Laboratory assessment of the anticoagulant effects of the next generation of oral anticoagulants

Authors


David Garcia, 1201 University Blvd, University of New Mexico School of Medicine, Albuquerque, NM 87131, USA.
Tel.: +1 505 925 0404; fax:+1 505 925 0408.
E-mail: davgarcia@salud.unm.edu

Abstract

Summary.  In contrast to vitamin K antagonists, which reduce the functional levels of several coagulation factors, the new oral anticoagulants specifically target either thrombin or factor Xa. These new agents have such predictable pharmacokinetics and pharmacodynamics that routine coagulation monitoring is unnecessary. However, there are still some situations in which measurement of anticoagulant effect may be required. The coagulation assays that are used to monitor heparin derivatives or vitamin K antagonists may not always accurately reflect the anticoagulant activity of the new oral anticoagulants, and specialized assays may be needed. In this article, we: (i) identify situations in which assessment of anticoagulant effect may aid treatment decisions; (ii) describe the effects of the new oral anticoagulants on the various coagulation tests; (iii) review the specialized coagulation assays that have been developed to measure the anticoagulant effects of the new oral anticoagulants; and (iv) provide a clinical perspective on the role of coagulation testing in the clinical management of patients treated with the new oral anticoagulants.

Introduction

The new oral anticoagulants fall into two main classes: direct thrombin (factor IIa) inhibitors and direct FXa inhibitors. Although agents in the two classes have distinct mechanisms of action, targeting distinct enzymes in the coagulation pathway, all of the new drugs have features in common that distinguish them from vitamin K antagonists, such as warfarin. These common features include a rapid onset of action, few drug–drug interactions, and a predictable anticoagulant response that enables fixed dosing for several indications and across a diverse range of patients with no need for routine coagulation monitoring.

Direct thrombin inhibitors bind to the active site of thrombin, thereby attenuating fibrin formation, preventing thrombin-mediated feedback activation of FV, FVIII, and FXI, and inhibiting thrombin-induced platelet activation. The direct thrombin inhibitor dabigatran etexilate (Pradaxa) has been approved as an alternative to warfarin for stroke prevention in patients with atrial fibrillation (AF) in the USA, Europe, Canada, Japan, and other regions. Dabigatran is also licensed in many countries, although not in the USA, for the prevention of venous thromboembolism (VTE) after elective hip or knee arthroplasty. Dabigatran was included as an option for both of these indications in the 2012 American College of Chest Physicians (ACCP) antithrombotic guidelines [1], and was included in the recent American Heart Association (AHA)/American Stroke Association (ASA) advisory as an alternative to warfarin for patients with AF [2].

Direct FXa inhibitors, such as rivaroxaban (Xarelto), apixaban (Eliquis), and edoxaban (Lixiana), target the active site of FXa, and so attenuate thrombin generation. Rivaroxaban is approved in the USA, Europe and Canada for VTE prophylaxis after elective hip or knee arthroplasty and for stroke prevention in AF. It has also completed trials and is under development for both the prevention of recurrent ischemia in patients with stabilized acute coronary syndrome (ACS) [3] and for the treatment of deep vein thrombosis (DVT) and pulmonary embolism (PE) [4]. Rivaroxaban is already licensed in Europe and Canada for DVT treatment, and approval for the PE indication is pending. In the USA, rivaroxaban was approved in November 2012 for VTE treatment, and it is likely that it will also be licensed for the ACS indication. Rivaroxaban was included as an option for VTE prevention after hip and knee surgery in the 2012 ACCP antithrombotic guidelines [1], and was included in the recent AHA/ASA advisory as an alternative to warfarin for patients with AF [2].

Apixaban has completed trials (and is currently approved in Europe and Canada) for VTE prophylaxis after elective hip or knee arthroplasty [5–7], and two ongoing trials are evaluating apixaban for VTE treatment. Although it is not approved for this indication in the USA, apixaban was included as an option for VTE prophylaxis after elective hip or knee arthroplasty in the 2012 ACCP antithrombotic guidelines [1]. Apixaban has completed two trials for stroke prevention in AF, one against aspirin [8] and the other against warfarin [9], and has been approved by the Food and Drug Administration and other regulatory agencies worldwide for this indication. Apixaban was included in the recent AHA/ASA advisory as an alternative to warfarin for patients with AF, and as an alternative to aspirin in patients deemed unsuitable for vitamin K antagonist therapy [2].

