Keywords
coagulation - surgical hemostasis - DNA aptamers - biomarkers - enzyme activity
Introduction
Coagulation is a dynamic and temporal-spatial controlled process that becomes activated
after vessel wall injury by complex formation between activated factor VII (FVIIa)
and tissue factor (TF).[1] The resulting extrinsic activation complex activates factor X that subsequently
catalyzes the formation of subnanomolar amounts of thrombin.[2] This initial thrombin recruits and activates platelets and augments further thrombin
formation through a series of acceleration steps resulting in peak thrombin levels
of approximately 800 nM as measured through in vitro monitoring of TF-induced coagulation
activation.[3] Since formation of a stable clot is achieved at thrombin concentrations between
10 and 20 nM, it is concluded that 96% of thrombin is generated after the wound-sealing
clot has been formed. Part of this thrombin is released into the flowing blood, since
thrombin, unlike other activated coagulation factors, contains no phospholipid-binding
sites. A fraction of this blood-born thrombin is rapidly inactivated by stoichiometric
inhibitors such as antithrombin (AT), heparin cofactor II, and α-2-macroglobin.[4] The resulting thrombin–AT (TAT) complex circulates in blood with a half-life of
44 minutes.[5] Another fraction of the blood-born thrombin binds to thrombomodulin (TM) on the
surface of endothelial cells where it becomes an anticoagulant through conversion
of protein C (PC) into the active enzyme activated protein C (APC).[6] Taken together, these thrombin-inhibiting mechanisms limit the half-life of thrombin
in the circulating blood to about 60 seconds.[7] Despite this short half-life, the results of numerical simulation of thrombin profiles
have predicted thrombin blood levels reaching low nanomolar concentrations downstream
a wounded area.[8]
[9] Hence, localized coagulation activation should induce a systemic coagulation response.
Recently, we measured systemic thrombin levels in the picomolar range in surgical
patients in a small pilot study.[10] To make these low thrombin concentrations measurable, a highly sensitive oligonucleotide
(aptamer)-based enzyme capture assay (OECA) was combined with a blood sampling technique
that protects thrombin from early inactivation by endogenous inhibitors through addition
of the reversible thrombin inhibitor argatroban to the blood sampling buffer.[10] The thrombin–OECA captures thrombin–argatroban complexes using a bivalent thrombin-specific
aptamer that binds to both exosite motifs of the enzyme. After immobilization, argatroban
is washed out and thrombin quantified through hydrolysis rates of a fluorogenic peptide
substrate. Using this approach, plasma levels of free thrombin have become measurable
with lower limits of detection and quantification of 0.46 and 1.06 pM, respectively.
Although the data obtained in a small series of patients undergoing hip- and knee-replacement
surgeries using this approach have given first evidence for the presence of low levels
of circulating thrombin, the influence of spatial coagulation activation on the systemic
coagulation reactions has not been studied in detail. To quantify the systemic coagulation
response and to identify factors that influence the activity levels, we performed
activity pattern analysis in patients undergoing a broad range of elective orthopaedic
surgeries. Besides thrombin, plasma levels of the prothrombin activation fragment
1.2 (F1.2) and TAT were measured to additionally assess the total amount of thrombin
formed and that of AT-inactivated thrombin, respectively. Plasma levels of APC were
quantified as a measure for thrombin-dependent activation of the protein C pathway.[11] Plasma levels of D-dimer were quantified as a measure for the formation of cross-linked
fibrin and subsequent degradation by plasmin. In addition, results of in vitro thrombin
generation kinetics were compared with in vivo thrombin generation kinetics to study
if changes in the systemic coagulation activities influence in vitro thrombin formation
kinetics.
Patients, Materials, and Methods
Patients, Materials, and Methods
Study Design
Patients scheduled for elective orthopaedic surgery and who have given informed consent
for study participation were eligible for the study. According to the guidelines for
perioperative care published by the American College of Cardiology (ACA) and the American
Heart Association (AHA), the surgeries were graded into minor and major surgeries
depending on magnitude, type, duration, blood loss, and transfusion requirements.[12] For prevention of thromboembolic events, all patients received enoxaparin once daily
at a dosage of 40 mg subcutaneously starting 4 to 8 hours postoperatively. The study
was approved by the local ethics committee of the University of Bonn (#001/13).
