 ]
[36 ]
Briefly, activation of coagulation results in the formation of thrombin, which hydrolyzes
the thrombin-specific fluorogenic substrate present in plasma. The setup records the
emerging fluorescence of the product of this hydrolysis (AMC) over the entire area
of the sample every 6 seconds. The result is a set of AMC profiles, each of which
characterizes the distribution of AMC in space at the corresponding moment in time.
The thrombin activity at each moment in time at each point in space is proportional
to the rate of AMC accumulation. This rate can be obtained by differentiating the
recorded AMC profiles. It should also be taken into account that the concentration
of AMC at any given moment in time consists not only of AMC formed as a result of
the substrate hydrolysis reaction by thrombin but also of AMC that has ended up at
this point due to its diffusion. Thus, to calculate the activity of thrombin, it is
necessary to use the reaction–diffusion equation [(1) ]:
where D
AMC is the diffusion coefficient of AMC, [S] and [IIa] are the concentrations of the
substrate and thrombin, respectively, and k
cat and K
M are the catalytic constant and Michaelis constant for the reaction of the substrate
hydrolysis by thrombin, which obeys the Michaelis–Menten equation (these constants
are known). From this equation, by solving the inverse problem, one can calculate
the concentration of active thrombin at each moment in time at each point in space
using equation [(2) ]:
It should be noted that the final calculation also takes into account the fact that
67% of the resulting AMC binds to plasma albumin, resulting in a decrease in its diffusion
rate and that the presence of fibrin also affects the fluorescence of AMC. All this
was shown in separate experiments.[8 ] We do not provide, here, a detailed description of the calculation of the thrombin
concentration from the measured AMС fluorescence (with all necessary corrections)
due to space limitations, and because this was described in great detail in the supplementary
material of the work by Dashkevich et al.,[8 ] as well as in the works.[19 ]
[22 ]
[36 ]
The following TD-4D parameters were measured.
T
lag is the time between coagulation activation upon contact with TF and the beginning
of fibrin formation on the activating surface;
V
i and V
st are the initial and steady-state clot growth rates measured at intervals of 2 to
6 minutes and 15 to 25 minutes after T
lag , respectively;
C
max and T
max are the height of the thrombin peak near the activator and the time to reach this
peak, respectively;
A
st is the height of the thrombin peak running in space 60 minutes after activation;
V
t and V
f are the rates of thrombin or fibrin propagation, respectively, in the interval of
45 to 55 minutes, which were shown to be equal ([Supplementary Fig. S1 ] [available in the online version]);
T
sp is the time at which spontaneous clots appear in the sample. The presence of such
clots indicates plasma hypercoagulation.[5 ]
For a detailed description of the reagents, sample preparation, method of TD-4D and
its parameters, as well as normal values, reproducibility, and interindividual variability
of these parameters, see [Supplementary Material, section 2 ] and [Supplementary Table S1 ] (available in the online version).
Ex vivo Study Design
The following exclusion criteria were used for the patient selection: (1) age less
than 18 years; (2) presence of hematological diseases; (3) taking anticoagulants within
2 weeks before hospitalization; (4) pregnancy; and (5) refusion to participate in
the study. No selected patients were withdrawn from the study during treatment.
The TD-4D test was performed at the following checkpoints within 8 to 9 days after
surgery.
For dabigatran and rivarixaban
Point 1: one hour before the operation (Day 1).
Points 2 and 3: the morning following the day after surgery, before and 2 to 3 hours
after the first dose of anticoagulant, respectively (Day 2).
Points 4 and 5: in the morning before and 2 to 3 hours after the second dose of anticoagulant,
respectively (Day 3).
Point 6: Day 7 (or 8), 2 to 3 hours after the scheduled anticoagulant intake.
Point 7: Day 8 (or 9), in the morning before the regular dose of anticoagulant.
For LMWH, which was administered two times a day subcutaneously, all samples were
taken on analogous days in the morning (before or 2–3 hours after administration of
the appropriate dose of anticoagulant).
A diagram of the anticoagulant administration times and the measurement points corresponding
to its maximum and minimum concentrations is shown in [Supplementary Fig. S3 ] (available in the online version).
On days 7 to 8 of hospitalization, the patients underwent an ultrasound of the lower
extremity veins to detect thrombotic complications in the postoperative period. The
data are presented in [Supplementary Table S2 ] (available in the online version).