Edoxaban has been approved for thromboprophylaxis after elective hip or knee arthroplasty in Japan, although it was not included as an option for VTE prophylaxis after elective hip or knee arthroplasty in the recent 2012 ACCP antithrombotic guidelines [1]; development for this indication in the USA and Europe has been discontinued. Edoxaban is currently undergoing phase III clinical trials for stroke prevention in patients with AF and for VTE treatment.

Although the new oral anticoagulants were designed to be administered in fixed doses without routine coagulation monitoring, there are situations in which assessment of the anticoagulant effects of the new drugs is important. This article: (i) identifies those situations in which assessment of anticoagulant effect may be helpful; (ii) provides guidance on the tests available to accomplish this goal; and (iii) describes how coagulation tests can distinguish the effects of oral thrombin inhibitors from those of FXa inhibitors.

Potential indications for assessment of anticoagulant effects of the new oral anticoagulants

The new oral anticoagulants show less intrasubject and intersubject variability in pharmacokinetic and pharmacodynamic responses than warfarin, and have a wider therapeutic index. These features eliminate the need for routine monitoring of the new drugs [10–12]. Nonetheless, there are still specific situations in which identifying the presence/absence of an anticoagulant effect or determining the concentration of drug may be helpful for patient management (Table 1).

Table 1. Situations in which identifying the presence/absence of an anticoagulant effect or monitoring may be helpful for patient management
To monitor adherenceAn assay that could confirm the presence or absence of drug may be useful to monitor patient adherence, and, in cases of thrombosis, would help to distinguish treatment failure from non-adherence
To determine the offset of activityQualitative assessment of the presence of drug in plasma at the time of presentation may impact on treatment decisions (e.g. thrombolysis for ischemic stroke)
If a patient requires a semi-urgent invasive procedure associated with an increased risk of bleeding or in cases of unexpected trauma, qualitative assessment could identify the presence of anticoagulant; a quantitative assessment of drug concentration or anticoagulant activity may be helpful so that the risk of bleeding can be weighed against the risk of delaying the procedure
To detect overdose or drug accumulationAssessment of the drug concentration or level of anticoagulant activity may help to identify cases of suspected overdose or drug accumulation in patients with impaired renal function. For potential accumulation assessment, collection of a sample at trough is likely to be most useful
To identify the mechanism of bleedingIn patients who present with bleeding, assessment of drug concentration or, if possible, the level of anticoagulant activity may help to identify the contribution of the anticoagulant to the bleeding event

One such situation is preparation for surgery or invasive procedures. For interventions associated with a low risk of bleeding (e.g. dental cleaning or extractions, skin biopsy, and cataract surgery), anticoagulant drugs do not normally need to be interrupted. In contrast, anticoagulants need to be stopped prior to procedures associated with an intermediate risk of bleeding (e.g. polyp resection at time of colonoscopy or laparoscopic cholecystectomy) or high risk of bleeding (e.g. joint arthroplasty, urologic surgery, neurosurgery, and abdominal or pelvic surgery, particularly in association with cancer). For patients undergoing interventions or surgery in which the risk of bleeding is high, or for those in whom neuraxial anesthesia is planned, it is advisable to confirm no residual anticoagulant effect prior to the procedure. Standard guidance for the new oral anticoagulants is to hold the drug for at least four to six half-lives prior to surgery, as a precaution. This option is feasible because of the short half-lives of the new oral anticoagulants relative to warfarin, although care should be taken in cases of renal impairment, which may significantly prolong the half-life of dabigatran, and to a lesser extent, those of rivaroxaban and apixaban. When a patient requires a semi-urgent invasive procedure, quantitative assessment of drug concentration or anticoagulant activity may be helpful so that the risk of bleeding can be weighed against the risk of delaying the procedure. Likewise, in patients taking anticoagulants who suffer major trauma, assessment of anticoagulant activity may be helpful.

The clinician may also wish to determine the anticoagulant effects of the new oral anticoagulants in patients with moderate or severe renal insufficiency (creatinine clearance of 30–49 or 15–30 mL min−1, respectively), or in patients with a body weight of < 50 kg or > 120 kg, because the number of individuals with extremes of body weight included in clinical trials is usually limited. Notwithstanding this, the current Summary of Product Characteristics (SmPC, issued by the European Health Authority) information for apixaban, rivaroxaban and dabigatran suggests that the effects of these comorbidities and cofactors are sufficiently small (Table 2) that routine monitoring of new oral anticoagulants is unnecessary [10–12].