Blood Sampling and Processing
Blood samples were taken 2 hours before surgery and 24 hours after surgery by puncturing
an antecubital vein using a 21-gauge winged blood collection set (Sarstedt, Nümbrecht,
Germany). During surgery, blood was taken from a vein of the lower arm via a 21-gauge
indwelling venous cannula (B. Braun, Melsungen, Germany) maintained patent by a continuous
infusion of physiological sodium chloride solution. During surgery, blood samples
were taken at incision of the skin (incision), at the middle of the surgery, and at
final suture of the skin. The middle of the surgeries was defined as removal of the
herniated intervertebral disc for nucleotomies, as implantation of the acetabular
component for total hip arthroplasties, or by definition of the surgeon intraoperatively
for all other types of surgeries. For routine coagulation analyses, blood was collected
into 3-mL trisodium citrate tubes (10.6 mM final concentration). For thrombin–OECA
or APC–OECA testing, the citrate tubes were prefilled with argatroban (final concentration
100 µmol/L) or aprotinin and r-hirudin (final concentrations 10 µmol/L and 15 µg/mL).
Filled blood tubes were stored and transported at room temperature or in crushed ice
(APC tubes) and centrifuged within 4 hours at 2,500 × g for 15 minutes, and plasma
samples were stored at –40°C until assayed.
Laboratory Analysis
F1.2 and TAT were measured using Enzygnost F1.2 (monoclonal) and Enzygnost TAT micro
enzyme-linked immunosorbent assay kits (Siemens Healthcare Diagnostics, Marburg, Germany).
The fibrin fragment D-dimer was measured using the Innovance D-dimer kit and a BCS
XP coagulation analyzer (both Siemens Healthcare Diagnostics). All analyses were performed
according to the manufacturer's instructions. Levels of thrombin and APC were measured
using the thrombin– and APC–OECA as previously described.[10]
[11] Thrombin generation kinetics were measured using calibrated automated thrombography
(CAT) (Thrombinoscope BV, Maastricht, the Netherlands) using standard reagents (Stago,
Düsseldorf, Germany) and a Fluoroskan Ascent FL plate reader (Thermo Scientific) as
described elsewhere.[13] Reference ranges for all parameters were previously established by analyzing plasma
samples of healthy individuals. Normal plasma levels of thrombin and APC were found
below the lower limit of quantification (LLOQ) of the assays.[10]
[11] Thus, the LLOQ was defined as the upper limit of the reference range of these parameters.
Statistical Analysis
The Wilcoxon signed-rank test with continuity correction (paired) was applied to assess
the statistical difference between plasma levels measured at the different time points.
The analysis was done online using R-based software as available on the wessa.net
web server (Wessa, P. [2018], Free Statistics Software, Office for Research Development
and Education, version 1.2.1, URL https://www.wessa.net/).
For the assessment of statistical differences between the percentages of values found
within or outside the reference ranges, 2 × 2 contingency tables and the Fisher exact
probability test were applied. Calculations were done online using the QuickCalcs
module on the graphpad.com homepage (http://graphpad.com/quickcalcs/contingency1.cfm).
Results
Study Patients
Thirty-five patients (14 females) with a mean age of 50.5 ± 17.4 years (mean body
mass index, 26.4 ± 2.6) were included in the study. At baseline, all patients showed
the following key coagulation parameters within the respective normal ranges: platelet
count, prothrombin time, activated partial prothrombin time, factor II activity, factor
VII activity, protein C activity and AT activity. The surgical procedures included
13 nucleotomies, 4 total hip arthroplasties, 8 arthroscopic interventions, 5 osteosynthesis
procedures, and 5 implant removals. Surgeries were categorized as minor (n = 18) or major (n = 17) according to ACH/AHA criteria. Mean interventions times (mean ± SD, 170 ± 95
vs. 60 ± 33 minutes) and mean blood loss (294 ± 351 vs. 17 ± 37 mL) were significantly
higher in the major surgery group (for details, see [Supplementary Tables 1A] and [1B]). Five patients of the major surgery group required transfusion of packed red blood
cells (transfusion trigger: Hb < 80 g/L), whereas none of the patients in the minor
surgery group required blood support.
None of the patients had a bleeding history indicative of an inherited or acquired
hemostasis disorder or a history of venous thromboembolism. Results of routine coagulation
analysis were within the respective reference ranges (data not shown).
