Patients on extracorporeal life support (ECLS) are at risk of developing thrombi in the ECLS circuit due to activation of the hemostatic system.1 Anticoagulants are used to prevent circuit thrombosis, the most common being unfractionated heparin.2 Patients on ECLS are also at risk of bleeding from cannula sites, chest tubes, operative sites, and other locations. Bleeding risk may be increased due to reduced coagulation factor levels including reduced vitamin K dependent factor levels in infants, consumptive coagulopathies, and other mechanisms related to underlying disease.
To balance the risk of bleeding versus circuit thrombosis, it is important to monitor anticoagulant activity and hemostatic factor levels. A variety of monitoring tests are available, including prothrombin time (PT), partial thromboplastin time (PTT), activated clotting time (ACT), and antifactor Xa heparin activity (aXa), which have variable sensitivity to coagulation factor levels and unfractionated heparin activity. The purpose of this study was to compare different hemostatic assays (PT, PTT, ACT, aXa) to determine their sensitivity to changes in the levels of coagulation factors, antithrombin, unfractionated heparin, platelets, and hematocrit. We evaluated these assays using 1) in vitro prepared samples with known levels of different coagulation factors, heparin, platelet counts, and hematocrits; 2) in vivo data from pediatric patients on ECLS; and 3) assay imprecision based on laboratory quality control and patient data.
Materials and Methods
This study was approved by the Seattle Children’s Hospital Human Subjects Review Committee. Outdated Group O red blood cells and platelets were obtained from BloodWorks Northwest (Seattle, WA). Patient data were obtained from retrospective review of the clinical record, and no blood was obtained from patients for this study. The anticoagulants used in this study were nine volumes of blood plus one volume of 0.105 mol/L (3.2%) trisodium citrate for coagulation tests and EDTA anticoagulated whole blood for platelet counts. Citrated samples were centrifuged at 4,600 g for 3 minutes resulting in platelet poor plasma with residual platelet counts of less than 10,000/µl. The citrated platelet-poor plasma was removed and frozen at −80°C.
Activated clotting times were measured using ACT-LR cartridges on a Signature Elite (Accriva Diagnostics, Bedford, MA). Prothrombin time using Neoplastin CI Plus reagent, PTT using PTT A reagents, anti-Xa unfractionated heparin activity using Liquid anti-Xa reagents, and kinetic fibrinogen using Fibrinogen reagent were measured on Stago Compact analyzers (Diagnostica Stago, Budd Lake, NJ). Owren-Koller (OK) buffer was obtained from Stago. Platelet counts and hematocrits were measured in EDTA anticoagulated blood on Sysmex XN-3,000 analyzers (Sysmex Inc., Lincolnshire, IL). Pooled normal citrate plasma and factor XII deficient plasma were obtained from PrecisionBiologics Inc. (Dartmouth, Nova Scotia, Canada). Antithrombin deficient plasma was obtained from Haematologic Technologies (Essex Junction, VT). Porcine unfractionated heparin was obtained from Fresenius Kabi USA, LLC (Lake Zurich, IL).
Preparation of In Vitro Samples
Two groups of in vitro samples were prepared for comparison of hemostasis assays. In the first group, only a single parameter was changed while all other parameters were kept constant. Varying a single parameter showed the optimum sensitivity of the assay to changes in a single parameter. In the second group, the same range of values for each parameter was used, but each sample had a randomized level of heparin, coagulation factor, platelet count, and hematocrit covering the overall range. Varying different parameters randomly simulated the clinical ECLS environment where multiple hemostatic factors are varying at the same time.