Data Analysis
The normality of the distributions of all the parameters was examined via the D'Agostino-Pearson
test (program MedCalc, version 14.12; MedCalc Statistical Software bvba, Ostend, Belgium).
The in vitro inhibitory effect of different concentrations of each anticoagulant on
TD-4D parameters was calculated as a percentage relative to the values of these parameters
in the system without inhibitor and averaged overall relevant samples.
The TD-4D test parameters were also compared in several pairs of adjacent points with
the minimum and maximum effect of the drug (before and 3 hours after taking the anticoagulant).
For this, the paired Wilcoxon's t -test (Wilcoxon signed-rank test) was used together with the Holm correction for multiple
comparisons, which was performed using the “stats” package in R (v 4.4.0). Differences
were considered significant at p -values <0.05.
Results
Effect of Anticoagulants on Clot Growth and Thrombin Distribution In Vitro
In vitro experiments avoid changes in the concentrations of added anticoagulants in
the sample, which is possible in ex vivo experiments because of the distribution of
the drug in the body and the different rates of its elimination in different patients.
To study the effects of LMWH, UFH, dabigatran, and rivaroxaban on different phases
of coagulation, increasing concentrations of these anticoagulants were added to the
plasma of different donors. The effects of these concentrations on all the TD-4D parameters
were compared. [Figure 2 ] shows typical curves for some TD-4D parameters obtained for individual donors in
the presence of various concentrations of nadroparin, dabigatran, and rivaroxaban.
Fig. 2 Typical examples of the effects of various anticoagulants in vitro on the size of
the fibrin clot, the thrombin distribution near the activator (ThrombinA ), and its spatial distribution at 60 minutes after activation (ThrombinS ) in the plasma of individual donors. A-C: LMWH nadroparin, F-D: dabigatran, and G-I: rivaroxaban. Curves corresponding to different concentrations of anticoagulants are
presented (see panels A , D, and G ). LMWH, low-molecular-weight heparin.
The averaged results of all experiments in vitro, obtained for the main parameters
of the clot growth and spatial thrombin generation, are presented in [Figs. 3 ] and [4 ], and [Supplementary Fig. S2 ] (available in the online version).
Fig. 3 Effects of LMWH nadroparin (A, B ), dabigatran (C, D ), and rivaroxaban (E, F ) on fibrin formation parameters T
lag , V
i , V
st , and V
f . The mean values ± standard deviations are presented. N = 8 for each group. The shaded areas represent the plasma concentrations that can
be obtained for each anticoagulant during therapeutic or prophylactic treatment (see
[Supplementary Material, section 1.3 ], available in the online version). The horizontal dashed lines represent the normal
range for the T
lag value. LMWH, low-molecular-weight heparin.
Fig. 4 Effects of LMWH nadroparin (A, B ), dabigatran (C, D ), and rivaroxaban (E, F ) on thrombin generation parameters T
max , C
max , V
t , and A
st . The mean values ± standard deviations are presented. N = 8 for each group. The shaded areas represent the plasma concentrations that can
be obtained for each anticoagulant during therapeutic or prophylactic treatment (see
[Supplementary Material, section 1.3 ], available in the online version). The horizontal dashed lines represent the normal
range for the T
max value. LMWH, low-molecular-weight heparin.
Nadroparin
The TD-4D parameters in the presence of different concentrations of nadroparin are
presented in [Figs. 3A, B ] and [4A, B ]. Increasing the LMWH concentration to 2.4 anti-Xa IU/mL did not increase the delay
in the onset of clot formation (T
lag ; [Fig. 3A ]). In contrast, spatial clot growth slowed greatly. In the range of nadroparin concentrations
of 0.3 to 2.4 anti-Xa IU/mL, an almost complete cessation of clot growth was observed
for V
f and V
st ([Fig. 3B ]). The sensitivity of various parameters describing the spatial rate of clot growth
to nadroparin increased with increasing distance from the coagulation-activating surface
(V
i < V
st < V
f ). The clot growth rate 45 to 55 minutes after activation was most sensitive to LMWH.
It decreased by about 80% at the LMWH concentration of 0.15 anti-Xa IU/mL and then
remained almost unchanged. In contrast, the decrease in the initial clot growth rate
(V
i ) with increasing LMWH concentration was the slowest, but this rate continued to gradually
decrease to a heparin concentration of 2.4 anti-Xa IU/mL ([Fig. 3B ]). Fibrin clot formation was still detectable at the nadroparin concentration of
4.8 anti-Xa IU/mL as a thin layer on the activator colocalized with residual thrombin
activity ([Supplementary Fig. S4A, B ] [available in the online version]), but a further increase in the heparin concentration
led to complete inhibition of clot formation.