In Vitro Thrombin Generation Kinetics Are Not Influenced by Surgical Trauma
To study whether surgery-induced activation of coagulation changes in vitro thrombin
generation kinetics, we measured the lag time, peak thrombin, and endogenous thrombin
potential values after TF-induced (5 pM final concentration) activation of coagulation
by CAT. The minor surgery group showed a moderate statistically significant decrease
of the median lag time (4.33 vs. 3.93 minutes, p = 0.019) at the middle of the surgery compared with baseline, while significant changes
to baseline were not observed at all other time points ([Fig. 1]). The major surgery group showed no significant changes to baseline for all three
parameters at any time point.
Fig. 1 In vitro thrombin generation kinetics in patients undergoing minor (light gray boxes) and major (dark gray boxes) orthopaedic surgeries. Changes in (A) lag time, (B) peak thrombin generation, and (C) endogenous thrombin potential (ETP) were measured at the indicated time points using
CAT as outlined in the text. Data are expressed as box and whiskers plots with median,
minimal and maximal values, interquartile ranges, and outliers. The dashed and dotted horizontal lines represent the established upper and lower reference limit. Significant differences
(p < 0.05) between blood collecting time points are indicated by a line dragged between
the groups (-*-).
Plasma Levels of Thrombin and APC Substantially Increase during Surgery
Plasma levels of F1.2 showed a significant increase to baseline at the middle and
end of the surgeries and returned to baseline 24 hours after end of surgery ([Fig. 2A]). The increase was comparable in patients undergoing minor and major surgeries.
TAT levels increased after start of surgery and reached median plasma levels of 148
and 245 pM at the middle and at end of surgery, respectively ([Fig. 2B]). There was no significant difference between both patient groups at any time point.
Mean thrombin levels increased from a median of 0.46 pM (limit of detection [LOD])
at baseline to 1.09 pM at incision and reached median values of 20.05 and of 11.41
pM at the middle and end of major surgeries, respectively ([Fig. 2C]). A nearly identical dynamic of thrombin formation kinetics was measured in minor
surgeries with the exception that the increase at incision was not statistically significant.
Thrombin levels returned to baseline 24 hours after surgery in both patient groups.
Fig. 2 Changes of in vivo coagulation biomarkers during orthopaedic surgeries. Plasma levels
of the (A) prothrombin fragment (F1.2), (B) thrombin–antithrombin complexes (TAT), (C) thrombin, (D) activated protein C (APC), and (E) D-dimer were measured at the indicated time points in patients undergoing minor
(light gray boxes) and major (dark gray boxes) orthopaedic surgeries. Data are expressed as box and whiskers plots with median,
minimal and maximal values, interquartile ranges, and outliers. The long- and short-dashed horizontal lines represent the established upper limit of the normal ranges and the LLOQ or LOD (thrombin,
APC) of the assays. Significant differences *p < 0.05, **p < 0.01, ***p < 0.001.
Plasma levels of APC also increased significantly during the surgical interventions
in both groups reaching median values of 6.62 and 12.76 pM at the middle and end of
the surgeries, respectively. Twenty-four hours after surgery, APC values returned
to baseline. With the exception of the time point at incision, there was no statistical
difference between plasma APC levels in both patient groups ([Fig. 2D]).
Plasma levels of D-dimer showed no increase to baseline at incision and at the middle
of the surgeries in both groups ([Fig. 2E]), but significantly increased at the end of surgery in the major surgery group.
Twenty-four hours after end of the surgery, D-dimer levels significantly increased
relative to baseline in both groups. There was no statistically significant difference
between both groups.
Levels of the different parameters in all analyzed samples are shown in [Supplementary Table 2].
Color-Coded Visualization of Surgery-Induced Changes in Systemic Coagulation Activity
To follow changes in coagulation activity on an individualized basis, a color-coded
scheme was applied ([Fig. 3]). Data were stratified in relation to the upper limits of the reference ranges.
Green indicates no increase and yellow indicates an up to twofold increase, while
data colored in dark yellow or red showed a more than twice or more than fivefold
increase, respectively.
Fig. 3 Digital dynamics of surgery-induced changes in activity patterns. The changes in
plasma levels of prothrombin fragment 1.2 (F1.2), thrombin–antithrombin (TAT) complexes,
thrombin (FIIa), APC, and D-dimer (DD) are expressed in relation to the respective
upper limit of the reference ranges. Green, no increase; yellow, up to twofold-increase; orange, up to fivefold-increase; red, more than fivefold-increase; gray, not measured. The number of stars shown behind the type of surgery represents the
number of RBCs given.