To evaluate the sensitivity of the PT, PTT, and heparin activity assays to unfractionated heparin, heparin was diluted with OK buffer then added to pooled normal plasma at concentrations of 0, 0.2, 0.4, 0.6, and 1.0 U/ml. To evaluate the sensitivity to antithrombin, antithrombin deficient plasma was mixed with pooled normal plasma to produce final antithrombin concentrations of 20%, 40%, 60%, 80%, and 95%, with a constant heparin concentration of 0.5 U/ml. To evaluate the sensitivity to coagulation factors levels, OK buffer was mixed with pooled normal plasma to produce coagulation factor levels of 50%, 60%, 70%, 85%, and 100%. To evaluate the sensitivity to factor XII, factor XII deficient plasma was mixed with pooled normal plasma to produce final factor XII concentrations of 10%, 20%, 50%, and 100%. To evaluate ACT testing, platelets were added to the plasmas prepared above to a concentration of 143,000/µl and hematocrit adjusted to 30%. Samples for ACT testing were recalcified just before running the assay (final calcium concentration 15 µM). Expected heparin, antithrombin, coagulation factor, and factor XII levels in the mock whole blood samples differed slightly from the levels in the mock plasma samples due to plasma present in the platelet concentrate. Packed red cells were added to adjust hematocrit which had minimal effect on plasma volume.
To evaluate the sensitivity of the ACT to platelet count, platelets were added to pooled normal plasma at concentrations of 30,000/µl, 50,000/µl, 80,000/µl, 120,000/µl, and 185,000/µl with a hematocrit of 30%. To evaluate the sensitivity of the ACT to hematocrit, red cells were added to pooled normal plasma to final hematocrits of 15%, 25%, 35%, 45%, and 55% with a platelet count of 143,000/µl.
Ten samples each with a randomized level of heparin, coagulation factor level, platelet count, and hematocrit were prepared (Table 1). Heparin levels ranged from 0 to 1 U/ml, coagulation factor levels from 50% to 95%, platelet count from 11,000/µl to 214,000/µl, and hematocrit from 15% to 60%. For PT, PTT, and heparin activity testing, plasmas were first prepared with one of the 10 random levels of heparin and one of the 10 separately randomized levels of coagulation factors. For ACT testing, samples were further adjusted with one of the 10 separately randomized levels of platelets and randomized hematocrit. Samples for ACT testing were recalcified just before running the assay (final calcium concentration 15 µM). Again, expected heparin, antithrombin, coagulation factor, and factor XII levels in the mock whole blood samples differed slightly from the levels in the mock plasma samples due to plasma present in the platelet concentrate.
Retrospective clinical data for PT, PTT, heparin activity, ACT, fibrinogen, and platelet count were obtained on 2,196 samples from 47 pediatric patients (aged 0–18 years) undergoing ECLS. Paired data (samples collected within minutes of each other) were collected for comparison of assays. For example, 1,031 paired ACT and anti-Xa heparin activity results were available with an average time between samples of 7 minutes. Demographics for the ECLS patients are shown in Table 2 and included age, sex, type of ECLS (veno-venous or veno-arterial), duration, and reason for ECLS.
Pearson correlation coefficient (r2), linear, and polynomial regression were calculated using Microsoft Excel. Polynomial regression analysis was used to evaluate data with an apparent curvilinear correlation (heparin activity versus antithrombin, ACT, PT, and PTT versus factor level). Correlations were considered statistically significant if p was <0.05. Demographic data for patients were described using the median and range of values for age and ECLS duration. Assay imprecision was estimated from between day quality control data (mean and standard deviation).
Sensitivity to Single Parameter Changes
Anti-factor Xa heparin activity was correlated with heparin levels in plasma (r2 = 0.97, Figure 1). PTT also correlated with heparin levels in plasma (r2 = 0.99). PT showed a small change increasing from 13.4 to 14.8 seconds from 0 to 1.0 U/ml heparin. ACT showed no change up to a level of 0.5 U/ml of heparin. At 0.86 U/ml heparin, the ACT increased from a baseline of ~200 seconds to >300 seconds. Heparin activity showed a curvilinear correlation with antithrombin levels (r2 = 0.99) in samples containing a constant 0.5 U/ml heparin. PTT in pooled normal plasma containing 0.5 U/ml heparin remained elevated at ~170 seconds as antithrombin levels decreased down to 60%, but then decreased as antithrombin levels were reduced to 20% of normal. PT showed no change as antithrombin decreased in heparin containing samples. There was no significant change in ACT as antithrombin decreased in samples containing 0.5 U/ml heparin (p = 0.55).