Nadroparin moderately affected the maximum value of the thrombin peak near the activator
surface (C
max ) but did not affect the time to reach this maximum peak (T
max ; [Fig. 4A, B ]). It also significantly inhibited other parameters of the thrombin propagation phase
(V
t and A
st ; [Fig. 4B ]).
The most sensitive parameter to LMWH was V
t . Its value decreased by approximately 85% at a nadroparin concentration of 0.15 anti-Xa
IU/mL, whereas A
st and C
max at the same LMWH concentration were reduced by approximately 61.6% and 28.8%, respectively.
Thus, the action of LMWH rapidly decreased the moving thrombin peak, simultaneously
with the inhibition of clot growth. At nadroparin concentrations above 0.15 anti-Xa
IU/mL, thrombin generation was still observed, but its propagation rate was significantly
reduced. The thrombin wave disappeared, and the thrombin distribution profile changed
to diffusional ([Fig. 2C ]).
Quite similar results were obtained in in vitro experiments with UFH ([Supplementary Fig. S2 ] [available in the online version]).
Dabigatran and Rivaroxaban
Compared with nadroparin, direct oral anticoagulants (DOACs; direct thrombin or FXa
inhibitors) had a significantly weaker effect on fibrin clot propagation ([Fig. 3D, F ]), although the onset of fibrin formation (T
lag ) slowed with increasing in their concentrations ([Fig. 3C, E ]). As in the case of LMWH, the rate V
f was the most sensitive to dabigatran or rivaroxaban concentration, but it decreased
only 20 to 30% from the initial value within the therapeutic range of their concentrations.
Increasing the concentration of both DOACs to 2.4 μM resulted in an additional reduction
of only 5 to 10%.
DOACs essentially affected the thrombin peak C
max at the activating surface ([Fig. 4D, F ]). The peak height decreased, and the time to peak (T
max ) was prolonged ([Fig. 4C, E ]), resulting in a delay in fibrin formation. The height of the moving thrombin peak
A
st decreased synchronously with C
max near the activator. An increase in the concentration of dabigatran above 2 μM led
to the transformation of the mobile thrombin peak into a plateau with a sharp decrease
in the thrombin concentration at its border ([Fig. 2F ]). Rivaroxaban also changed the shape of the thrombin peak, lengthening the thrombin
inhibition process, which led to a widening of the peak ([Fig. 2H, I ]). Unlike LMWH, with increasing rivaroxaban concentration, the thrombin peak transformed
into a plateau rather than into a diffusional distribution ([Fig. 2I ]). As with dabigatran, this plateau could still spread in space and control the formation
of a fibrin clot.
The differences between LMWH and the DOACs were also confirmed by in vitro experiments
where the anticoagulant concentrations were much higher than therapeutic ones. AT-dependent
nadroparin at a concentration of 4.8 anti-Xa IU/mL reduced thrombin formation and
spread ([Supplementary Fig. S4B ] [available in the online version]), as well as the spatial clot growth rate ([Supplementary Fig. S4A ] [available in the online version]) to almost zero. On the other hand, reversible
AT-independent anticoagulants dabigatran and rivaroxaban also reduced the spatial
clot growth but were significantly weaker than heparin. They suppressed thrombin production
more than clot growth ([Supplementary Fig. S4D, F ] [available in the online version]), especially in the case of dabigatran ([Supplementary Fig. S4C, D ] [available in the online version]). The clot continued to grow even at dabigatran
concentrations of 5, 9.6, and 20 μМ ([Supplementary Figs. S4C ] and [S5E, F ] [available in the online version]). A similar pattern was also observed in the presence
of a high (9.6 μM) concentration of rivaroxaban ([Supplementary Fig. S4E ] [available in the online version]).
These data indicate that strong specific inhibition of thrombin does not prevent the
formation and spread of fibrin in space completely even at very high DOAC concentrations.
