Coagulation assays and anticoagulant monitoring

Tens of millions of patients worldwide are on short- or long-term anticoagulant therapy. A variety of fast-acting and extended prophylaxis agents are available, all of which may be associated with bleeding complications. Each anticoagulant varies in its effect on routine and specialty coagulation assays and each drug may require distinct laboratory assay(s) to measure drug concentration or activity. It is thus imperative that clinicians know the appropriate assay(s) to monitor therapy, determine drug concentration, and recognize physiological or laboratory-related variables that may influence results. Furthermore, patients on anticoagulant therapies may require evaluation for either thrombophilia or bleeding disorders. Clinicians should have some awareness of how the different anticoagulants interfere with routine and specialty coagulation laboratory results.¹  This review presents the more clinically significant pearls and pitfalls of the coagulation laboratory in relation to both conventional and new anticoagulant therapies.

Unfractionated heparin (UFH) and low-molecular-weight heparin (LMWH) are fast-acting anticoagulants consisting of mixtures of negatively charged glycosaminoglycans, and fondaparinux is a synthetic pentasaccharide that recapitulates heparin's active site.²  The anticoagulant effect of UFH, LMWH, and fondaparinux is mediated through antithrombin (AT), which greatly accelerates the inhibition of activated factor II (FII), FIX, FX, FXI, and FXII; FX and FII; and FX, respectively. These agents act as nonspecific inhibitors in the laboratory and thus can potentially interfere with a host of clot-based assays.

UFH results in significant prolongation of the thrombin clotting time (TCT), with elevation of the activated partial thromboplastin time (APTT) and little to no effect on the prothrombin time (PT). (Table 1) The widely accepted view that neither LMWH nor fondaparinux prolongs the APTT is incorrect, and whether this occurs depends on the sensitivity of the reagent and the plasma concentration of the drug.^3,4  These agents do not prolong the PT. When UFH, LMWH, or fondaparinux is present in a plasma sample, a heparin neutralizer can be added in vitro before performing those assays, which may show heparin interference (eg, APTT-based factor assays, mixing studies, lupus anticoagulant [LA] screen, and confirmatory assays). Up to 1 to 2 IU/mL of heparin can be neutralized using heparinase (Hepzyme; Dade Behring/Siemens Healthcare Diagnostics), an enzyme that destroys the heparin effect. Heparinase will also neutralize the effect of both LMWH and fondaparinux. Polybrene is a positively charged material that will neutralize UFH, and this is a component of many clot-based reagents. Polybrene is added to the majority of PT reagents (clinicians may want to confirm that the PT reagent used in their laboratory contains polybrene) and is also present in many assays used in the detection of a LA, such as the dilute Russell's viper venom time (dRVVT) and Staclot LA (Diagnostica Stago) kits. If UFH is not completely neutralized in a plasma sample, it can mimic an LA by recapitulating the first 3 diagnostic laboratory criteria of LA: specifically, prolongation of the APTT, incomplete correction in mixing studies, and a positive phospholipid neutralization result.^5,6  The addition of platelets or phospholipid to a heparinized sample will often result in shortening of the clotting time due to platelet factor 4 (PF4), a potent heparin neutralizer that is found in platelets. Neither LMWH nor fondaparinux tend to cause false positive results in LA confirmatory assays. APTT-based factor assays (FVIII, FIX, and FXI) may be spuriously decreased in the presence of UFH, LMWH, or fondaparinux, and this artifact may be apparent by nonparallel curves when the samples are diluted. The PT is not prolonged with LMWH or fondaparinux, and is infrequently so with UFH, and therefore PT-based assays are rarely affected by these drugs.

