Share this article on:

Thromboelastography as a Perioperative Measure of Anticoagulation Resulting from Low Molecular Weight Heparin: A Comparison with Anti-Xa Concentrations

Klein, Stephen M. MD*; Slaughter, Thomas F. MD*; Vail, Parker T. MD; Ginsberg, Brian MB, ChB, FFA (SA)*; El-Moalem, Habib E. PhD; Alexander, Ratan MBBS, FRCA; D’Ercole, Francine MD*; Greengrass, Roy A. MD, FRCP*; Perumal, Thangam T. M. MBBS; Welsby, Ian MBBS, FRCA*; Gan, Tong J. MBBS, FRCA, FFARCS(I)*

doi: 10.1213/00000539-200011000-00009
Cardiovascular Anesthesia

Low molecular weight heparin (LMWH) is commonly used to prevent postoperative thromboembolism. Currently, there is no convenient test to measure the degree of anticoagulation from LMWH. This prospective study examines the relationship of thromboelastography and serum anti-Xa concentration in patients treated with enoxaparin. Twenty-four adult patients scheduled for orthopedic surgery using epidural anesthesia were enrolled. Epidural catheters were removed the morning after surgery before the commencement of subcutaneous enoxaparin 30 mg twice daily. Venous blood samples were obtained at 1) the induction of anesthesia (baseline), 2) immediately before the third dose of enoxaparin postoperatively (Day 2-trough), 3) 4 h after the third dose postoperatively (Day 2-peak), and 4) immediately before the fifth dose postoperatively (Day 3-trough). Whole blood samples were obtained for thromboelastography, activated clotting time, and anti-Xa level analyses at each of the four time intervals. At the four sample intervals, the r time (mean ± sem). (20 ± 1, 25 ± 2, 51 ± 6, 31 ± 3 mm) and the k time (9 ± 0.7, 12 ± 1, 27 ± 5, 14 ± 2 mm) of the thromboelastograph were significantly correlated with the expected peak and trough levels of LMWH and serum anti-Xa levels (P < 0.05). At the Day 3-trough, thromboelastograph r times exceeded the normal range in 6 of 25 patients (25%). Prolongation of r time and k time on postoperative Day 3 may indicate an exaggerated response to LMWH. Thromboelastography is a test that could potentially correlate with the degree of anticoagulation produced by low molecular weight heparin.

Implications Thromboelastography is a test that could potentially correlate with the degree of anticoagulation produced by low molecular weight heparin. The r time from the thromboelastogram correlates with serum anti-Xa concentration.

Departments of *Anesthesiology, †Surgery-Division of Orthopedics, and ‡Division of Biometry, Duke University Medical Center, Durham, North Carolina

June 27, 2000.

Address correspondence and reprint requests to Tong J. Gan, MD, Box 3094, Duke University Medical Center, Durham, NC 27710.

This study was, in part, supported by a grant from Rhone Poulenc Rorer, Inc.

Enoxaparin is a low molecular weight heparin (LMWH) used for prophylaxis against deep venous thrombosis (DVT). Prospective studies indicate that enoxaparin is very effective in preventing DVT in high-risk patients (1,2). Compared with unfractionated low-dose heparin, LMWH offers a higher and more predictable bioavailability after subcutaneous injection, a longer biologic half-life, and a reduced influence on platelet function and lipolysis (3). As a result, the use of LMWH has become commonplace and is frequently the agent of choice for DVT prophylaxis in orthopedic and trauma patients.

Numerous reviews and case reports have linked the increased use of LMWH with the occurrence of epidural hematomas after epidural and spinal blockade (4–7). One factor underlying these complications is that standard perioperative tests of coagulation do not reflect the extent of anticoagulation resulting from LMWH (8). LMWHs primarily inhibit coagulation factor Xa. Unfortunately, measuring anti-Xa is not readily available in the United States, and an alternative test has yet to be validated. In the absence of a quantitative measure of anticoagulation, timing of needle insertion and catheter manipulation to minimize bleeding complications is difficult. A clinically useful test to further help define the degree of anticoagulation would be useful.