Effects of Various Anticoagulants on Ex Vivo Coagulation
To confirm the in vitro data on the effect of various anticoagulants on the activation
and propagation phases of coagulation, plasma samples from patients after elective
knee or hip replacement who received anticoagulant prophylaxis in the postoperative
period were studied. The study lasted 8 to 9 days while the patients were in the clinic
and included 75 patients who were divided into three groups treated with LMWH nadroparin
(N = 15), dabigatran etexilate, dabigatran prodrug (N = 27), and rivaroxaban (N = 33). Prophylactic doses of each drug were used, which were approximately equal
in efficacy.[25 ]
[26 ]
[27 ] The study design, patients, and doses are described in the “Materials and methods”
section and [Supplementary Material, sections 1, 3, 4 ] (available in the online version).
The results obtained ex vivo for various anticoagulants confirmed the findings of
the in vitro experiments.
Nadroparin Prophylactic Therapy
All the results obtained for patients treated with nadroparin are presented in [Fig. 5 ]. To avoid overloading the figure, only the significance of differences between adjacent
points with maximum and minimum drug effects is presented. These are pairs of points:
3/2, 3/4, 5/4, and 6/7. The p -values for the significance of the differences between all studied pairs of points
for all anticoagulants are presented in [Supplementary Table S3 ] (available in the online version).
Fig. 5 Effect of nadroparin on TD-4D parameters at different time points before and after
surgery (see “Materials and methods” section, and [Supplementary Fig. S3 ] [available in the online version]). The values of the lag period before the fibrin
appearance (A ), the initial (B ) and stationary (C ) clot growth rates, the maximum thrombin concentration near the activator (D ), the rate of thrombin propagation in the time interval 45 to 55 minutes after activation
(E ), and the height of the thrombin peak 60 minutes after activation (F ) are presented. The shaded zone corresponds to the range of normal values for each
parameter. The arrows indicate the time of nadroparin administration. Box sizes correspond
to the range from 25th to 75th percentiles of all the measured values. The medians
are represented by horizontal lines, the length of the whiskers corresponds to the
range from the 5th to the 95th, and the “ × ” signs from 1st to 99th percentiles of
all the measured values. *The value is significantly different from this value at
the previous point (Wilcoxon signed-rank test with the Holm correction for multiple
comparisons, p < 0.05). TD-4D, Thrombodynamics-4D.
Nadroparin did not result in a significant change in T
lag ([Fig. 5A ]) but did result in a significant decrease in initial and especially stationary rates
of thrombus growth below normal in most patients at the peak of LMWH action (points
3, 5, and 6). At the end of the nadroparin dose action (12 hours after administration),
the values of these parameters returned to the hypercoagulation (point 4) and normal
coagulation (point 7) regions. Trends in the parameters characterizing thrombin generation
at different points of the study are clearly visible in [Fig. 5D–F ]. It turned out that the maximum concentration of thrombin near the activator (C
max ) did not differ significantly at different points of the study ([Fig. 5D ]). This also applies to A
st values, for which there was a decreasing trend at the maximum nadroparin action (points
3, 5 and 6). However, because of large variations, these changes were not reliable,
and the value of the parameter A
st for all points remained within the normal range ([Fig. 5F ]). V
t and A
st values, always decreased at the points of maximum drug action (points 3, 5, and 6),
but increased again by the end of the drug dose action (points 4 and 7). However,
the differences between these values were significant only after the second and subsequent
administrations, although the trend V
t after the first administration was the same ([Fig. 5E ]). Thus, the significance of the differences between the maximum and minimum points
of nadroparin dose action was more pronounced for the V
t parameter than for A
st .
The reduction in hypercoagulability after LMWH administration was also confirmed by
a decrease in the percentage of samples with spontaneous clots observed ([Supplementary Figs. S6 ] and [S7 ] [available in the online version]). The number of spontaneous clots in the samples
after nadroparin administration was reduced at the peak of drug action (points 3,
and 5), but increased again at the end of its action (point 4)
Thromboprophylaxis with Direct Oral Anticoagulants
The second and third groups of patients received DOACs. These were dabigatran etexilate,
a prodrug that quickly turns in the body into a direct reversible thrombin inhibitor
dabigatran (N = 27), and rivaroxaban, a reversible inhibitor of FXa (N = 33). Changes in TD-4D parameters at different time points in all patients receiving
these anticoagulants are shown in [Figs. 6 ] and [7 ], respectively. Unlike LMWH, these DOACs cause a significant increase in T
lag ([Figs. 6A ] and [7A ]) and a decrease in C
max ([Figs. 6D ] and [7D ]) after each dose of the drug (points 3 and 5). Although the amplitude of thrombin
near the activator changed significantly after the administration of dabigatran etexilate
or rivaroxaban, almost all C
max values remained within the normal range ([Figs. 6D ] and [7D ]).
