Direct oral anticoagulant monitoring: what laboratory tests are available to guide us?

An 85-year-old white female presents to the emergency department at 1800 hours with major lower gastrointestinal bleeding and confusion. Her hemoglobin on admission is 7.5 g/dL, and her platelet count is 145 × 10⁹/L. She states that she takes 6 pills in the morning after breakfast and 6 pills after dinner, and she is on a blood thinner and baby aspirin for some heart condition. She weighs only 45 kg. Her serum creatinine is 1.2 mg/dL, and her blood urea nitrogen is 22 IU/mL. She is unable to tell whether she is on a vitamin K antagonist (VKA) or a direct oral anticoagulant (DOAC). Her point-of-care international normalized ratio (INR) in the emergency department is 2.1. You are paged to help decide what type of reversal or hemostatic agent she should be given for her ongoing bleeding.

The following questions arise in your mind. Because her INR is 2.1, is it due to VKA or one of the DOACs that may affect prothrombin time (PT)/INR? Which DOAC might be she on—a direct thrombin inhibitor or a factor Xa inhibitor (FXaI)? What tests should I order “Stat” to know which specific reversal/hemostatic agent I would recommend?

DOACs are increasingly used in clinical practice because of several advantages over VKA that include significantly less intracranial hemorrhage, no need for routine monitoring, short onset and offset of action, and patient convenience because of no or less interference from diet and drugs. Large randomized clinical trials have shown that DOACs are effective in the treatment and prophylaxis of venous thromboembolic events and thromboembolic prophylaxis of atrial fibrillation. Unfortunately, clinical trials have many inclusion and exclusion criteria that are not applicable in daily practice. Therefore, there are many clinical scenarios that may require knowledge of either presence (qualitative) or actual levels (quantitative) of DOACs in patients’ plasmas. Because clinical trials were conducted with the advantage of “no need for laboratory monitoring” because of favorable pharmacokinetics and pharmacodynamics of DOACs, there were no tests or specific assays proposed to look at the drug effect/level. Although DOACs levels were measured using liquid chromatography mass spectrometry (LCMS) methods on stored plasma samples for future analysis, this method is impractical in daily use if DOAC levels are to be assessed for any rapid clinical decision making. There are clinical scenarios where it might be useful to know either (1) accurate peak and trough DOAC levels in patients at extremes of body weights, including children, elderly patients (>80 years of age), and patients developing thrombosis despite on therapy; or (2) a rough estimate of drug activity in patients presenting with major bleeding or needing urgent surgery that may require use of specific reversal or nonspecific hemostatic agent.^1,2  There are many recommendations published on laboratory assessment of DOACs.^3-5 

Dabigatran etexilate given twice a day is a prodrug that is converted to its active drug metabolite dabigatran by proteases in stomach, liver, and plasma. The active metabolite is only 6% to 9% of the original prodrug. Dabigatran binds tightly to free and bound thrombin. It is mainly excreted by kidneys, and hence, impaired renal function significantly increases plasma drug levels. Because it inhibits thrombin, potentially all clotting-based coagulation tests/assays can be affected.³  In patients with normal renal function, PT is often unaffected (depending on reagent’s sensitivity), whereas activated partial thromboplastin time (aPTT) is likely to be often prolonged at the peak; at trough, aPTT is likely to be prolonged only if that particular aPTT reagent is sensitive to dabigatran. Thrombin time is extremely sensitive to even a minute amount of dabigatran in the plasma. Therefore, if aPTT is normal, you cannot rule out the presence of a sufficient amount of dabigatran in the plasma; however, a normal thrombin time can confidently rule out the presence of a significant amount of dabigatran.⁶  Thus, a patient on dabigatran who needs surgery with a normal thrombin time can be taken to the operating room without concern for excessive bleeding because of anticoagulation. However, if the thrombin time and/or aPTT is prolonged, then it may be useful to know the exact amount of dabigatran for either use of idarucizumab for reversal before urgent surgery or waiting for dabigatran levels to decrease below a certain level for routine surgery. Similarly, in a patient with major bleeding with prolongation of both thrombin time and aPTT, idarucizumab may be considered for reversal, whereas if only thrombin time is prolonged and aPTT is normal, then supportive care, especially with normal renal function, can be provided.⁶ 