At baseline, seven patients showed no increase of all five parameters. Five patients
showed a moderate increase in thrombin formation as characterized by increased levels
of at least one of the thrombin markers. In seven patients, an underlying thrombin
activity was associated with increased APC levels, while D-dimer was within the reference
range. Six patients showed moderately increased (n = 5) or increased D-dimer levels without an increase in thrombin or APC. In four
patients, thrombin and D-dimer levels but not APC plasma levels were increased, and
in six patients, the biomarker pattern indicates increased thrombin formation associated
with increased APC formation and increased D-dimer levels.
The activity patterns measured during surgeries indicate a strong increase in plasma
levels of thrombin. This increase was not influenced by the activity status at baseline
and was associated with increased TAT levels in the majority of patients, whereas
F1.2 levels showed only a moderate increase. Among the 22 (63%) patients showing APC
levels below the LLOQ at baseline, only 1 patient showed no increase in APC plasma
levels despite a significant increase in thrombin and TAT levels. Among the nine (26%)
patients showing moderately increased APC levels at baseline, seven patients showed
a significant increase in APC levels during surgery, while APC levels in two patients
remained stable. Stable APC levels throughout surgery were also observed in the four
(11%) patients showing a more than fivefold APC increase at baseline.
Among the 19 patients showing D-dimer levels within the reference range at baseline,
10 showed increased D-dimer levels at end of surgery. The strongest increase in D-dimer
was observed in the group of 16 patients showing increased levels at baseline. Twenty-four
hours after end of surgery, 24 (69%) patients showed elevated D-dimer levels.
Discussion
This study demonstrates enhanced but balanced systemic coagulation activity in response
to localized coagulation activation as evidenced by increased levels of active thrombin
and APC in the venous circulation of patients undergoing elective orthopaedic surgeries.
To assess whether or not the size of the wounded area influences the rate of the systemic
coagulation activation, patients undergoing major and minor surgeries were included
into the study. The higher severity of surgical trauma in the major surgery group
as stratified following ACH/AHA criteria was confirmed by significantly longer intervention
times, increased rates of blood loss, and the differences in the need of red blood
cell transfusions between both study groups. While one might expect that these differing
conditions will yield divergent biomarker kinetics, no distinct patterns were observed
between groups or in patients that received transfusions ([Fig. 3]). However, while perioperative infusion volumes had not been monitored, significantly
higher infusion volumes should have been applied within the major surgery group. Thus,
it cannot be ruled out that, to some extent, the presented results are affected by
dilution effects. To rule out that the kinetics of injury-related coagulation activation
were influenced by critically low plasma levels of coagulation factors, only patients
showing plasma levels of coagulation factors within the reference ranges were included
into the study.
Fully active thrombin becomes detectable in the circulating blood immediately after
start of surgery and remains increased during the time of surgery. This indicates
that despite the high pressure of endogenous inhibiting mechanisms, the amount of
thrombin that is released into the circulation is high enough to yield plasma levels
of thrombin ranging between 10 and 100 pM. Although these concentrations are below
the threshold values of 500 and 800 pM that are required to activate platelets and
factors V and XIII,[3] one can expect that these plasma levels induce a hypercoagulable state when reaching
low flow areas or areas of circulating flow such as in the valve pockets.[14]
[15]
To estimate the relation between the total amount of thrombin formed and the amount
of thrombin that is released into the circulating blood, plasma levels of F1.2 were
compared with plasma levels of thrombin and TAT. Plasma levels of F1.2 significantly
increased during surgery, whereas plasma levels of TAT were measured within the same
molar ranges as F1.2. Considering the longer half-life of F1.2 of approximately 120
minutes as compared with 40 minutes of TAT,[5] significantly higher plasma levels of F1.2 should be expected. These data suggest
that plasma levels of F1.2 underestimate the amount of thrombin actually formed. This
finding supports previously published data.[16] One explanation for the difference between expected and measured F1.2 concentrations
could be that parts of F1.2 remain bound to phospholipid surfaces within the growing
clot and are not released into the circulation. Another explanation is that parts
of F1.2 are altered by further proteolytic processing, whereas TAT represents a stable
end product.[17] Regardless of the underlying mechanisms, these data demonstrate that F1.2 levels
should be taken with caution when data on the amount of thrombin formation in response
to localized coagulation activation are required. Hence, we were not able to assess
the amount of thrombin that is formed at the wounded area. As expected, the increase
in thrombin was associated with a rapid increase in plasma levels of APC, indicating
that a substantial part of thrombin that is released from the wounded area into the
circulation binds to TM on the endothelial cell surface. Plasma levels of APC showed
a high degree of interindividual variability. This indicates that other factors besides
thrombin modulate the APC formation capacity such as the density of the endothelial
protein C receptor and TM on the endothelial cell surface. Moreover, plasma levels
of APC are influenced by concentrations of APC inhibitors such as the protein C inhibitor
and α-2-macroglobulin. However, one could speculate that a disbalance between the
amount of thrombin formed and the subsequent APC response will induce a hypercoagulable
state and thereby contributes to the increased risk of thrombosis in individual patients
undergoing surgical interventions.