Thromboelastography is a bedside blood test, which can be used to define the viscoelastic properties of blood. In addition, thromboelastography provides information about platelet activation, fibrin formation, and clot retraction (9). As a result, thromboelastography has the potential to be a clinical test for predicting anticoagulation from LMWH. Currently, there is little data exploring the relationship of thromboelastography and LMWH therapy. To investigate whether thromboelastography correlates with the activity of LMWH, we examined the relationship of thromboelastography and serum anti-Xa levels in postoperative patients treated with the LMWH, enoxaparin. We hypothesized that the r values of the thromboelastogram would correlate with anti-Xa levels.

Back to Top | Article Outline


This study was approved by the institutional review board, and written, informed patient consent was obtained. Twenty-four patients classified as ASA physical status I–III, aged 18 yr or older, participated in this prospective clinical trial. All patients were scheduled for unilateral total knee or hip arthroplasty using lumbar epidural anesthesia. Patient exclusion criteria included preoperative use of anticoagulant drugs (including concurrent use of nonsteroidal antiinflammatory drugs), morbid obesity, or contraindications to epidural anesthesia.

All patients received an epidural placed at the lumbar 3–4 or 4–5 level by using an 18-gauge Tuohy needle and loss of resistance technique. Each epidural was used to provide the primary anesthetic and left in place the night of surgery. The epidural catheters were removed the morning after surgery (2 h) before the commencement of subcutaneous enoxaparin 30-mg injections administered twice daily.

Venous whole blood samples were obtained at:

  • the induction of anesthesia (baseline),
  • immediately before the third dose of enoxaparin (postoperative Day 2-trough),
  • 4 h after the third dose (postoperative Day 2-peak), and
  • immediately before the fifth dose (postoperative Day 3-trough).

Whole blood samples were obtained simultaneously for thromboelastography, activated clotting time (ACT), and anti-Xa analyses at each of the four time intervals. Venous blood was collected by venipuncture from an antecubital vein by using a tourniquet and a two-syringe collection technique. The blood samples were collected into Vacutainer tubes (Becton Dickinson, Rutherford, NJ) containing sodium citrate 3.8% (9:1 vol/vol). The initial 5 mL of blood was discarded. Whole blood was transferred within 4 min to prewarmed cuvettes of a Thromboelastograph Coagulation Analyzer (TEG®; Haemoscope Corp, Skokie, IL). All thromboelastographic tests were performed in duplicate. Variables recorded were r time (normal range = 19–28 mm), k time (normal range = 8–13 mm), maximum amplitude (MA) (normal range = 48–60 mm), and α angle (normal range = 29°–43°) (Figure 1). In addition to TEG® testing, a celite ACT (Hemochron; International Technodyne, Inc, Edison, NJ) (normal range = 90–120 sec) was performed in duplicate and the mean value reported. Trained research assistants performed all tests, and the TEG® and ACT were calibrated and maintained according to institutional laboratory standards.

Figure 1

Figure 1

The remaining blood from each sample interval underwent immediate centrifugation at 3000 g, and the plasma was stored at −70°C until analysis. Anti-Xa concentrations were determined on an automated analyzer (MDA-180; Organon Teknika, Durham, NC) by using a chromogenic assay (Chromostrate MDA; Organon Teknika). The coefficient of variation for the assay was 2.4% at heparin concentrations of 0.45 U/mL and 6.2% at 0.08 U/mL. Duplicate measures were performed for each sample, and the assays were performed in parallel with manufacturer-provided standards. The average value between the two measurements was used.

Descriptive statistics for patient demographics were recorded. The statistical analysis used a generalized estimating equation (GEE) approach to handle the repeated measurements (10). The GEE approach is a novel way of handling correlated data. The method is similar to a linear mixed model but with no assumption about the distribution of the response vector. The dependent variables on the left side of the regression equation are stacked in one column, yielding clusters of data for each individual. The GEE models the expected value and covariance structure of this column. The interpretation of the variable estimates is similar to the case of ordinary linear regression. The Cochran-Mantel-Haenszel test was used to test for trend in time of the variables. Kendall tau b correlation was used at each time point to assess strength of relationship between r time and anti-Xa. Power calculation revealed a sample size of 24 patients was required to detect a 30% difference in baseline r values at a significance level of 0.05 and a beta of 0.2. A P value of less than 0.05 was considered statistically significant.