Due to the potential of PF4 neutralization of UFH, whole-blood samples containing UFH are unstable and must be transported to the laboratory and centrifuged within approximately 1 hour of collection.⁷  Samples drawn to monitor UFH that cannot be received in the laboratory within 1 hour must be centrifuged and the plasma removed from the cellular component. If samples are maintained as whole blood for as little as 2 hours, up to 50% of the UFH will be neutralized by the PF4 released from platelets, leading to both a factitiously short APTT and a low heparin concentration as measured using an anti-Xa assay.⁸  Unlike samples containing UFH, whole-blood samples containing LMWH and presumably fondaparinux (no published data) are stable for 24 hours.⁹  This enhanced stability compared with UFH is because the interaction of PF4 with LMWH (and presumably fondaparinux) is weaker and therefore so is its neutralizing potential.

UFH can be measured using an anti-Xa (heparin) assay, which, when available, may be preferable to the APTT because it provides a direct measure of heparin activity.¹⁰  Most of the commercially available heparin assays used in the United States do not contain added AT and rely on endogenous AT levels in the patient sample. Although uncommon, heparin resistance (when patients require unusually high doses of heparin to achieve therapeutic anticoagulation) may occur in patients with significant AT deficiency or polymorphisms involving the heparin-binding site.^2,11–13  Heparin assays that contain AT are not recommended, because they will overestimate the level of anticoagulation in patients with AT deficiency.¹⁴  Clinicians may want to inquire with their laboratory to determine whether the heparin assay in use has added AT. If the APTT is used to monitor UFH, factitious or in vitro (laboratory-based) heparin resistance may occur in patients who have significantly elevated FVIII and fibrinogen levels, as typically occurs with any acute phase response.¹³  In a population of second- and third-trimester pregnant women studied in our laboratory, 75% of those who had therapeutic UFH levels using a heparin assay had subtherapeutic levels based on an APTT assay. A pitfall in using anti-Xa assays for monitoring not only UFH, but also LMWH and fondaparinux (see next paragraph), is the lack of assay standardization and poor comparability between commercially available kits, with differences of up to 30% in UFH levels demonstrated.¹⁵ 

Neither LMWH nor fondaparinux requires routine monitoring, although the assessment of plasma concentration may be indicated in patients at extremes of body weight, during long-term therapy, in pregnancy, or in those with impaired renal function.²  LMWH and fondaparinux cannot be monitored using the APTT assay, but rather, require an anti-Xa assay that uses an appropriate standard curve (LMWH or hybrid standard [reported in U/mL] and a fondaparinux standard [reported in mg/L], respectively). A standard curve specific for each type of LMWH is not needed.¹⁶  Patient AT levels may affect the measured concentrations in those assay systems that do not contain added AT. LMWH levels should be drawn 4 hours after subcutaneous injection, and the recommended therapeutic range for once-daily administration is approximately 1.0-2.0 U/mL, with the therapeutic range for twice-daily therapy at 0.5-1.0 U/mL.¹⁷  Fondaparinux levels should be drawn 3 hours postdose, and the observed mean peak steady-state plasma concentration is approximately 1.20-1.26 mg/L in patients receiving treatment doses (eg, 5, 7.5, or 10 mg daily).¹⁸ 

Vitamin K antagonists (VKAs) are oral, extended thromboprophylaxis anticoagulants that are derivatives of 4-hydroxycoumarin. With either vitamin K deficiency or antagonism, FII, FVII, FIX, and FX are synthesized but lack procoagulant activity because they are devoid of the posttranslational, vitamin K–dependent gamma-carboxylation needed for calcium binding and proper orientation on a phospholipid surface.¹⁹  These noncarboxylated, and therefore nonfunctional, vitamin K–dependent factors are called “proteins induced in vitamin K absence or antagonism” (PIVKAs). When stabilized on VKA therapy, reduced FII, FVII, FIX, and FX activities results in prolongation of both the PT and APTT while the TCT remains normal. In VKA samples, normal plasma-mixing studies based on the APTT show correction into the normal range, but PT-mixing studies typically demonstrate only near correction, likely due to PIVKA interference. Plasma levels of PIVKA FII can be measured using a specific immunoassay. PIVKA levels are very sensitive to vitamin K deficiency or antagonism, and a PIVKA FII assay can be useful in determining the etiology of certain factor deficiencies or the presence of vitamin K deficiency. Vitamin K deficiency may be detected with a PIVKA FII assay even before elevation of the PT.²⁰ 