Fig. 6 Effect of dabigatran etexilate on TD-4D parameters at different time points before
and after surgery (see “Materials and methods” section, and [Supplementary Fig. S3 ] [available in the online version]). The values of the lag period before the fibrin
appearance (A ), the initial (B ) and stationary (C ) clot growth rates, the maximum thrombin concentration near the activator (D ), the rate of thrombin propagation in the time interval 45 to 55 minutes after activation
(E ), and the height of the thrombin peak 60 minutes after activation (F ) are presented. The shaded zone corresponds to the range of normal values for each
parameter. The arrows indicate the time of drug administration. Box sizes correspond
to the range from 25th to 75th percentiles of all the measured values. The medians
are represented by horizontal lines, the length of the whiskers corresponds to the
range from the 5th to 95th, and the “ × ” signs from 1st to 99th percentiles of all
the measured values. * The value is significantly different from this value at the
previous point (Wilcoxon signed-rank test with the Holm correction for multiple comparisons,
p < 0.05). TD-4D, Thrombodynamics-4D.
Fig. 7 Effect of rivaroxaban on TD-4D parameters at different time points before and after
surgery (see “Materials and methods” section and [Supplementary Fig. S3 ] [available in the online version]). The values of the lag period before the fibrin
appearance (A ), the initial (B ) and stationary (C ) clot growth rates, the maximum thrombin concentration near the activator (D ), the rate of thrombin propagation in the time interval 45 to 55 minutes after activation
(E ), and the height of the thrombin peak 60 minutes after activation (F ) are presented. The shaded zone corresponds to the range of normal values for each
parameter. The arrows indicate the time of drug administration. Box sizes correspond
to the range from 25th to 75th percentiles of all the measured values. The medians
are represented by horizontal lines, the length of the whiskers corresponds to the
range from the 5th to 95th, and the “ × ” signs from 1st to 99th percentiles of all
the measured values. *The value is significantly different from this value at the
previous point (Wilcoxon signed-rank test with the Holm correction for multiple comparisons,
p < 0.05). TD-4D, Thrombodynamics-4D.
The clot growth rates (V
i and V
st ) also decreased after taking the drugs, but unlike rivaroxaban, where the decrease
was significant after each dose of the drug, dabigatran etexilate had the same general
trend, but significantly reduced V
i only after the first and second doses, and V
st only after the first dose of the drug.
Measures characterizing the distribution of thrombin (V
t and A
st ) showed a weak downward trend at the points of maximum effect of both drugs. However,
for dabigatran etexilate, this effect was unreliable, and in the case of rivaroxaban,
it was reliable only after the first dose of the drug for V
t , and after the first and second doses of the drug for A
st . Importantly, after a week of taking either of these two drugs, the difference between
the parameters at the peak and the end of the drug dose action was erased, that is,
almost all the parameters (except T
lag ) were within the normal range. Thus, after approximately 1 week, hemostasis from
postoperative hypercoagulation returned to normocoagulation in most patients. We call
hypercoagulation a condition in which any of the indicators characterizing the intensity
of blood clotting shows an increase in this intensity that goes beyond the normal
values.
Comparison of the Effects of Different Anticoagulants on Thrombodynamics-4D Parameters
In Vitro and Ex Vivo
The mean changes in various TD-4D parameters (in percentages) relative to conditions
without anticoagulants were calculated for in vitro and ex vivo experiments at fixed
anticoagulant concentrations ([Fig. 8 ]). To calculate parameters in in vitro experiments, concentrations close to those
obtained in ex vivo experiments were used (0.6 anti-Xa IU/ml, 0.4 μM and 0.3 μM for
LMWH, dabigatran, and rivaroxaban, respectively; [Fig. 8A ]). Calculations for ex vivo experiments were performed using the anticoagulant concentrations
obtained in this study during patient prophylaxis. For this, the values at point 5
(after the second or third dose of anticoagulant for DOACs or LMWH, respectively)
were calculated as a percentage of the values at the corresponding point 2 (after
operation, but before anticoagulant administration; [Fig. 8B ]). The V
f values are not shown in [Fig. 8 ], because, according to [Supplementary Fig. S1 ] (available in the online version), they are equal to the V
t values.