Table 2. Sources of variability in pharmacokinetic profiles of the novel oral anticoagulants
Rivaroxaban 
 Baseline variabilityInterindividual variability (%CV) ranging from 30% to 40%Note: variability in exposure may be up to 70% on the day after surgery
CofactorEffect on rivaroxaban exposure
 Renal impairmentCrCl 30–49 mL min−1: 1.5-fold increase in exposure as compared with individuals with normal CrCl
CrCl 15–29 mL min−1: 1.6-fold increase in exposure as compared with individuals with normal CrCl
 Hepatic impairmentModerate hepatic impairment (Child–Pugh B) increased mean exposure by 2.3-fold
 GenderNo clinically relevant differences in pharmacokinetics between male and female patients
 AgeElderly patients showed higher plasma concentrations than younger patients, with mean AUC values being ∼ 1.5-fold higher
 Extremes of body weightExtremes in body weight (< 50 kg or > 120 kg) had only a small influence on rivaroxaban plasma concentrations (< 25%)
 PolypharmacyKetoconazole (400 mg once daily)Exposure increased by 2.6-fold; Cmax increased by 1.7-fold
Ritonavir (600 mg twice daily)Exposure increased by 2.5-fold; Cmax increased by 1.6-fold
Clarithromycin (500 mg twice daily)Exposure increased by 1.5-fold; Cmax increased by 1.4-fold
Erythromycin (500 mg three times daily)Both exposure and Cmax increased by 1.3-fold
RifampicinExposure decreased by 50%; Cmax decreased by 22%
Dabigatran 
 Baseline variabilityNot given in SmPC
CofactorEffect on dabigatran exposure
 Renal impairmentCrCl 30–50 mL min−1: 2.7-fold higher exposure as compared with individuals with normal CrCl
CrCl 10–30 mL min−1: six-fold higher exposure (two-fold increase in the plasma half-life)
 Hepatic impairmentModerate hepatic insufficiency (Child–Pugh B): no evidence of a consistent change in exposure or pharmacodynamics
 GenderIn atrial fibrillation patients, females had 1.3-fold higher trough and post-dose concentrations
 AgeElderly subjects showed increases of 40–60% in exposure and > 25% in Cmax
Age ≥ 75 years: 31% higher Cmin as compared with subjects between 65 and 75 years
Age < 65 years: 22% lower Cmin as compared with subjects between 65 and 75 years
 Extremes of body weightBody weight > 100 kg: 20% lower Cmin as compared with patients with body weight of 50–100 kg. Limited clinical data are available for patients with body weights of > 120 kg or < 50 kg
 PolypharmacyRifampicin (600 mg once daily) Cmax increased by 65.5%; exposure increased by 67%
PantoprazoleExposure decreased by 30%
VerapamilExposure increased by 1.5-fold; Cmax increased by 1.8-fold
Clarithromycin (500 mg twice daily)Exposure increased by 19%; Cmax increased by 15%
QuinidineBoth exposure and Cmax increased by > 1.5-fold
Amiodarone (600 mg)Exposure increased by 60%; Cmax increased by 50%
Ketoconazole (400 mg once daily)Exposure and Cmax increased by approximately 1.5-fold
  1. AUC, area under the plasma concentration–time curve; Cmax, maximum observed plasma concentration; Cmin, minimum observed plasma concentration; CrCl, creatinine clearance; CV, coefficient of variance; SmPC, Summary of Product Characteristics. Created from information taken from the Eliquis, Xarelto and Pradaxa SmPCs [10–12], issued by the European Health Authority (CHMP).