Warfarin therapy may interfere with a host of specialty coagulation assays. APTT-based activated protein C resistance (APCR) ratios, even those with added FV-deficient plasma, may be falsely elevated or decreased in patients on AVK.²¹  Anticoagulant proteins C, S, and Z are also vitamin K dependent, and these activities cannot be reliably measured on VKA therapy because the levels will be decreased. Some advocate the use of protein C and total protein S Ag to FVII or FIX Ag ratios to screen for deficiencies; however, this is discouraged because its validity has never been substantiated. VKA may cause a false-positive dRVVT ratio and therefore may lead to false-positive diagnosis of LA.²²  APTT-based phospholipid-neutralization assays such as Staclot LA tend not to show spurious positive results in VKA plasmas.

The VKA dose required to achieve therapeutic anticoagulation varies between patients and therapy must be regularly monitored. The use of warfarin pharmacogenetics specifically determining the CYP2C9 and VKORC1 genotype to better predict warfarin dose requirements is not currently considered standard of care. Warfarin therapy is monitored using the international normalized ratio (INR), a mathematical conversion of the PT that was established as a means of normalizing results regardless of different thromboplastins (PT reagents) used. An INR of 2-3 is considered therapeutic. It was anticipated that the use of the INR, as well as the introduction of recombinant human thromboplastin reagents, would improve consistency of clotting assay results within and between laboratories; however, clinically significant variability still exists for a variety of reasons.^23,24  The INR is not a reliable indicator of the level of anticoagulation until approximately day 7 of VKA therapy. The INR is validated (shown to report values with acceptable precision and accuracy) for INR levels between 1.0 and 4.5 only.²⁵  Above an INR of approximately 5 or 6, results determined with different reagent/instrument systems can vary significantly. This is, in part because the sensitivity to reductions of the different vitamin K–dependent factors varies between different thromboplastin reagents, and the decrease in FII, FX, and FVII activity is not equivalent or constant, even during stable anticoagulation.²⁶  It has also been reported that point-of-care instruments tend to exhibit a positive bias in INR results of approximately 0.3 INR units and that this bias increases as INR values increase.²⁷  INR values of > 4.0 as determined using a point-of-care device should be confirmed using a standard laboratory-determined INR, especially before a change in drug dose is recommended.

The PT typically does not prolong in the presence of an antiphospholipid Ab (eg, LA) because thromboplastin reagents tend to contain a higher phospholipid concentration than APTT reagents and this neutralizes the LA effect. Although the majority of thromboplastins are not, a few PT reagents have been shown to be LA sensitive and these include: IL-PT HS (Instrumentation Laboratory), Innovin (Dade Behring/Siemens Healthcare Diagnostics), and thromboplastin C (Dade Behring/Siemens Healthcare Diagnostics).²⁸  These reagents may cause a spuriously elevated PT/INR when LA is present. If LA interference is suspected, the INR should be measured using a different, LA-insensitive thromboplastin. A chromogenic FX activity assay can also be used to monitor VKA therapy in the presence of LA because it does not show LA interference, although the therapeutic range is not well established. A chromogenic FX activity assay measures the activity of FX (a vitamin K–dependent factor) using a chromophore specific for the enzyme's active site and, in the absence of LA, is correlated with a one-stage FX activity assay. This chromogenic FX assay must be distinguished from a chromogenic anti-Xa assay because the latter is used to determine the concentration of anticoagulants that demonstrate activity against activated factor X (eg, heparin).

