Ciraparantag, an anticoagulant reversal drug: mechanism of action, pharmacokinetics, and reversal of anticoagulants

Since the discovery of dicoumarol in 1940, the vitamin K antagonists, primarily warfarin, have been the only oral anticoagulants available. In 2010, the first of a new class of oral anticoagulants was introduced and approved by the US Food and Drug Administration: an oral direct thrombin or factor IIa (FIIa) inhibitor (eg, dabigatran), followed shortly thereafter by 4 oral direct factor Xa (FXa) inhibitors (eg, apixaban, betrixaban, edoxaban, and rivaroxaban). These new direct oral anticoagulants (DOACs; also referred to as non–vitamin K oral anticoagulants or NOACs) are as effective as warfarin, and evidence strongly suggests they are safer.^1,2  Even though major bleeding rates are lower, there is still a significant risk, up to 4%, of major bleeding, and an even higher rate of clinically relevant nonmajor bleeding.²  Between 2013 and 2014, rivaroxaban and dabigatran were the 5th and 10th most common drugs, respectively, to cause emergency room visits for adverse drug events in older adults.³  Hospitalization rates for DOAC-related adverse drug events were also similar to that of warfarin. These drugs were developed without a specific reversal agent to treat active bleeding or to reverse anticoagulation in the face of urgent surgery, and the absence of a reversal agent is a limitation for most currently used antithrombotic agents. This limitation is now partially remedied with a humanized murine monoclonal antibody Fab fragment (idarucizumab; Praxbind; Boehringer Ingelheim) targeted to bind to and reverse dabigatran⁴  and a second agent, andexanet alfa (Andexxa; Portola Pharmaceuticals Inc), a modified, enzymatically inactive recombinant FXa functioning as a decoy FXa targeted to bind to and reverse the oral FXa inhibitors as well as enoxaparin.⁵  Andexanet alfa was recently approved by the US Food and Drug Administration, but has had challenges to its widespread use.⁶  A third reversal agent, ciraparantag, with several favorable pharmacokinetic (PK) and pharmacodynamic (PD) characteristics, is in development and is currently emerging from phase 2 clinical trials.

This report summarizes the mechanism of action of ciraparantag, its PK and PD characteristics, and its early proof-of-concept ability to reverse anticoagulant-induced bleeding in animal models. This information has been presented in abstract form, but has not been published in the peer-reviewed literature.

Standard dynamic light scattering (DLS) methodology⁷  was used as a screening tool to determine the binding to enoxaparin and the DOACs in aqueous solution, other drugs commonly used in patients receiving anticoagulant therapy (eg, cardiovascular drugs) and selected coagulation factors (eg, FXa). DLS is a physical technique used to characterize the effective size of particles in suspension via quasielastic light scattering. When 2 molecules combine, the hydrodynamic radius increases, reflecting an association between the entities. The use of DLS methodology was necessary to characterize the binding of ciraparantag because nonaqueous solvents disrupt physical interactions, the DOACs and ciraparantag are comparable in size, and the DOACs are poorly water soluble, rendering measurements by equilibrium dialysis, mass spectrometry, and isothermal titration calorimetry unsuitable.

The PK and metabolism of ciraparantag were explored in both animal and human experiments.

In the first experiment, quantitative whole-body autoradiography (QWBA) was used to determine the PK, rates, and routes of excretion, with mass balance and tissue distribution (determined by QWBA of radioactivity), after a single IV administration of [¹⁴C]ciraparantag. Male Sprague-Dawley rats (N = 19) received a single dose of 10 mg/kg radio-labeled ¹⁴C-ciraparantag (200 μCi/kg; 1.8 mg/kg ¹⁴C-ciraparantag plus 8.2 mg/kg ciraparantag). Samples for radioanalysis were collected at designated time points up to 96 hours after administration and included whole blood and serum (with and without carbon dioxide inhalation), excreta (urine and feces), cage residue, and expired radiolabeled carbon dioxide (using a metabolism cage and bubbler flask). At the terminal time point of 96 hours after the dose, some animals were euthanized for a residual radioanalysis of the carcass. In addition, at each time point of 1, 8, 12, 24, and 72 hours after the dose, some animals were euthanized, and carcasses were prepared and examined by QWBA analysis.

In the second experiment, part of a phase 1 first in-human study (www.clinicaltrials.gov #NCT01826266),^8,9  the PK of ciraparantag was studied in adult volunteers (N = 70) aged 18 to 45 years. Ten subjects were enrolled in each of 7 dose cohorts (5, 15, 25, 50, 100, 200, and 300 mg ciraparantag) and randomized in an 8:2 ratio to receive ciraparantag or placebo (saline), respectively. The 25-mg cohort was repeated because of changes in coagulation testing during the experiment. Each dose cohort consisted of 2 time periods, an infusion of a single IV dose of ciraparantag or placebo (period 1) and the administration of a single 60-mg oral dose of edoxaban, followed by the infusion of either ciraparantag or placebo (period 2). Serial plasma measurements were collected for PK assessment of ciraparantag and its primary metabolite, 1,4-bis(3-aminopropyl) piperazine (BAP), in both time periods (0, 5, 10, 20, 30, 45, 60, and 90 minutes and 2, 4, 6, 12, and 24 hours after the dose). Urine samples were also assessed for ciraparantag at regular intervals relative to study drug administration (∼4 samples per subject per period). Details of the study design have been published.⁹  Serum and urine PK parameters for ciraparantag and BAP were computed from drug-concentration time data, using noncompartmental approaches (WinNonLin Phoenix, version 6.1; Pharsight Corporation, STATA 12; College Station, TX). PK parameters included maximum time (T_max) to maximum plasma concentration (C_max), half-life (t_1/2), area under the concentration curve (AUC), plasma clearance (CL), volume of distribution, amount of drug excreted in urine (Ae), fraction of metabolite excreted in urine (Fm), and renal clearance (CLr).