Apixaban 
 Baseline variabilityIntrasubject variability of ∼ 20% CVInter-subject variability of ∼ 30% CV
CofactorEffect on apixaban exposure
 Renal impairmentCrCl 30–50 mL min−1: 1.29-fold higher exposure as compared with individuals with normal CrCl
CrCl 15–29 mL min−1: 1.44-fold higher exposure as compared with individuals with normal CrCl
 Hepatic impairmentMild hepatic impairment (Child–Pugh A) or moderate hepatic impairment (Child–Pugh B) did not alter exposure, with pharmacokinetics similar to those in healthy subjects
 GenderApproximately 18% higher exposure in females than in males
 AgeElderly patients (> 65 years) show approximately 1.3-fold higher exposure
 Extremes of body weightBody weight > 120 kg: 30% lower exposure as compared with exposure in subjects with body weight of 65–85 kg
Body weight < 50 kg 30% higher exposure as compared with exposure in subjects with body weight of 65–85 kg
 PolypharmacyKetoconazole (400 mg once daily)Exposure increased by two-fold; Cmax increased by 1.6-fold
Diltiazem (360 mg once daily)Exposure increased by 1.4-fold; Cmax increased by 1.3-fold
Naproxen (500 mg)Exposure increased by 1.5-fold; Cmax increased by 1.6-fold
RifampicinExposure decreased by 54%; Cmax decreased by 42%
Atenolol (100 mg)Exposure decreased by 15%; Cmax decreased by 18%

Assessment of the anticoagulant effects of the new oral anticoagulants can also be helpful when the administration of tissue-type plasminogen activator to patients presenting with an acute ischemic stroke is being contemplated, and/or to determine whether the stroke was the result of non-adherence. Likewise, in patients with major bleeding, assessment of the anticoagulant effects of the new oral anticoagulants may help to identify the mechanism responsible for the bleed. Finally, assessment of the anticoagulant effect can provide reassurance that the new oral anticoagulants are working.

Assessment of the anticoagulant effects of the new oral anticoagulants

Unlike warfarin and heparin derivatives, the new oral anticoagulants target only a single clotting enzyme. The global tests of coagulation that are routinely used to monitor the anticoagulant activity of conventional anticoagulants, such as the prothrombin time (PT), standardized as the International Normalized Ratio (INR), and the activated partial thromboplastin time (APTT) do not precisely reflect the plasma concentrations (or the anticoagulant effect) of new oral agents [13,14]. Also, unlike the situation with warfarin, no correlation has yet been identified between any test of coagulation and the clinical outcome in patients taking any of the new oral anticoagulants. However, when the clinician needs to determine only whether an anticoagulant effect is present, absent (e.g. when deciding whether to proceed with an invasive procedure), or at supratherapeutic levels, which could occur through accumulation or overdose, traditional coagulation assays, such as the PT and APTT, may still be helpful. A summary of assays and their expected utility with dabigatran, rivaroxaban or apixaban is shown in Table 3.

Table 3. A summary of different assays and their utility with apixaban, rivaroxaban, and dabigatran
  1. APTT, activated partial thromboplastin time; dPT, dilute prothrombin time; dTT, dilute thrombin time; mPT, modified prothrombin time; PT, prothrombin time; TT, thrombin time. *Assays or reagents may not be approved for patient care purposes; check with your local laboratories before ordering the test.

 Assay availability*ApixabanRivaroxabanDabigatran
Coagulation assays    
 PTWidely availableNot usefulUseful for qualitative assessmentNot useful
   dPTNot widely availableData not availableData not availableData not available
   mPTNot widely availableUseful for qualitative assessmentData not availableData not available
 APTTWidely availableNot usefulNot usefulUseful for qualitative assessment
 TTWidely available, but turnaround time may varyNot usefulNot usefulUseful for qualitative assessment but may be abnormal even at clinically insignificant concentrations
    dTT/HEMOCLOTNot widely availableNot usefulNot usefulUseful for quantitative assessment
Chromogenic assays
 Anti-FXa assayWidely available, but turnaround time may varyUseful for quantitative assessmentUseful for quantitative assessmentNo effect
 Anti-FIIa assayNot widely availableNo effectNo effectUseful for quantitative assessment
 Ecarin anti-FIIa assayNot widely availableNo effectNo effectUseful for quantitative assessment

Dabigatran

The APTT shows a two-phase correlation with dabigatran: a greater than linear increase with concentrations up to 200 ng mL−1, and a linear relationship at concentrations above 200 ng mL−1 [15]. In the study by Lindahl et al. [15], the effect of dabigatran on the APTT was found to vary by only 1.26-fold between the most and least sensitive reagent tested with several different kits for this assay. The authors concluded that measurement of the APTT may be useful to identify an excess anticoagulant effect; however, the SmPC for dabigatran highlights the limited sensitivity of the test to dabigatran, making it unsuitable for precise quantification of anticoagulant effect, particularly at high dabigatran concentrations [11]. A normal APTT indicates little or no dabigatran effect.