Serum warfarin levels can be measured directly using mass spectrometry or HPLC, and this may be useful in the evaluation of a patient who claims compliance with warfarin therapy but fails to demonstrate appropriate or proportional prolongation of the PT/INR.²⁹  Likewise, warfarin levels may be useful in the investigation of a patient with an elevated PT and APTT and normal TCT who is not prescribed warfarin but may be receiving the drug due to an improperly filled prescription or is taking the drug surreptitiously (eg, Munchausen syndrome). Serum warfarin levels will not detect the presence of superwarfarin pesticides (rat poison) such as brodifacoum or difenacoum, and these VKAs must be detected using HPLC or mass spectrometry assays designed specifically to detect these superwarfarins.³⁰  Superwarfarins, developed because rats became resistant to standard warfarin, are not only significantly more potent than warfarin, but have extremely longer half-lives due to their storage and continued release from fat stores.³¹  Their anticoagulant effect can last many months and often requires high doses of daily vitamin K as an antidote. Superwarfarin ingestion as a one-time occurrence (not infrequent in a pediatric population) generally does not require therapy, but must be distinguished from poisoning, which is associated with coagulation abnormalities and bleeding. The quintessential clinical presentation of rat poisoning is a patient, often with a psychiatric history or disorder, who presents with a greatly elevated APTT and PT, but normal TCT and is found to have very low levels of the vitamin K–dependent factors. The clotting times will normalize with a single administration of vitamin K, only to have the patient redevelop a prolonged PT and APTT within approximately 24-48 hours unless additional vitamin K therapy is administered.

Dabigatran etexilate (Pradaxa; Boehringer-Ingelheim Pharma), apixaban (Eliquis; Bristol-Myers Squibb/Pfizer), and rivaroxaban (Xarelto; Ortho-McNeil and Bayer HealthCare) are relatively new oral anticoagulant agents that do not require therapeutic monitoring. Dabigatran is a thrombin inhibitor and apixaban and rivaroxaban inhibit activated FX.

Patients on dabigatran, apixaban, or rivaroxaban may require, at some point in their therapy, laboratory investigation using one or more coagulation assays. These may be performed to: (1) determine the degree of anticoagulation in a patient with major or minor hemorrhage or requiring an invasive procedure or prone to drug accumulation due to altered metabolism of the drug, such as in the presence of seriously impaired renal function; (2) determine whether there is an underlying coagulopathy leading to the bleeding episode; and (3) evaluate for underlying acquired or hereditary thrombophilia to aid in the determination of whether anticoagulant therapy can be discontinued safely.

Dabigatran, apixaban, and rivaroxaban affect a variety of routine and specialty coagulation assays.^32,33  The APTT is more responsive to dabigatran than the PT, whereas rivaroxaban and apixaban prolong the PT to a greater extent than the APTT.³⁴  Similar to UFH and warfarin, different APTT and PT reagents show varying responsiveness to each of these agents, and therefore results in seconds cannot be standardized across laboratories.^32–34  Routine TCTs are exquisitely sensitive to the presence of dabigatran and, using some reagent systems, trough levels still result in no clot detected.³⁵  Direct Xa inhibitors will not prolong the TCT and, likewise, dabigatran will not interfere with an anti-Xa assay.

Dabigatran will interfere with most APTT-based and some PT-based assays, depending on the drug concentration, with only very high levels causing PT assay interference. Intrinsic factor assays (ie, FVIII, FIX, FXI, and FXII) show a dose-dependent spurious decrease in activity, which cannot be adequately diluted (D.M.F. and R. Gosselin, unpublished observations, February 2012). Activities of FVIII, FIX, and FXI, when measured using one-stage clotting assays, are all factitiously abnormal, ranging from < 10 to 40 IU/dL. (Table 1) Due to its activity as a thrombin inhibitor, dabigatran results in incomplete correction in mixing studies. For this reason, dabigatran may mimic a specific intrinsic factor inhibitor and can be mistaken specifically for an FVIII inhibitor. Therapeutic levels of dabigatran will yield spurious, low-positive, APTT-based (eg, FVIII, FIX, and FXI) Bethesda titers. A chromogenic FVIII activity assay is not affected by dabigatran. APCR assays may be falsely elevated, leading to misclassification of patients with FV Leiden mutations as normal. Dabigatran may cause false-positive screening and confirmatory LA assays. Most fibrinogen assays are not affected by the presence of dabigatran, although this is notably method dependent.³³ 