The LCMS method to measure dabigatran level is not feasible in routine clinical practice. Therefore, there are 2 other options for measuring dabigatran levels using appropriate calibrators and controls.³  Ecarin venom–based assays are very accurate where the venom converts prothrombin in patient’s plasma to meizothrombin that is neutralized by dabigatran. The end point of meizothrombin neutralization by dabigatran can be detected by either (1) clot-based assay (ecarin clotting time), where free meizothrombin converts fibrinogen to a fibrin clot (however, the results may be affected by low prothrombin and fibrinogen levels in the patient’s plasma), or (2) chromogenic assay (ecarin chromogenic assay; the reagent is supplemented with prothrombin), where free meizothrombin acts on a chromogenic substrate to give a color and thus, is unaffected by fibrinogen level in the patient’s plasma. Dilute thrombin time–based assay is also very accurate in measuring dabigatran levels. All above assays use commercial dabigatran calibrators and appropriate controls, with measurable range very similar to LCMS. Unfortunately, none of these assays are Food and Drug Administration (FDA) approved, and hence, each laboratory would have to perform its own validation as a laboratory-developed test. Table 1 summarizes laboratory assessment of dabigatran.

There are several FXaIs in the market, including rivaroxaban, apixaban, edoxaban, and betrixaban. Rivaroxaban was the first FXaI approved for clinical use. Thereafter, apixaban, edoxaban, and recently, betrixaban have been FDA approved for similar indications. They all bind tightly to free FXa and FXa present within the prothrombinase complex, and some also bind FXa within the thrombus. Rivaroxaban is mostly cleared by kidneys (67%) followed by edoxaban (50%), whereas apixaban is mostly cleared by the fecal route. Because they all have similar rapid onset and offset of action, it is easy to remember that the general onset of action is at a mean of 2 hours (range = 1-3 hours), mean peak is at ∼4 hours, and the mean half-lives are 12 hours (range = 10-14 hours). Unlike dabigatran, they are mostly protein bound, and hence, dialysis is not an option for removal of these DOACs.

Although they inhibit FXa akin to low-molecular weight heparin, they do not require antithrombin for their action. Thus, FXaI also generally does not affect aPTT while on therapy range. They may prolong PT depending on the sensitivity of each PT reagent, and prolongation is often concentration dependent—the higher the amount of FXaI in the plasma, the longer the PT. However, a normal PT may not rule out clinically relevant amounts of FXaI (30-75 ng/mL) in the plasma.^7,8 

Accurate measurement can be performed with LCMS, but this is not practical for real-life scenarios. There are commercial (non–FDA-approved) rivaroxaban/apixaban assays available that are based on chromogenic anti-Xa methods that use appropriate calibrators and controls to provide accurate drug levels comparable with LCMS. Some medical centers are able to provide rapid turnaround time (TAT; <30 minutes) for rivaroxaban/apixaban assay; however, the majority are not able to provide rapid TAT because of the time needed to prepare and run calibrators and controls before testing patient plasma. These assays are useful in nonemergent cases where either failure of therapy because of noncompliance or other reasons is suspected or when critical location surgery is planned (eg, neurosurgery, spinal surgery, etc). However, for rapid assessment of the presence of an FXaI in a bleeding patient or for emergent surgery, a routine heparin anti-Xa assay may be useful to guide use of either a specific reversal agent (andexanet) or a nonspecific hemostatic agent (prothrombin complex concentrate).

The principle of anti-Xa assay for both heparin assay and specific FXaI assay is identical. Currently, the majority of heparin assays use a single curve to measure both unfractionated and low-molecular weight heparins using the same calibrators and controls. The results are given in international units per milliliter, with a comment stating that the therapeutic ranges for unfractionated heparin and low-molecular weight heparin are 0.3 to 0.7 and 0.5 to 1.0 IU/mL, respectively. In this assay, heparin (unfractionated or low molecular weight) in patient’s plasma binds to endogenous antithrombin to form heparin + antithrombin complexes that neutralize a known amount of FXa present in the reagent. The remaining free FXa then acts on a chromogenic substrate to release p-nitroaniline to give a color, which is read photometrically against a calibrated curve. Rivaroxaban/apixaban assays use the same principle, where FXaI directly neutralizes FXa present in the reagent without requiring antithrombin and where the amount of p-nitroaniline released is measured and compared against rivaroxaban/apixaban calibrators.