Fig. 8 Comparison of the effect of various anticoagulants on TD-4D parameters characterizing
fibrin distribution (T
lag , V
i , V
st ) and thrombin formation (T
max , C
max , V
t , and A
st ) in in vitro and ex vivo experiments. Effect values are presented as a percentage
of the corresponding values in the absence of anticoagulants. The mean values and
standard errors of the means are presented. (A ) Changes in TD-4D parameters in in vitro experiments. To compare in vitro and ex
vivo experiments, in vitro anticoagulant concentrations similar to those obtained
ex vivo for the studied prophylactic doses of anticoagulants were used (see calculation
in [Supplementary Material, section 1.2 ] [available in the online version]). They were as follows: LMWH = 0.6 anti-Xa IU/mL,
dabigatran = 0.4 μM, and rivaroxaban = 0.3 μM (N = 8 for each group). (B ) TD-4D parameters in ex vivo experiments at point 5 (on the third day of the experiment,
3 hours after the administration of the corresponding anticoagulant) in relation to
point 2 (after surgery but before the anticoagulant administration) in patients treated
with LMWH nadroparin (5,700–8,550 anti-Xa IU/day, N = 15), dabigatran etexilate (220 mg/day, N = 27), or rivaroxaban (10 mg/day, N = 33). LMWH, low-molecular-weight heparin; TD-4D, Thrombodynamics-4D.
For all anticoagulants in vitro, the degree of inhibition of clot growth in space
increased with increasing distance from the activating surface (V
i < V
st < V
f ). The levels of inhibition of all these rates were almost the same for both DOACs,
but higher for LMWH ([Fig. 3B, D, F ]). This was also true for similar parameters in ex vivo experiments (except for V
i , where inhibition was similar for all anticoagulants; [Fig. 8B ]). The parameters characterizing thrombin generation were fundamentally different
for LMWH and DOACs (AT-dependent and independent inhibitors of thrombin and FXa, respectively).
In the case of LMWH, inhibition was reduced in the following order: V
t > A
st > C
max ([Fig. 4B ]), whereas for dabigatran and rivaroxaban, the decreasing order was reversed: C
max > A
st > V
t ([Fig. 4D, F ]). Thus, the most and least sensitive parameters to DOAC were the concentration of
thrombin near the activator (C
max ) and the rate of thrombin propagation away from the activator (V
t ), respectively. Moreover, despite a comparable decrease in C
max for in vitro experiments, the effects of the studied anticoagulants on V
t were qualitatively different, which indicates that thrombin activity does not fully
determine this difference and, accordingly, the clot growth rate ([Fig. 8A ]). A qualitatively similar pattern was also observed in ex vivo experiments.
Thus, the results obtained ex vivo were fully qualitatively consistent with the in
vitro experimental data ([Fig. 8 ]).
Discussion
Currently, DOACs are considered good replacements for heparins, for the treatment
and prevention of thrombotic complications. However, to date, facts are gradually
accumulating, indicating that not only bleeding but also hypercoagulation disorders
are possible with the use of DOACs. It has been shown[37 ]
[38 ]
[39 ]
[40 ]
[41 ]
[42 ] that when UFH or DTIs, dabigatran, ximelagatran, lepirudin, etc., are discontinued,
a relapse of hypercoagulation may occur. The authors explain this by the fact that
thrombin can both activate coagulation and inhibit it by activating the protein C
system and the formation of prostacyclin.[39 ]
[40 ]
[41 ]
[42 ] In addition, in a rat model of disseminated intravascular coagulation, low doses
of melagatran have been shown to increase intravascular coagulation by increasing
thrombin formation.[38 ] FXa inhibitors are not thought to have a similar effect because they do not affect
the protein C system. However, in this study, we showed that after 1 week of treatment
with the studied anticoagulants, asymptomatic deep vein thrombosis was more common
in patients receiving the FXa inhibitor rivaroxaban (29%), less common in patients
receiving dabigatran etexilate (11%), and was absent from the nadroparin group ([Supplementary Table S2 ] [available in the online version]).
No symptomatic thrombotic events were observed in the patients included in the study.
In all cases of asymptomatic thrombosis, ultrasound signs of occlusive distal thrombosis
of the vein soleus were observed, predominantly at the level of the upper and middle
third ([Supplementary Table S2 ] [available in the online version]). Based on the available data, we cannot reliably
state how the number of asymptomatic thromboses in these groups quantitatively correlates.