Dabigatran prolongs the PT, often reported as an INR, in a concentration-dependent manner, but the effect is minimal with dabigatran concentrations up to 200 ng mL−1 (expected drug exposure: 50–300 ng mL−1), with INR values ranging from 0.9 to 1.2 [15]. In addition, INR determinations performed at the point of care in dabigatran-treated patients can differ from those performed in central laboratories [16–18]. Although INR values are elevated with high concentrations of dabigatran, substantial differences are noted, depending on the thromboplastin reagent used to perform the test [15]. Consequently, the PT/INR in its current form is an unsuitable test for assessment of the anticoagulant effect of dabigatran.

For the thrombin time (TT), clotting is triggered by the addition of exogenous thrombin to plasma, usually in the presence of calcium. The TT is significantly prolonged even with low concentrations of dabigatran. Despite being exquisitely sensitive to the presence of dabigatran, the traditional TT assay lacks standardization between reagents and laboratories. In contrast, a properly calibrated dilute TT (dTT) can yield a linear correlation with the plasma concentrations of dabigatran that span the therapeutic range. The HEMOCLOT (Aniara, Mason, OH, USA) assay is a modified version of the dTT assay in which the TT is measured in a 1 : 16 dilution of patient plasma with normal pooled plasma. The dTT provides a rapid method for quantitative determination of the dabigatran concentration [19]. The HEMOCLOT assay is provided with dabigatran calibrators and yields a linear relationship (R = 0.99) with dabigatran concentrations from approximately 50 to 2000 ng mL−1 [20].

As the TT and dTT assays measure thrombin activity downstream from FXa, they are insensitive to the effects of FXa inhibitors [21]. Thus, although excessive concentrations of any of the new oral agents can prolong both the PT and APTT, the TT may help to identify which of the new anticoagulants is present; in the absence of confounding effects from multiple anticoagulants, a prolonged TT indicates dabigatran, whereas a normal TT is indicative of an FXa inhibitor.

Rivaroxaban and apixaban

Inhibition of FXa activity by rivaroxaban or apixaban prolongs the PT in a dose-dependent manner, although the changes observed in the PT with approved dose regimens (apixaban 2.5 mg twice daily; rivaroxaban 10, 15 or 20 mg once daily) can be small.

Rivaroxaban has a greater effect than apixaban on the PT, and when neoplastin is used as the test reagent, the PT shows a good correlation with rivaroxaban concentration [22]. However, the changes observed in the PT of patients taking approved doses of oral FXa inhibitors can vary significantly, because many different thromboplastin reagents are available and the assay has not been standardized for this indication [23]. Indeed, the limited sensitivity of the PT to FXa inhibitory activity, coupled with the variability in the sensitivity of different thromboplastin reagents to FXa inhibitors [12,24], may limit the utility of the PT as a method for quantitative assessment of the anticoagulant effect of oral FXa inhibitors. However, the PT may still be useful to qualitatively identify the presence or absence of rivaroxaban. Because apixaban has little effect on the PT when given in approved doses, the PT is unlikely to be useful for indicating the presence or absence of apixaban. The dilute PT assay, which uses a 16-fold dilution of the thromboplastin reagent to increase the sensitivity of the test, shows less interindividual but more intraindividual variability than the standard PT assay for determination of the anticoagulant effect of oral FXa inhibitors [25]. The modified PT (mPT) assay, which also involves dilution of the thromboplastin agent (2.25-fold in 100 mm CaCl2) [26], appears to have sufficient sensitivity and variability to determine whether or not a drug effect is present with all oral FXa inhibitors, although it correlates better with apixaban concentrations than with rivaroxaban concentrations [27]. If it is adequately standardized between thromboplastin reagents, the mPT may provide a method for qualitative assessment of the presence of anticoagulant effects of all oral FXa inhibitors, including apixaban, with existing laboratory assay facilities.

Conversion of PT results to an INR increases assay variability with the new oral anticoagulants [22]. Appropriate standardization and calibration of thromboplastin reagents will improve the interpretability of the PT for FXa inhibitors, but it is preferable to report the results in seconds or PT ratio, rather than as an INR [28]. Attempts have been made to normalize PT results with rivaroxaban by calibrating different thromboplastin reagents against rivaroxaban itself, but the effect of rivaroxaban is quite variable from one PT reagent to the next [23,29].