Back to Top | Article Outline


The patient age, weight, sex, and ASA physical status are presented in Table 1. The mean ± se for the TEG® variables, ACT, and anti-Xa concentration are shown in Table 2. The r time correlated with the expected peak and trough levels of LMWH and anti-Xa concentrations (P < 0.05) (Figure 2). ACT values (124 ± 4, 124 ± 6, 141 ± 4, 147 ± 5 s) increased with additional doses of LMWH, similar to that seen by additional doses of unfractionated heparin.

Table 1

Table 1

Table 2

Table 2

Figure 2

Figure 2

At the Day 3-trough, 6 of 24 patients (25%) had an r time greater than normal, mean ± sd of 39 ± 5 mm. The remaining 18 patients had an r value of 22 ± 7 mm. The k time at the Day 3-trough was greater than normal in 9 of 24 patients (38%), 22 ± 7 mm. The remaining 15 patients had a k value of 7 ± 3 mm. The MA demonstrated the least variation of the four TEG® variables. Each of the four time measurements were nearly identical, reflecting LMWH’s minimal effects on platelet function as measured by using thromboelastography. No clinically significant adverse bleeding events were detected in any of the patients.

Back to Top | Article Outline


The results of this study suggest that the thromboelastographic r value correlates significantly with the peak and trough levels of anti-Xa activity after the administration of enoxaparin (P < 0.05). LMWHs are unique in their ability to preferentially inhibit factor Xa more than factor IIa. The degree of factor Xa inhibition is directly related to its anticoagulation effect. As a result, this correlation with anti-Xa levels may also reflect the relationship of thromboelastography and clinical anticoagulation.

LMWH has many effects on the coagulation cascade that can vary with individual drug formulations. Besides preferentially inhibiting factor Xa, platelet function, tissue factor-endothelial cell interactions, and antithrombin activity are also effected. As a result, measuring each individual variable of the coagulation cascade is impractical. To better evaluate clinical anticoagulation, a broad measure of the cascade is necessary. Thromboelastography is a unique test that provides an overall global assessment of the clotting process from fibrin formation to clot lysis. A major advantage to the use of thromboelastography is the ability to directly assess clot formation in real time as a point-of-care measure of coagulation.

Variables measured by thromboelastography have been previously defined (11). The r value represents the rate of initial fibrin formation and has been linked functionally to plasma clotting factors and circulating inhibitor activity. Frequently, the r time is prolonged in patients with factor deficiencies, heparin anticoagulation, or severe hypofibrinogenemia (11). Therefore, the fact that the r value is closely associated with the level of anti-Xa activity is not surprising. Gottlieb et al. (12) recently supported this conclusion with an in vivo study examining the ability of thromboelastography to predict anti-Xa activity. R value, k value, and α angle correlated with three different laboratory anti-Xa controls. In addition, a standard thromboelastogram can usually be obtained within 20 minutes, making the test convenient and clinically accessible. Because of its comprehensive nature, its ready availability, and correlation with anti-Xa concentration, thromboelastography could offer an advantage if it also correlated with clinical anticoagulation from LMWH and bleeding.

In this study, the ACT progressively increased at each of the postoperative sample intervals measured. However, the values did not correlate with the specific anti-Xa activity. ACT has been a traditional test for in vitro clot formation and a measure of unfractionated heparin. It increases linearly with increasing doses of heparin. Unfortunately, as this study has demonstrated, the ACT lacks sensitivity to monitor LMWH administration.

LMWHs can produce therapeutic anticoagulant effects within 2–4 hours after subcutaneous administration (4). Peak plasma levels occur 4 hours after subcutaneous administration and decrease to 50% of peak values 12 hours after injection. Based on these findings, it has been suggested that a 10- to 12-hour interval occur between the insertion of and removal of epidural needles and catheters and the previous administration of a LMWH dose (4). Despite these guidelines, bleeding risk cannot be predicted from available, standard laboratory tests. Lack of point-of-care measures for LMWH anticoagulant effects makes it difficult to predict which patients may be at particularly high risk for bleeding complications. In this study, there was a subgroup of patients (25%) with a prolonged r time on Day 3, which corresponded to the normal trough level of enoxaparin. Given the correlation of r value and anti-Xa level in this study, this increased value may reflect an exaggerated response to LMWH and theoretically the potential for an increased bleeding risk in these patients. The mechanism underlying an increased anti-Xa level and prolonged r time and k-time in these patients is unclear but may be related to patient-to-patient variability of the effects of LMWH.

