Thromboelastometry versus free-oscillation rheometry and enoxaparin versus tinzaparin: an in-vitro study comparing two viscoelastic haemostatic tests’ dose-responses to two low molecular weight heparins at the time of withdrawing epidural catheters from ten patients after major surgery

Ten patients who had undergone oesophageal resection were included in the study after informed and signed consent. The study was approved by the local ethics committee in Lund (DNR 2010/482-100).

Blood was sampled from each patient’s indwelling central venous catheter on the day that their epidural catheter was removed, using 4.5 mL BD Vacutainer® citrate tubes. All patients had been routinely sampled the day before to assure normal renal function (creatinine, urea), routine coagulation parameters: activated partial thromboplastin time (aPTT), prothrombin time international normalized ratio (PT-INR) and platelet count (PLT). All patients had received standard thrombosis prophylaxis with enoxaparin 40 mg at 8 p.m. the evening before blood sampling, which took place between 10 a.m. and 2p.m, 14–18 h after the last dose of enoxaparin.

Enoxaparin (Klexane, Sanofi-Aventis, Guildford, UK) and tinzaparin (Innohep, Leo Pharma, Ballerup, Denmark) were diluted with isotonic saline (9 mg/mL NaCl: Fresenius Kabi, Bad Homburg, Germany) to concentrations of 10, 20 and 30 IU/mL. 60 μL aliquots of saline containing 0, 10, 20 or 30 IU/mL enoxaparin or tinzaparin were then added to 2 mL portions of pre-warmed (37 ° C) citrated blood from each patient to obtain plasma concentrations of 0, 0.5, 1.0 and 1.5 IU/mL of enoxaparin and tinzaparin, respectively, assuming that the blood samples had a haematocrit of 40 %. The samples were incubated for 10 min at 37 °C.

The concentrations of LMWH between 0 and 1.5 IU/ml in this study encompass both thromboprophylactic levels of 0.2-0.4 IU/mL, and higher anti-FXa levels that are above recommended levels [17].

Clot formation and lysis was studied using thromboelastometry (ROTEM®, Pentapharm, Munich, Germany) and FOR (ReoRox G₂®_, MediRox, Nyköping, Sweden). Analyses were run at 37 °C within 1 h of sampling.

The ‘R’ Statistical environment (version 3.1.3: www.​r-project.​org) was used for statistical calculation and to create diagrams. Correlations were tested using Spearman’s test. The significances of differences between results for different levels of heparinization using the same LMWH and different LMWH’s at the same level of heparinization were tested using the Wilcoxon signed rank test. A P-value of <0.05 was considered significant. Friedman’s analysis of variance was used to detect significant differences in the distributions of results for enoxaparin and tinzaparin, taking into account inter-individual variation and differing concentrations of LMWH. Box and whisker diagrams were constructed using R’s boxplot function. Boxes span the interquartile range and the whiskers encompass the data point furthest from the box yet within 1.5 times the length of the box from the box.

Measures of initiation of coagulation as assessed by ROTEM and FOR were significantly prolonged by increasing concentrations of both LMWH’s (ROTEM-CT, FOR-COT1 and FOR-COT2: see Table 1 and Fig. 2a-c), with significant correlation coefficients (Spearman’s Rho) of between 0.54 and 0.77, but there was a wide spread of results, with the lowest measured ROTEM-CT in the presence of 1.5 IU/mL tinzaparin being shorter than the longest ROTEM-CT in the control group (0 IU/mL tinzaparin). The two LMWH’s values of ROTEM-CT correlated to each other significantly, as did their values of FOR-COT1 and FOR-COT2 (see Figs. 3a,b and d).

ROTEM-alpha angle, ROTEM-CFT and FOR-Slope were not affected by increasing concentrations of LMWH (see Table 1 and Fig. 2d and g). FOR-(COT2-COT1), which is the time delay between when viscosity starts to increase (COT1) and when elasticity starts to increase (COT2), was significantly prolonged by increasing doses of LMWH. The median (COT2-COT1) in the presence of 1.5 IU/mL tinzaparin and enoxaparin were 50 % and 30 % respectively longer than in the absence of added LMWH, giving correlation coefficients (Spearman’s Rho) of 0.47 and 0.79 respectively (P < 0.05). Although the differences between (COT2-COT1) for the two LMWH’s at each concentration were not significantly different, there was a significant whole-data difference in the results for tinzaparin and enoxaparin (see Table 1 and Fig. 2h). ROTEM-MCF and FOR-G’max for enoxaparin and tinzaparin showed good correlation (see Fig. 3e-f) but were not affected by increasing doses of either LMWH (see Table 1).

Neither our ROTEM-ML nor FOR-ClotSR results were outside the reference ranges for a normal level of fibrinolysis, and there were no significant differences between enoxaparin and tinzaparin at any concentration, or in ANOVA whole-data analysis. There was, however, a significant but weak negative correlation between the dose of enoxaparin and ROTEM-ML (σ = −0.36, P < 0.05); and the dose of tinzaparin and FOR-Clot SR (σ = −0.41, P < 0.05), but not between the dose of enoxaparin and FOR-Clot SR or tinzaparin and ROTEM-ML (see Table 1 and Fig. 2e and f).

We found that both LMWH substances prolonged both instruments’ measures of clot initiation in a significant dose-dependent manner, suggesting possible usefulness for postoperative monitoring. FOR measures both the time to initiation of increasing viscosity (COT1), reflecting early clot initiation and the time to increasing elasticity (COT2), which corresponds to ROTEM-CT. All these parameters increased significantly with increasing doses of LMWH’s, which is in agreement with previous research [16].