Although both apixaban and rivaroxaban prolong the APTT in a concentration-dependent manner, the changes are small and variable. In a healthy volunteer study, single oral apixaban doses of 25 or 50 mg prolonged the APTT by only 1.2-fold [30]. In another study, the concentrations of rivaroxaban that doubled the APTT ranged from 389 to 617 ng mL−1, depending on the APTT reagent [31]; these are concentrations that exceed the expected peak concentrations of rivaroxaban (∼ 200 ng mL−1). Therefore, the APTT lacks sufficient sensitivity for accurate assessment of the anticoagulant effect of FXa inhibitors.

Searching for a solution: novel tests

As the new oral anticoagulants become more widely used in routine clinical practice, traditional PT, APTT and TT measurements will not always meet the requirements of clinicians, who sometimes require a quantitative assessment of drug concentration. Until accurate assays that are specific for use with these new oral agents become widely available, one or more of the aforementioned modified clotting assays may permit qualitative assessments in most laboratories that already perform coagulation testing. However, because these modified tests have not been approved for patient care use in the USA, and because interlaboratory variation is likely to occur, calibration and protocols for interpreting the test results are necessary.

Chromogenic assays for anti-FXa and anti-FIIa activity

Unlike that of warfarin, the plasma concentrations of the new oral anticoagulants correlate closely with their anticoagulant activities. Bioassays that are specific for measurement of anti-FXa or anti-FIIa (thrombin) activity have been used as surrogates for drug concentration assays; such testing is feasible because the new agents target only a single enzyme (either FXa or thrombin) and do not have any known clinically important metabolites.

Multiple commercial anti-FXa kits are currently available for measuring the activity of heparin and low molecular weight heparin [32]. Because these assays use an FXa-directed chromogenic substrate to monitor the activity of exogenous FXa added to patient plasma, the endpoint is color generation, rather than clot formation. Nonetheless, chromogenic assays are amenable for use with either the automated coagulometers available in most coagulation laboratories or with relatively low-cost manual spectrometers. However, these chromogenic assays are not currently approved for patient care purposes, and additional validation is required before they can be widely adopted by physicians for use in making treatment decisions.

The anti-FXa assay has long been recommended [33] for monitoring therapeutic doses of low molecular weight heparin, and is available in many clinical laboratories. For rivaroxaban and apixaban, anti-FXa activity correlates linearly with plasma concentration throughout the ranges expected in therapeutic use [32,34–36], and kits with rivaroxaban calibrators and controls are now commercially available (e.g. STA-Liquid anti-Xa [Diagnostica Stago, Parsipanny, NJ]; TechnoChrom [Technoclone, Vienna, Austria]). Although the anti-FXa assay will not detect the anticoagulant effect of dabigatran, it is not specific for any particular FXa inhibitors, and will also be affected by other agents with anti-FXa activity, such as heparin, low molecular weight heparin, and fondaparinux [37]. This may add to the complexity of interpreting anti-FXa assay results in patients receiving more than one anticoagulant.

Like the anti-FXa assays for oral FXa inhibitors, anti-FIIa chromogenic assays may be used to assess dabigatran plasma concentrations with the same equipment as used for the chromogenic anti-FXa assay [38,39]. Kits for anti-FIIa assays that can measure dabigatran levels are commercially available (e.g. BIOPHEN Heparin anti-IIa [Aniara, West Chester, OH, USA]; ACTICHROME Heparin [anti-IIa] [American Diagnostica, Stamford, CT, USA]), but few clinical laboratories perform this test.