As average coagulation factor levels decreased from 100% to 50% of normal, the PT and PTT prolonged and showed a curvilinear correlation (r2 = 0.99 for both, Figure 2). Similarly, the ACT showed a progressive increase in clotting times as coagulation factor levels decreased (r2 = 0.99, Figure 2). The PTT prolonged and showed a strong curvilinear correlation as factor XII levels decreased from 100% to 10% of normal (r2 = 0.99). In contrast, the PT showed no change as factor XII levels varied. The ACT showed a progressive increase in clotting times as factor XII activity decreased (r2 = 0.88). In these samples with no heparin, aXa activity remained undetectable with changes in coagulation factor level and factor XII. The ACT progressively prolonged as platelet counts decreased from 185,000/µl to 30,000/µl (r2 = 0.93). The ACT showed no significant change as hematocrit decreased from 55% to 15%.
Sensitivity to Random Parameter Changes
Ten samples each with a randomized level of heparin, coagulation factor level, platelet count, and hematocrit were prepared (Table 1). The aXa assay was correlated with heparin level (r2 = 0.99, Figure 3) in samples with randomization of hemostatic parameters. PTT showed a rapid rise in clotting times in samples with randomized levels of heparin and clotting factor levels up to heparin levels of 0.4 U/ml; above this the PTT was over the measurement limit of the assay (>236 seconds). There was no significant change in PT in response to different levels of heparin in randomized samples. The ACT showed no significant correlation with heparin levels in randomized samples (p = 0.47). The PT showed a progressive rise in clotting times as coagulation factor levels fell (r2 = 0.88) in randomized samples. The PTT showed no significant correlation with coagulation factor levels in samples with randomized heparin and coagulation factor levels (p = 0.87). The ACT showed no significant correlation with coagulation factor level in randomized samples (p = 0.42). Similarly, the ACT showed no significant correlation relative to platelet counts or hematocrit in randomized samples.
Comparison of Assays Using Patient Data
There was no correlation between aXa measurements and ACT in samples from pediatric patients on ECLS (r2 = 0.0007, Figure 4). The PTT increased with higher aXa measurements (r2 = 0.13). ACT showed the best correlation with PTT in patient samples (r2 = 0.46, Figure 4). The ACT decreased with higher platelet counts (r2 = 0.13, Figure 5) and higher fibrinogen levels (r2 = 0.14). Both PTT (r2 = 0.07) and ACT (r2 = 0.28) increased as the PT increased.
Figure 6 shows between-day control imprecision data compared to the therapeutic range for heparin activity, PTT, and ACT. The 2SD heparin activity assay imprecision (0.41–0.51, 0.46 ± 0.027 U/ml, n = 44) was approximately half of the width of the therapeutic range 0.3–0.6 (Figure 6). The 2SD PTT imprecision (78.6–87.4 seconds, 83 ± 2.2 seconds, n = 42) was about one-third the width of the PTT heparin therapeutic range 70–100 seconds (Figure 6). In contrast, the ACT imprecision (172–240 seconds, 206 ± 17 seconds, n = 88) was greater than the ACT ECLS therapeutic range 180–220 seconds. The bottom graph in Figure 6 shows ACT data from one patient demonstrating the random variation in clinical ACT data is similar to what is seen for ACT quality control results.
During ECLS it is important to monitor both anticoagulant activity and coagulation factor levels to help minimize the risk of both bleeding and circuit thrombosis. Currently, monitoring of hemostasis during ECLS is not standardized.3 One approach is to use an assay that is sensitive to both coagulation factor levels and anticoagulants like the ACT or PTT as a measure of overall hemostasis. An alternate approach is to focus on tests that are primarily sensitive to only coagulation factor levels like PT or only anticoagulant activity like aXa, but not both. In this study, we evaluated the sensitivity of the PT, PTT, ACT, and aXa assays to a series of hemostatic parameters including levels of coagulation factors and unfractionated heparin.