Whether the great inter-individual variation that we observed precludes these techniques use in monitoring LMWH’s depends on whether the results actually reflect the coagulation status of the patients or not: although methodological variation may account for some of the variation, the complex coagulation status occurring after major surgery is also likely to cause variation in patients’ response to any given dose of LMWH and it is possible that viscoelastic tests have a place in identifying ‘sensitive’ patients for whom a ‘normal dose’ is actually an overdose: preoperative malnourishment results in a reduced capacity to produce vitamin K dependent coagulation factors, and a major inflammatory response to surgery can be expected to cause shifts in plasma levels of coagulation factors. Shifts in fluid balance in the aftermath of haemorrhage with or without excessive transfusion can cause unpredictable variations in renal function and thereby pharmakokinetics. The postoperative state generally predisposes to hypercoagulation [29]. It is tempting to attribute the great inter-individual variation detected in this study exclusively to varying ‘postoperative factors’, but the results are actually in agreement with results from healthy volunteers given a direct factor Xa inhibitor by Casutt et al. in 2012 [30]. This is clearly an under-researched area of perioperative medicine and deserves more attention.

We observed no significant correlation between the concentration of LMWH and maximum clot strength (ROTEM-MCF and CFT, and FOR-Slope and G’max), see Fig. 2 and Table 1. This confirms previous work by Feuring et al., who observed that ROTEM-MCF was only affected by supratherapeutic levels of dalteparin; but is in contrast to Gerotziafas et al. who found that therapeutic doses of enoxaparin did indeed affect TEG-MA (thrombelastography maximum amplitude, corresponds to ROTEM-MCF) in healthy volunteers [25, 31]. LMWH consistently impedes clot initiation as measured by viscoelastic tests but not clot propagation or structure, but this does not necessarily mean that LMWH does not affect clot propagation in vivo since both the ex vivo viscoelastic tests discussed in this article are flawed by the fact that they monitor coagulation in a stagnant container. In vivo coagulation takes place within or beside blood vessels in which there is blood flow. If clot initiation is too slow in a microenvironment where there is constant flow, the clot may be ‘washed away’ before it has even formed. In contrast, even a very slow-forming clot in a viscoelastic test container is able to contribute to the cell mediated positive-feedback loops that maintain propagation.

We had hypothesised that FOR might be more sensitive to LMWH’s possible attenuations of clot propagation and maximum amplitude, but this could not be confirmed by our results. Our hypothesis was based on the knowledge that the shear forces applied by rotational thromboelastometry are known to exceed the linear viscoelastic properties of clots and may therefore in themselves weaken the developing clot [32]. FOR, however, does not apply shear force to the sample: it applies a short oscillation every 2.5 s instead, which allows measurement of both viscosity and elasticity, and should also disturb the clot less than ROTEM [33]. Other differences between the techniques that may lead to differing patterns of contact activation are that the reagents used to initiate coagulation are different: ROTEM® uses thromboplastin phospholipid and ellagic acid whereas ReoRox® uses thromboplastin alone, potentially resulting in different patterns of activation. The surfaces in the ROTEM® chamber are plastic while ReoRox® is gold-plated, which may affect initiation of coagulation and reduce the tendency of the clot to loosen from the cup wall giving a false impression of fibrinolysis.

Although the statistical significance of the dose-effect of LMWH’s on measures of fibrinolysis were only tentative, Fig. 2e and f show a negative dose–response that may deserve further investigation. Previous findings suggest that LMWH’s increase rather than decrease fibrinolysis [34]: it is an interesting hypothesis that the postoperative coagulative environment may provide conditions where the inverse is true.

We found significant correlations between tinzaparin and enoxaparin for several ROTEM and FOR parameters. However, all the parameters for which a dose–response could be demonstrated in this study showed that tinzaparin had a stronger anticoagulant effect than a corresponding dose of enoxaparin in international units per millilitre (see Fig. 2 and 3). This is in line with our previous findings where tinzaparin has been shown to prolong aPTT and impede thrombin generation to a greater degree than enoxaparin, and as explained previously is due to tinzaparin having a lower anti-FXa/anti-FIIa ratio than enoxaparin. If the two LMWH’s are dosed in equal units of anti-FXa activity (‘international units’), the tinzaparin will have a stronger overall anticoagulant effect due to the anti-IIa activity which accompanies each unit of anti-FXa activity [11]. There is also evidence that UFH and LMWH’s with larger molecular weight (>2 kDa) exert an anticoagulant effect through plasma tissue factor pathway inhibitor [35].

At many institutions, including our own institution, the anti-FXa activity assay has become the clinical ‘gold standard’ for monitoring LMWH’s. Although anti-FXa activity is likely a reliable measure of LMWH concentration [36], we would advise against relying on this assay alone to titrate the dose of LMWH: we suggest that several assays (anti-FXa, aPTT, antithrombin, viscoelastic tests, possibly thrombin generation) should be run concurrently and in series. Laboratory results should be combined with clinical judgement to dose LMWH’s in patients at risk of thromboembolic or haemorrhagic complications, particularly in patients where haemorrhage could be catastrophic, such as those whose epidural catheter is due to be withdrawn. This is not particularly new: in 2009 Van et al. observed that thrombelastography was a better predictor of deep vein thrombosis than anti-FXa activity in trauma and surgical patients [14].

A criticism of the method could be that we did not test for the samples’ haematocrits and adjust the doses of LMWH accordingly: a lower haematocrit means a greater fraction of plasma in the sample, and therefore a greater ‘volume of distribution’ for the LMWH that we added. We nevertheless decided to administer LMWH to our samples in standard doses because this is what happens in clinical practice: LMWH is either prescribed in standard doses or by weight, which are rarely adjusted for renal function or haematocrit.