The importance of the timing of the last dose in assessment of the anticoagulant activity of the new oral anticoagulants

In patients taking daily warfarin or other vitamin K antagonists, the timing of INR testing relative to administration of the last dose of drug is unimportant, because these medications have long pharmacodynamic half-lives. This is, in part, because they cause the synthesis of dysfunctional clotting factors; once created, these dysfunctional proteins can circulate for days, and the anticoagulant effects of vitamin K antagonists will not vary significantly from one hour to the next. In contrast, the direct anticoagulant activity and resulting rapid onset/offset of anticoagulant effect associated with the new oral anticoagulants, coupled with their relatively short half-lives, make it critically important to determine when the last drug dose was taken when evaluating the results of coagulation tests. Plasma concentrations of the new oral anticoagulants can vary by as much as four-fold with twice-daily administration, and there can be a 10-fold to 20-fold difference between peak and trough concentrations with once-daily administration, even in healthy individuals [35]. Therefore, without knowledge of the exact time when the blood sample was collected relative to the administration of the last dose of medication, results of laboratory tests, such as the dTT or anti-FXa activity, will be difficult to interpret. For assessment of adherence, these tests only provide information about recent drug consumption. For assessment of anticoagulant activity in patients with moderate to severe renal impairment or at extremes of body weight, trough values obtained immediately prior to the next scheduled dose are likely to be the most informative.

Finally, even when the timing of the last dose of drug is known and a sample is collected when the plasma concentration is at peak or trough, there are no data showing a correlation between any given level of anticoagulation effect, including anti-FXa or anti-thrombin activity, and efficacy or bleeding outcomes at the individual patient or population level. This is in contrast to the fairly well-established relationships (between high INR values and bleeding, or between low INR values and thrombosis) that apply to patients taking vitamin K antagonists; such relationships have been established over > 60 years of clinical experience with these agents [40].

How should physicians properly use/interpret coagulation test results?

Although a method with which to assess the anticoagulant effects of the new oral anticoagulants is desirable to identify potential treatment failures or the appropriate timing of invasive procedures, there are potential downsides to the widespread availability of such assays. The seemingly harmless decision to assess the anticoagulant effect of a new oral anticoagulant in an asymptomatic, stable patient on chronic therapy could be hazardous if it prompted the clinician to deviate from the evidence-based, regulatory agency-approved dose of the medication. Indeed, clinicians may overestimate the significance of an assay measurement that slightly exceeds (or falls below) the laboratory-defined ‘normal range’. Although pharmacokinetic studies with these novel agents have identified an expected range of drug concentrations for particular clinical trial patient populations, this range does not necessarily define the limits beyond which the bleeding or thrombosis risk would increase significantly for a particular patient. In contrast to warfarin [41], the clinical significance of coagulation assay results that fall outside the expected range with the new oral anticoagulants is unknown. In addition, unlike a warfarin-treated patient, whose INR can be interpreted irrespective of when it is drawn, a patient taking one of the new oral anticoagulants will have a plasma concentration (and anticoagulant effect) that will vary over time. Thus, even if an evidence-based dose adjustment strategy were available, the potential to misinterpret (or act improperly upon) a single measurement would be significant.

Summary

The new oral anticoagulants do not require monitoring during routine clinical use. However, the anticoagulant effects of these drugs will need to be determined, sometimes rapidly, in selected situations. Thus, there is still a need for widely available, technically straightforward and standardized assays with which to assess the anticoagulant effects of the new oral anticoagulants. Currently, however, the interpretation of any results is complicated by the timing of drug ingestion, the large number of available assays/reagents, possible confounding effects of other anticoagulants, and the difficulties associated with monitoring agents that have such wide variations between peak and trough values.

Clinicians (both specialists and primary care providers) will require substantial, ongoing education about when to assess the anticoagulant effects of the new oral anticoagulants and how to interpret laboratory results. It will be important to gather clinical experience with these assays to: (i) determine when measurements of anticoagulant effect will be helpful; (ii) define how assay results will translate into clinical decisions; and (iii) correlate anticoagulation levels with efficacy and safety. Clinicians and clinical laboratory specialists will need to collaborate to establish the role of coagulation testing in the management of patients taking the new oral anticoagulants.

Acknowledgements

Professional medical writing and editorial assistance was provided by A. Shepherd at Caudex Medical, funded by Bristol-Myers Squibb and Pfizer.

Disclosure of Conflict of Interests

D. Garcia has chaired or served on advisory boards for Boehringer Ingelheim, Bristol-Myers Squibb (BMS), Daiichi Sankyo, and Pfizer. Y. C. Barrett and E. Ramacciotti are employees of BMS. J. I. Weitz has served as a consultant and has received honoraria from BMS, Pfizer, Boehringer Ingelheim, Bayer, Janssen Pharmaceuticals, Daiichi Sankyo, Merck, and Takeda.

Ancillary