In vivo and in vitro heparin accelerates the inhibition of coagulation by antithrombin. The active product in blood during heparin therapy is heparin/antithrombin complex. The aXa assay measures the rate of factor Xa inactivation by heparin/antithrombin complexes in plasma. The aXa measurements are routinely used for monitoring of unfractionated heparin levels and have advantages over monitoring using the PTT.4,5 The aXa correlates with heparin dose during ECLS, unlike ACT which shows no correlation with heparin dose.6 In this study, the aXa assay was sensitive to both heparin and antithrombin levels in plasma, but not to average coagulation factor levels or factor XII levels, making it the most specific for heparin activity measurements in both ideal conditions with only one factor changing and in simulated ECLS conditions with multiple hemostatic parameters changing.
The PT is a citrate plasma clotting time that uses tissue factor and phospholipid (complete thromboplastin) to activate factor VII and in turn the coagulation system. Most PT reagents contain a heparin neutralizer like polybrene to make the assay relatively insensitive to unfractionated heparin levels up to about 1.0 U/ml. Under ideal conditions with only a single factor changing and simulated ECLS with multiple factors changing, the PT showed a strong correlation with coagulation factor levels, but minimal change with heparin levels. The PT was also unaffected by changes in antithrombin or factor XII. Overall, the PT was the most specific test for changes in coagulation factor levels.7
The ACT is a whole blood clotting time that uses a contact system activator, like kaolin or cellite, to activate factor XII which in turn activates the coagulation system. The ACT is used during cardiac surgery to demonstrate adequate heparinization before going on cardiopulmonary bypass.8 The levels of heparin used during cardiopulmonary bypass are typically 10-fold higher (3–5 U/ml) than the levels used during ECLS (0.2–0.4 U/ml), resulting in ACT values > 400 seconds after bypass heparinization. While the ACT is commonly used to monitor heparin during ECLS,3 recent studies have reported that ACT correlates poorly with anti-Xa heparin activity or heparin dose during ECLS.2,6,9–13 ACT was not predictive of survival, bleeding, or need for circuit change.2,14,15
In this study, we found that the combination of ACT insensitivity to low dose ECLS heparin, ACT sensitivity to multiple other factors, and inherent high ACT variability lead to essentially random variations in ACT during ECLS that were not useful for heparin or hemostasis monitoring. Under optimal conditions, when only a single factor was changing, the ACT showed no change in response to increasing unfractionated heparin levels up to 0.5 U/ml, but did increase to >300 seconds at a heparin level of 0.86 U/ml. When different hemostatic factors were randomly changed in vitro, simulating the changes that occur during ECLS, the ACT showed no significant correlation with heparin levels or coagulation factor levels. Using in vivo data from pediatric ECLS patients, the ACT showed no correlation with heparin activity (r2 = 0.0007). While the ACT was sensitive to coagulation factor levels when only that parameter was changed, there was no correlation when multiple factors were changing and only a weak correlation with PT in vivo (r2 = 0.28). The ACT showed a moderate correlation with PTT (r2 = 0.46), likely due to their shared contact activation step. ACT also showed a weak correlation with platelet count (r2 = 0.13) with the highest ACT values (>300 seconds) when platelet counts were below 100,000/µl, consistent with the ACT depending in part on platelet phospholipid for its activation step.16
Prior studies have reported that ACT results are highly variable.8,17,18 We found that ACT imprecision was greater than the typical ACT heparin therapeutic range. Due to this inherent imprecision and the low sensitivity of ACT to heparin activity, an ACT value outside of the heparin therapeutic range was more likely due to random error than an actual change in heparin activity. While useful for documenting high dose heparinization during cardiopulmonary bypass, ACT is not useful for monitoring hemostasis during ECLS.