This is due to the small number of patients in each group, especially those receiving
nadroparin. The only thing we are ready to state is that thromboses occurred in both
groups of patients receiving DOACs. This result differs from data previously described
in the literature and deserves special discussion.
Heparins irreversibly inhibit thrombin, FXa, FIXa, FXIa, and XIIa, enhancing AT activity,
especially in the case of FIXa (about 10,000 or 100,000 times for nadroparin and UFH,
respectively).[43 ] DOACs specifically inhibit only thrombin or FXa.
For the clot to begin to grow, the first small concentrations of thrombin must appear,
which activate factors V and VIII, after which clotting accelerates explosively. At
that, the concentration of the initially formed thrombin must exceed a certain threshold,
since otherwise it will be completely inhibited by natural anticoagulants (AT, α2 -macroglobulin, etc.).[44 ] The time required for the formation of the thrombin concentration required to start
accelerated coagulation determines the value of the clotting delay period (T
lag ). In our study, T
lag values changed very slightly in the presence of heparins (at physiologically possible
concentrations), but lengthened in the presence of low-molecular-weight DOACs (dabigatran
or rivaroxaban). We explain this by the fact that the rate of direct second-order
inhibition reactions is always directly proportional to the concentrations of the
target factor, the inhibitor, and the reaction rate constant (k
on ) [(3) ]:
It follows that heparins (both nadroparin and UFH) in complex with AT more slowly
intercept the first formed molecules of thrombin and FXa than DOACs, since, on the
one hand, the rate constants k
on for the thrombin or FXa association with heparin–AT complexes are approximately 2
to 10 times lower than the corresponding constants for DOACs,[43 ]
[45 ] and on the other hand, the molar concentration of heparins at their therapeutic
concentrations is much lower than for DOACs. For nadroparin, taking into account its
molecular weight (4,500 kDa) and specific activity (90 anti-Xa IU/mg),[32 ] this concentration at dose 0.6 anti-Xa IU/mL is 1.5 nM, while for dabigatran and
rivaroxaban in prophylactic doses −0.4 μM and 0.3 μM, respectively (see [Supplementary Material, section 1.2 ] [available in the online version]). We were unable to find the k
on constant for the binding of dabigatran to thrombin, but similar constants for the
binding of some other low-molecular-weight inhibitors (LB30870, melagatran, argatroban)
to thrombin are reported in the literature. The values of all these constants range
from 0.9 × 107 to 1.3 × 107 M−1 ·s−1 ,[46 ] while a constant for the binding of the AT–nadroparin complex to thrombin is 5.3 × 106 M−1 ·s−1 .[43 ] For the binding of FXa to rivaroxaban or the AT–nadroparin complex, the corresponding
constants are 1.7 × 107 M−1 ·s−1 and 1.3 × 106 M−1 ·s−1 , respectively.[43 ]
[45 ] Thus, DOACs more quickly intercept the first formed molecules of thrombin and FXa,
which lengthens the time to reach their concentrations necessary for the onset of
accelerated coagulation, lengthening the lag period of clotting compared to heparins.
The final size of the clot is determined by how far thrombin spreads in space. this
depends not only on how much thrombin was formed but also on the anticoagulant used.
The main role in the spatial distribution of thrombin is played by FIXa and FXI. FIXa
has been shown to be a major factor determining the spatial distribution of coagulation,
since unlike thrombin and FXa, it has a long plasma lifetime and can diffuse quite
far from its site of formation.[2 ]
[14 ]
[47 ] Factor XI is evenly distributed in plasma. Thrombin activates FXI at all points
where it appears and can trigger a coagulation cascade at these new points, leading
to thrombus growth. Since heparins significantly enhance FIXa inhibition, this may
explain the decreased rate of fibrin propagation and the formation of smaller clots
in their presence, in contrast to DOACs that do not inhibit FIXa.
Reversible DOACs form inactive complexes with the target factor, which can diffuse
in space much further than free factors. However, since there is an equilibrium between
inhibitor-bound and unbound forms of the factor, the breakdown of the factor–inhibitor
complex can result in a free factor away from the activator (where the factor–DOAC
complex reaches during diffusion) and trigger a coagulation cascade. According to
our hypothesis, this is the main mechanism influencing the occurrence of recurrent
hypercoagulation in a patient with a decrease in the concentration of any DOAC in
the blood. The binding of the active factor to the DOAC gradually decreases. There
is no longer a free inhibitor in the blood, but for some time the complex factor with
DOAC remains in the plasma, and its dissociation can lead to the recurrence of thrombosis.