To study the kinetics of the surgery-related fibrinolytic response, we measured plasma
levels of D-dimer. D-dimer is a composite measure for reactions downstream of thrombin
generation including formation of cross-linked fibrin and its subsequent degradation
by fibrin. Throughout the surgeries, D-dimer levels remained stable and significantly
increased relative to baseline only at the end of surgeries in the group of major
surgeries. This can be explained by the greater size of the wounded area in major
surgeries resulting in higher concentrations of cross-linked fibrin. In both groups,
D-dimer levels were found to be significantly increased 24 hours after end of surgery.
This delay between increase in plasma levels of thrombin and APC on the one site and
the increase in D-dimer on the other site indicates that the fibrinolytic response
is substantially slower when compared with the fast pro- and anticoagulant responses.
Unexpectedly, the increase in all three thrombin markers showed no significant differences
between the minor and major surgery groups. This indicates that the level of coagulation
activation over time is comparably high in both groups. Most likely this can be explained
by the surgical strategy that favors atraumatic techniques and restricts the tissue
damage per time unit to a limited area even in major surgeries.
One might suspect that the rate of coagulation activation in response to the surgical
trauma will be influenced by the activity levels at baseline. By analyzing the activity
patterns on an individualized basis, a group of six patients was identified who showed
increased coagulation activities at baseline as characterized by enhanced levels of
thrombin markers, and/or APC and D-dimer. In this small subgroup, the rate of subsequent
coagulation activation during surgery was comparable to the increase observed in patients
showing no or only moderate increased basal coagulation activities. This demonstrates
that the level of coagulation activity before start of surgery only marginally influence
the coagulation activity that is achieved during surgery in patients without clinically
relevant coagulation disorders. Factors that occur during surgery such as induction
of TF expression in circulating monocytes might be more relevant.[18]
To study whether changes in the systemic activity levels of coagulation occurring
during surgery can be monitored by in vitro thrombin generation testing, we additionally
performed CAT analysis. One might expect that the increase in systemic activity levels
will shorten the lag time or will result in higher peak values. Overall, however,
CAT parameters remained within reference ranges and unaltered relative to baseline
throughout the surgeries. These results demonstrate that in vitro thrombin generation
testing, applying a final TF concentration of 5 pM, cannot detect surgery-related
changes in systemic coagulation reactions. This is supported by previously published
data and is probably a consequence of rapid inactivation of thrombin and APC by endogenous
inhibitors in the blood sampling tube rather than by the concentration of TF used.[19] In the OECA setting, reversible inhibitors of thrombin and APC are added to the
sampling tube to block these inhibiting reactions.
Another potential limitation of the study might be that a peripheral venous access
device was used for blood sampling during surgery, whereas blood samples before and
after surgery were collected by puncturing an antecubital vein. While it seems unlikely
that the different collection sites affect the results, it cannot be excluded that
the biomaterial surface led to an artificial increase of coagulation markers although
the flush then waste method was used during blood sampling. Moreover, the flush and
waste method limits the number of blood sampling points due to the loss of blood.
The data presented here demonstrate that controlled surgical trauma induces a rapid
change in systemic coagulation reactions as mainly characterized by increased levels
of thrombin in the circulating blood. Maintenance of the hemostatic balance is achieved
through rapid activation of the protein C pathway and inactivation of thrombin through
stoichiometric inhibitors such as AT. The fibrinolytic response starts in the early
postoperative period as indicated by a significant increase of D-dimer levels after
end of surgery. There was no difference in the rate of systemic coagulation activation
among the groups of patients undergoing major and minor surgeries. In general, activity
pattern analysis could be a helpful tool to estimate the prothrombotic potential of
surgical interventions on an individualized basis.