The PTT is a citrate plasma clotting time that uses a contact system activator, similar to the ACT, but includes phospholipids to replace platelets as the surface needed for coagulation activation. While the PTT is commonly used to monitoring unfractionated heparin therapy, multiple studies have reported discordance between PTT and anti-Xa heparin activity measurements during heparin monitoring.19–22 The PTT on heparin is affected by the baseline PTT value. Low coagulation factor levels, low contact system factor levels, lupus inhibitors, and other effects can result in prolonged baseline PTT values that in turn produce discordant high PTTs during heparin monitoring. High factor VIII levels can produce short baseline PTTs leading to higher heparin doses relative to anti-Xa heparin activity monitoring.23 Despite the discordance with anti-Xa heparin activity, the PTT is better correlated with heparin therapy than the ACT.9,11
In this study, we found that while the PTT was correlated with heparin level and coagulation factor level when only a single factor changed, when multiple factors were changing the PTT was more variable. Under ideal conditions with only one factor changing, the PTT was sensitive to and showed a strong correlation with increasing heparin concentrations. The PTT was also sensitive to antithrombin, coagulation factor, and factor XII levels. Under simulated ECLS conditions with multiple factors changing, the PTT showed a complex pattern with increasing clotting times from 0 to 0.4 U/ml unfractionated heparin and maximum clotting times (>236 seconds) for heparin levels above 0.4 U/ml. There was no significant correlation between coagulation factor levels and PTT when multiple factors were changing. Using in vivo data from pediatric ECLS patients, the PTT showed a weak correlation with aXa heparin activity (r2 = 0.13), and an even lower correlation with PT (r2 = 0.07), the best surrogate marker for coagulation factor levels.
If the PTT were sensitive to only coagulation factors and heparin activity, it might serve as a balance between bleeding risk due to low coagulation factors and thrombotic risk treated with heparin, but its sensitivity to contact activation interference (e.g., low factor XII), lupus-like inhibitors, and other causes makes interpretation of the PTT for many patients on heparin difficult.20 A prior study evaluating laboratory tests for estimation of coagulation factor levels in trauma found that PT was a better overall estimate of factor level than PTT, which suffered from higher false positive and false negative results.7
Limitations for this study include evaluation of only pediatric patients, possible bilirubin interference, and evaluation of only one instrument/reagent concentration for each type of hemostasis assay. Further work is needed to determine if all the conclusions in this study apply to adult patients as well. Bilirubin levels above 13.8 mg/dl can interfere with the Stago heparin activity assay (package insert November 2014 version). There were transient elevations of bilirubin above 13.8 mg/dl in four of 47 subjects in this study potentially leading to falsely low heparin activity measurements in approximately 1% to 2% of samples. Different instruments and hemostasis reagents can show substantial differences in their sensitivity to heparin levels, coagulation and contact system factor levels, platelet count, and hematocrit. Individual institutions with different instrument/reagent systems would need to evaluate the analytic sensitivity and specificity of their systems.
In summary, anti-Xa heparin activity was the most specific for measuring heparin effect in plasma, whereas PT was most specific for estimating coagulation factor levels. ACT was too insensitive to heparin activity, sensitive to too many other factors, and too imprecise to be helpful in monitoring heparin or hemostasis. PTT showed good correlation with heparin and coagulation factor levels when only one parameter was changed, but was highly variable when multiple factors were changing in vitro and when tested on in vivo samples. Based on these data, our institution primarily monitors anticoagulation and hemostasis for patients on heparin and ECLS using the aXa and PT assays with PTT used as a secondary measure in the presence of aXa assay interference such as high levels of plasma free hemoglobin or bilirubin. The ACT is being phased out as a measure of hemostasis during ECLS at this institution.
1. Urlesberger B, Zobel G, Zenz W, et al. Activation of the clotting system during extracorporeal membrane oxygenation in term newborn infants. J Pediatr 1996.129: 264–268.
2. Bembea MM, Schwartz JM, Shah N, et al. Anticoagulation monitoring during pediatric extracorporeal membrane oxygenation. ASAIO J 2013.59: 63–68.
3. Bembea MM, Annich G, Rycus P, Oldenburg G, Berkowitz I, Pronovost P. Variability in anticoagulation management of patients on extracorporeal membrane oxygenation: an international survey. Pediatr Crit Care Med 2013.14: e77–e84.
4. Guervil DJ, Rosenberg AF, Winterstein AG, Harris NS, Johns TE, Zumberg MS. Activated partial thromboplastin time versus antifactor Xa heparin assay in monitoring unfractionated heparin by continuous intravenous infusion. Ann Pharmacother 2011.45: 861–868.
5. Vandiver JW, Vondracek TG. A comparative trial of anti-factor Xa levels versus the activated partial thromboplastin time for heparin monitoring. Hosp Pract (1995) 2013.41: 16–24.