A similar situation cannot be observed in the case of the heparins introduction.
Thus, the clot grows only while thrombin, followed by fibrin, spreads in space, that
is, the clot growth rate can determine its size. We assume that this parameter is
critical for hemostasis and should be stabilized. This study confirmed that, unlike
the TGT parameters, the clot growth rate actually has a decreased coefficient of interindividual
variability.[18 ]
[48 ]
[49 ] Besides, it can be assumed that most of the thrombin, formed after the appearance
of a thrombus, is necessary not only for the activation of platelets, the protein
C system, thrombin activatable fibrinolysis inhibitor, etc. but also for maintaining
the clot growth rate in space.
A limitation of our study is the relatively small size of the groups studied ex vivo
(33, 27, and 15 patients took rivaroxaban, dabigatran etexilate, and nadroparin, respectively),
and the small number of anticoagulants explored. Therefore, we currently consider
our results as qualitative. However, we also obtained similar results for UFH in vitro
(see [Supplementary Material, section 2.6 ] and [Supplementary Fig. S2 ] [available in the online version]), confirming the fundamental difference between
heparins and DOACs.
Conclusion
Our work was the first to demonstrate a fundamental difference in the mechanisms of
action of heparins and DOACs at the stages of activation and spatial propagation of
clotting. Simultaneous measurement of parameters characterizing the formation and
spatial distribution of thrombin and fibrin became possible due to the TD-4D test.
The main findings of this work, obtained in vitro and confirmed ex vivo, are as follows:
Heparins (both UFH and LMWH [nadroparin]) at physiologically possible concentrations
have a very weak effect on the activation stage of coagulation, characterized by such
parameters as T
lag and T
max , and relatively weak on the maximum activity of thrombin formed on the activator
(C
max ). On the other hand, heparins strongly inhibit the spatial spread of thrombin and
fibrin (parameters V
t , A
st , and V
i , V
st , V
f , respectively).
DOACs, on the contrary, significantly prolong the onset of coagulation (T
lag and T
max ) but have a weaker effect on its spatial distribution than heparins (V
i , V
st , V
f , as well as V
t and A
st ). They also have a moderate effect on C
max .
With a decrease in the concentration of any DOAC (both a thrombin inhibitor and an
FXa inhibitor), thromboembolic complications may occur, which are associated with
the breakdown of complexes of these reversible inhibitors with their target factors
and the release of free active factors far from the activator.
The clinical significance of all these findings is not yet fully clear. Considering
that a decrease in the concentration of DOAC may lead to a relapse of the hypercoagulation
initially observed in the patient, one can try to discontinue DOAC against the background
of a minimum dose of heparin (for 1–2 days) in order to maintain the inhibitory capacity
of the patient's own AT and after 2 to 3 days transfer the patient only to LMWH.
Since re-thrombosis is possible due to the breakdown of factor–DOAC complexes from
which the factor is released active, one could try to create antibodies or another
type of “antidote” that would prevent the breakdown of these complexes. This would
make the anticoagulant irreversible and eliminate the cause of coagulation initiation
away from the activator. In any case, the clinical application of the presented results
requires further in-depth studies.
What is known about this topic?
Various reactions are involved in the phases of coagulation activation and propagation
of coagulation in space. How different anticoagulants act on these phases has not
been studied.
Various heparins inhibit coagulation effectively but require intravenous or subcutaneous
administration and monitoring.
DOACs have approximately the same pharmacokinetics for most patients, so they can
be used in the same doses without monitoring.
What does this paper add?
The formation of a fibrin clot is determined not only by the amount of thrombin formed
but also by its ability to spread in space. The same amount of thrombin can form fibrin
clots of significantly different sizes depending on the mechanism of thrombin inhibition
(AT-dependent or -independent).
Heparins delay spatial propagation of coagulation, but not coagulation activation
since they enhance the factor IXa inhibition by AT.
Dabigatran and rivaroxaban inhibit the activation of clotting more than its spatial
spread, since they do not inhibit FIXa, but intercept the first-formed thrombin and
FXa molecules faster than AT–heparin complexes, which prolong the period of clotting
delay.