There are limited published data on the effect of rivaroxaban or apixaban on specialty coagulation assays (Table 1). It has been reported that rivaroxaban will factitiously decrease both one-stage and chromogenic FVIII activity assays and that this effect can be highly individualized (Dager WE, Gosselin RC, Kitchen S, Dwyre D, Multicenter analysis of dabigatran effect on international normalized ratio (INR), activated partial thromboplastin time (APTT), thrombin time and fibrinogen, Ann Pharmacother, manuscript accepted). Based on the known action of rivaroxaban and apixaban, it is likely that PT-based factor assays (ie, FVII, FX, FV, and FII) will show a dose-dependent, spurious decrease in activity; clot-based protein C and protein S results will be overestimated; Clauss fibrinogen and free protein S Ag assays will show no effect; and APCR ratios will be overestimated, although published data are lacking. Rivaroxaban and likely apixaban can cause false-positive assays in the evaluation of LA. Because these agents are inhibitors of FXa, mixing studies will lead to incomplete correction and these agents therefore may mimic a specific inhibitor of FVII, FX, FV, or FII.

Levels of the new direct thrombin or anti-Xa anticoagulant agents can be estimated using a variety of assays. In general, global assays such as the PT, APTT, and routine TCT are not ideal assays with which to measure these agents because they tend to be either too sensitive, too insensitive, or fail to show the appropriate dose response to either direct thrombin or Xa inhibitors. When using the APTT or PT to gauge the activity of any anticoagulant, a prolonged or shortened baseline will interfere with this determination. The ecarin-clotting time (ECT), a snake venom TCT-like assay, demonstrates good linearity in response to increasing concentrations of dabigatran, and a ratio of 4.0 (ie, the patient result in seconds/mean of the ECT normal range) is generally correlated with dabigatran plasma concentrations ranging from 200-300 ng/mL.³⁷  Both clot-based and chromogenic ECT assays are available commercially, although they are currently labeled for research use only.

Levels of the new oral anticoagulants are best determined using companion diagnostic assays, which use calibrators and controls specific for (or referenced against) the anticoagulant to be measured. Mass spectrometry can be used to measure all 3 drugs over a broad concentration range and with tremendous accuracy and precision, although these assay systems are not widely available, especially in most routine clinical laboratories. Dabigatran can be monitored using a modified TCT such as the Hemoclot Thrombin Inhibitor Assay (Hyphen BioMed), a prothrombinase-induced clotting time assay (PefakitPICT: DSM IP Assets), or an ECT assay, and each assay can be established using a specific dabigatran standard curve.^37–39  These assays demonstrate variation between instrument/reagent systems, and results are not as precise or reproducible as those obtained with mass spectrometry. Apixaban and rivaroxaban can be monitored using modified chromogenic anti-Xa assays that use specific apixaban or rivaroxaban standards.^36,40  Rivaroxaban can also be measured using the PT combined with rivaroxaban calibrators and controls.⁴¹  Determination of an appropriate therapeutic range for each of these new, oral agents has been difficult and has not been forthcoming from the pharmaceutical companies nor the various clinical trials. Currently, there are no US Food and Drug Administration (FDA)–approved assays to measure either the inhibitors of thrombin or FXa. Some laboratories may choose to use kits that are designated for research use only or to implement laboratory-developed tests.

Because anticoagulants are characterized by heterogeneous mechanisms in their physiological means of clot inhibition, they also vary in their effect on common and specialty coagulation assays. Therefore, an array of assays is needed to monitor or measure activities of the conventional and new oral anticoagulant agents. Interpretation of coagulation assay results in patients on anticoagulant therapy requires knowledge of their specific effect, and whether this impact is physiological or spurious, to avoid patient misdiagnosis and mismanagement.

Conflict-of-interest disclosure: The author declares no competing financial interests. Off-label drug use: None disclosed.

D. (Adcock) Funk, 8490 Upland Dr, Englewood, CO 80112; Phone: 720-568-4328; Fax: 720-568-4314; e-mail: adcockd@labcorp.com.