6. Nankervis CA, Preston TJ, Dysart KC, et al. Assessing heparin dosing in neonates on venoarterial extracorporeal membrane oxygenation. ASAIO J 2007.53: 111–114.
7. Yuan S, Ferrell C, Chandler WL. Comparing the prothrombin time INR versus the APTT to evaluate the coagulopathy of acute trauma. Thromb Res 2007.120: 29–37.
8. Martindale SJ, Shayevitz JR, D’Errico C. The activated coagulation time: suitability for monitoring heparin effect and neutralization during pediatric cardiac surgery. J Cardiothorac Vasc Anesth 1996.10: 458–463.
9. Khaja WA, Bilen O, Lukner RB, Edwards R, Teruya J. Evaluation of heparin assay for coagulation management in newborns undergoing ECMO. Am J Clin Pathol 2010.134: 950–954.
10. Panigada M, Artoni A, Passamonti SM, et al. Hemostasis changes during veno-venous extracorporeal membrane oxygenation for respiratory support in adults. Minerva Anestesiol 2016.82: 170–179.
11. Sulkowski JP, Preston TJ, Cooper JN, et al. Comparison of routine laboratory measures of heparin anticoagulation for neonates on extracorporeal membrane oxygenation. J Extra Corpor Technol 2014.46: 69–76.
12. Koerber JM, Smythe MA, Begle RL, Mattson JC, Kershaw BP, Westley SJ. Correlation of activated clotting time and activated partial thromboplastin time to plasma heparin concentration. Pharmacotherapy 1999.19: 922–931.
13. Moynihan K, Johnson K, Straney L, et al. Coagulation monitoring correlation with heparin dose in pediatric extracorporeal life support
. Perfusion 2017.32: 675–685.
14. Baird CW, Zurakowski D, Robinson B, et al. Anticoagulation and pediatric extracorporeal membrane oxygenation: impact of activated clotting time and heparin dose on survival. Ann Thorac Surg 2007; 83: 912–919, discussion 919.
15. Irby K, Swearingen C, Byrnes J, Bryant J, Prodhan P, Fiser R. Unfractionated heparin activity measured by anti-factor Xa levels is associated with the need for extracorporeal membrane oxygenation circuit/membrane oxygenator change: a retrospective pediatric study. Pediatr Crit Care Med 2014.15: e175–e182.
16. Moorehead MT, Westengard JC, Bull BS. Platelet involvement in the activated coagulation time of heparinized blood. Anesth Analg 1984.63: 394–398.
17. Ojito JW, Hannan RL, Burgos MM, et al. Comparison of point-of-care activated clotting time systems utilized in a single pediatric institution. J Extra Corpor Technol 2012.44: 15–20.
18. Doherty TM, Shavelle RM, French WJ. Reproducibility and variability of activated clotting time measurements in the cardiac catheterization laboratory. Catheter Cardiovasc Interv 2005.65: 330–337.
19. Takemoto CM, Streiff MB, Shermock KM, et al. Activated partial thromboplastin time and anti-xa measurements in heparin monitoring: biochemical basis for discordance. Am J Clin Pathol 2013.139: 450–456.
20. Warkentin TE. Anticoagulant failure in coagulopathic patients: PTT confounding and other pitfalls. Expert Opin Drug Saf 2014.13: 25–43.
21. Adatya S, Sunny R, Fitzpatrick MJ, et al. Coagulation factor abnormalities related to discordance between anti-factor Xa and activated partial thromboplastin time in patients supported with continuous-flow left ventricular assist devices. J Heart Lung Transplant 2016.35: 1311–1320.
22. Hanslik A, Kitzmüller E, Tran US, et al. Monitoring unfractionated heparin in children: a parallel-cohort randomized controlled trial comparing 2 dose protocols. Blood 2015.126: 2091–2097.
23. Levine MN, Hirsh J, Gent M, et al. A randomized trial comparing activated thromboplastin time with heparin assay in patients with acute venous thromboembolism requiring large daily doses of heparin. Arch Intern Med 1994.154: 49–56.