Another reason that may account for this increased level of anticoagulation is the twice-daily dosing regimen currently used in the United States. In Europe, the incidence of epidural hematoma appears to be less than that recently seen in the United States. Some authors have suggested that the decreased incidence of epidural hematoma results from European reliance on a single daily dose of enoxaparin (13). This method of administration would presumably result in lower trough levels. The fact that in our study the normal trough level of enoxaparin on Day 3 corresponded to a prolonged r time, k time, and an increased anti-Xa concentration may be evidence that the twice-daily dosing schedule produces an exaggerated anticoagulant effect in a subset of patients.

Despite the high correlation seen among r time, k time, and anti-Xa levels in these patients treated with enoxaparin, it is difficult to draw a definitive association about all patients treated with LMWH. All LMWH have different anti-Xa/anti-IIa ratios and impact platelet function differently. Although thromboelastography is a broad measure of the coagulation cascade, extrapolating the results from these patients treated with enoxaparin to other LMWH formulations is premature. In addition, because of the complex mechanism of LMWH action, it is possible that a single test will never provide a complete measure of anticoagulant activity and, hence, bleeding risk. However, if subsequent studies can confirm the relationship between thromboelastography and anti-Xa concentrations from a clinical perspective, then thromboelastography may help define patients in whom regional anesthesia should be avoided.

This study demonstrates a correlation between r time, as determined by thromboelastography and plasma anti-Xa activity in surgical patients treated with enoxaparin. A subset of patients retained a prolonged r time and increased anti-Xa activity before LMWH administration on the third postoperative day, which may be an indication of an exaggerated response to enoxaparin. Because thromboelastography evaluates a broad spectrum of the coagulation cascade, is readily available, and correlates with anti-Xa concentration, it could be a useful test to measure LMWH activity.

We would like to thank Chris Lesar, Diana Kucmeroski, and Ben Greenberg for their assistance with the study.

Back to Top | Article Outline


1. Levine M, Hirsh J, Gent M, et al. Prevention of deep vein thrombosis after elective hip surgery: a randomized trial comparing low molecular weight heparin with standard unfractionated heparin. Ann Intern Med 1991; 114: 545–51.
2. Turpie A, Levine M, Hirsh J, et al. A randomized controlled trial of low-molecular-weight heparin (enoxaparin) to prevent deep-vein thrombosis in patients undergoing elective hip surgery. N Engl J Med 1986; 315: 925–9.
3. Buckley M, Sorkin E. Enoxaparin: a review of its pharmacology and clinical applications in the prevention and treatment of thromboembolic disorders. Drugs 1992; 44: 465–97.
4. Horlocker TT, Heit JA. Low molecular weight heparin: biochemistry, pharmacology, perioperative prophylaxis regimens, and guidelines for regional anesthetic management. Anesth Analg 1997; 85: 874–85.
5. Bergqvist D, Lindblad B, Matzsch T. Risk of combining low molecular weight heparin for thromboprophylaxis and epidural or spinal anesthesia. Semin Thromb Hemost 1993; 19: 147–51.
6. Sternlo J, Hybbinette C. Spinal subdural bleeding after attempted epidural and subsequent spinal anesthesia in a patient on thromboprophylaxis with low molecular weight heparin. Acta Anaesthesiol Scand 1995; 39: 557–9.
7. Hynson J, Katz J, Bueff H. Epidural hematoma associated with enoxaparin. Anesth Analg 1996; 82: 1072–5.
8. Walenga J. Practical issues on the laboratory monitoring of low molecular weight heparins and related polysaccharides. Semin Thromb Hemost 1993; 19: 69–72.
9. Rogers MC. Principles and practice of anesthesiology. In: Murray DJ, ed. Monitoring of hemostasis. Boston: Mosby Year Book, 1993: 853.
10. Zeger S, Liang K. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 19869;42:121–30.
11. Mallett SV, Cox DJA. Thromboelastography. Br J Anaesth 1992; 69: 307–13.
12. Gottlieb A, Mraovic B, Sprung J, et al. The usefulness of thromboelastography in determining the coagulation effects of low-molecular weight heparin. Anesth Analg 1999.
13. Tryba M. European practice guidelines: thromboembolism prophylaxis. Reg Anesth 1998; 23: 178–82.
© 2000 International Anesthesia Research Society