There seems to be a good correlation between anti-Xa activity measured by heparin assay as international units per milliliter and rivaroxaban levels measured as nanograms per milliliter with rivaroxaban calibrators in 1 assay,⁹  whereas there was a poor correlation in other assays, perhaps owing to the different plasma dilutions used.¹⁰  Absence or very minimal anti-Xa activity rules out clinically relevant amounts of FXaI and therefore, would suggest against a need to use a reversal/hemostatic agent for a bleeding patient (<0.5 IU/mL) or before emergent surgery (< 0.3 IU/mL). This notion is probably supported by the Annexa-4 full study report, where patients on FXaI with levels <75 ng/mL were excluded from the hemostatic efficacy evaluation, because levels <75 ng/mL were not considered to be clinically sufficient for reversal.⁸  Of the 352 patients who received andexanet alfa, 254 patients (72%) had anti-FXa activity of >75 ng/mL, meaning that 98 (28%) patients did not have sufficient drug levels and hence, did not need the reversal agent. With the lower-dose andexanet alfa priced at ∼$26 000 per dose, for every 10 patients presenting with major bleeds because of FXaI, 3 patients potentially would not need the agent (ie, $78 000 savings). Andexanet alfa given to a patient with underlying prothrombotic condition in the absence of a clinically relevant amount of FXaI may have a higher potential to cause thrombotic complications because of its tissue factor pathway inhibitor inhibitory effect.¹¹  Therefore, based on the full study report, it may seem obligatory to have access to rapid assessment of FXaI levels >75ng/mL for appropriate use of an expensive drug that may have potential prothrombotic effects as indicated by the FDA’s black box warning.¹² 

Measuring anti-Xa activity by heparin assay is also supported by edoxaban measurement in the ENGAGE-AF-TIMI-48 study (Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation–Thrombolysis in Myocardial Infarction 48).¹³  The trough edoxaban levels were measured in 6780 of the 14c069 patients by the LCMS method as nanograms per milliliter, and they were compared with anti-Xa activity measured by heparin assay (Rotachrome; Stago) expressed as international units per milliliter, with a lower limit of quantitation of 0.1 IU/mL. The study showed an excellent correlation between edoxaban concentration by LCMS and anti-Xa activity by heparin assay (r = 0.96; 95% confidence interval, 0.95-0.96; P < .0001). Table 2 summarizes laboratory assessment of FXaIs.

In our patient, STAT PT/INR, aPTT, heparin anti-Xa activity, and thrombin time were performed. The PT/INR was confirmed to be prolonged at 19 and 1.9, respectively. The aPTT was 32.5 seconds (reference interval, 23-33.5 seconds), and thrombin time was 17 seconds (reference interval, 16-25 seconds), ruling out the presence of dabigatran. Her anti-Xa activity was 0.7 IU/mL, suggesting that there was a low level of an FXaI. Her Extem clotting time was 75 seconds (normal, <82 seconds) in the rotational thromboelastometry study simultaneously performed and showed no effect of FXaI. Because of her adequate renal function, she was successfully managed with supportive care.⁶  Currently, there is insufficient information on the effect of DOACs on viscoelastic measurements of clot by thromboelastography or rotational thromboelastometry, although studies show dose-dependent prolongation of clotting times.^14-16 

There are insufficient data supporting routine monitoring of DOAC levels to predict either thrombotic or bleeding risk in a given patient. There are a few studies that tried to use laboratory assessment for these reasons; however, DOAC measurements were performed only once or twice, and they were often remote from the clinical event to correlate with drug levels.^1,17-19  Because of short half-lives of DOACs, it is difficult to rely on these laboratory values, because skipping even 1 dose alters the levels greatly.

With increasing need for laboratory assessment of DOACs, especially FXaI, it is imperative to develop a universal assay that can measure direct anti-Xa activity of all FXaI (and even heparins) using a single-calibration curve as proposed by van Pelt et al.²⁰  The principle of this assay would be to detect the rate of direct FXa inhibition by anticoagulants and express results in a uniform way. Preferably, this assay should use routinely available heparin assay platforms for appropriate management of patients with major bleeding or patients requiring urgent surgery. Thrombin generation assay is another option that needs rapid validation for clinical use, because this assay could be useful in assessing thrombin generation potential of anticoagulated patients for specific therapy.

Routine coagulation tests of PT and aPTT are not very useful to assess DOAC effect because of high variability in their sensitivity. Thrombin time is very sensitive to dabigatran, and when normal, it suggests a clinically insignificant amount of dabigatran in plasma. Specific DOAC-calibrated non-FDA approved assays are available; however, not many hospitals can provide rapid TAT (<30 minutes). Routinely used, some heparin assays may provide sufficient information on FXaI activity to make some clinical decisions. Ultimately, user-friendly rapid assays are urgently needed to fulfill clinical need of complex patients on DOACs.

Ravi Sarode, Department of Pathology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390; e-mail: ravi.sarode@utsouthwestern.edu.