A series of experiments were conducted to investigate the ability of ciraparantag and/or BAP to reverse the effects of various anticoagulants (dabigatran, apixaban, edoxaban, rivaroxaban, unfractionated heparin [UFH], and enoxaparin) in in vivo rat models. In these preclinical experiments, we thought it important to focus on actual bleeding (the most important outcome) rather than correction of a coagulation assay. Ciraparantag is cationic, and it binds to the anionic substances in standard blood collection tubes used for coagulation testing (eg, sodium citrate, EDTA, oxalate, and heparin) and to the activators used in the traditional coagulation assays, such as activated partial thromboplastin time (aPTT; eg, kaolin and Celite), making plasma-based assays, because of reagent interference and insensitivity, unsuitable for measurement of ciraparantag’s effect. In subsequent clinical experiments, we used the whole-blood clotting time (WBCT), which measures clotting in reagent-free collection equipment as the key clinical PD measurement for ciraparantag.

In rat tail transection models¹⁰  (supplemental Figure 1, available on the Blood Web site), ciraparantag/BAP or saline was administered at the reported T_max of the anticoagulant, followed 20 minutes later by full transection of the rats’ tails. The volume of shed blood and the blood pellet weight, used to determine the amount of blood loss, were assessed and analyzed post hoc by 1-way analysis of variance.

Approval for animal experiments was granted by the Brown University Institutional Care and Use Committee (Providence, RI), and the MPI Research Institutional Care and Use Committee (now part of Charles River Laboratories). Approval for PK experiments was granted by the Copernicus Group Independent Review Board (Durham, NC). The study was conducted in accordance with the Declaration of Helsinki.

Ciraparantag arose from a program of intentional molecular design that searched for molecules that would bind to free UFH through noncovalent, charge-charge interaction. Multiple structures were hypothesized, and each was tested on an energy minimization computer model to predict noncovalent binding to fondaparinux, enoxaparin, and UFH. Several of these candidate structures were selected and synthesized, and one of the entities, ciraparantag (Figure 1), was found to directly bind the most strongly to these heparins. Subsequently and unexpectedly, energy minimization modeling similarly predicted binding to dabigatran, apixaban, edoxaban, and rivaroxaban.

Strong ionic bonding between ciraparantag and heparin was confirmed, using a commercially available heparin affinity chromatography column. Results from this experiment showed that ciraparantag was eluted from the heparin-coated column at a later time point compared with the acetate control (7.10 vs 2.47 minutes, respectively) when at 0.55 M sodium chloride (Figure 2). An increase in sodium chloride concentration from 0.55 to 0.75 M led to a gradual decrease in the elution time of ciraparantag to 4.39 minutes at 0.75 M. These data confirm that ciraparantag demonstrated a strong ionic interaction with heparin and that the affinity between ciraparantag and heparin is largely due to electrostatic interactions. Ciraparantag also showed higher affinity to heparins than did the ciraparantag metabolites monoarginine piperazine, BAP, and 3-(4-(3-aminopropyl)piperazin-1-yl)propanoic acid (carboxylated BAP) (data not shown).

DLS experiments that assessed binding of ciraparantag to anticoagulants in aqueous solution, as illustrated in Figure 3 for the DOACs, indicated that physical, noncovalent binding occurred between ciraparantag and all of the anticoagulants evaluated (additional data in supplemental Figure 2).

When binding to other drugs and blood proteins was assessed by DLS methodology, ciraparantag demonstrated no binding to FIIa, FXa (Figure 4A-B), or cardiac drugs (diltiazem, digoxin) (Figure 4C-D). Similar data were obtained for other cardiac drugs (propranolol, lidocaine, metoprolol, lisinopril, propafenone, hydrochlorothiazide, triamterene, and clopidogrel), antiepileptic drugs (carbamazepine, gabapentin, lamotrigine, phenytoin, and valproate), diabetic drugs (insulin and metformin), or other commonly used drugs (aspirin, atorvastatin, azithromycin, and streptokinase). Illustrative graphs are shown in supplemental Figure 3. In addition, ciraparantag did not bind to plasma proteins in human plasma compared with compounds with known moderate-to-high plasma protein binding (propranolol and warfarin, respectively; supplemental Table 1).

In the QWBA experiment, the T_max (peak of radioactivity) of ciraparantag in whole blood was observed at 5 minutes after the dose, the earliest time point sampled. The plasma C_max was followed by a steep decline through 4 hours after the dose (distribution phase), and a more gradual decline thereafter (elimination phase; supplement Figure 4A). At 72 hours after the dose, whole-blood-to-serum ratios for QWBA were calculated and ranged from a low of 0.49 at 1 hour after the dose, to a gradually increasing high of 2.53 at 72 hours after the dose. A blood-to-serum value <1.0, as observed at 1 hour after the dose, was indicative of low partitioning to erythrocytes, whereas a value >1.0, as observed at time points >1 hour, was indicative of high partitioning to erythrocytes.

Drug-derived radioactivity was widely distributed at 1 hour after the dose, the first QWBA time point analyzed. Tissue concentrations of radioactivity were highest in the kidney, urinary bladder, and urine. The main route of elimination of ¹⁴C-ciraparantag was urinary excretion (supplement, Figure 4B). Most of the total dose (∼93%) was recovered within the first 24 hours and ∼82.44% was recovered within the first 8 hours. Elimination via other routes was minimal. Twenty-five tissues were positive for radioactivity at 1 hour after the dose. At 72 hours after the dose, the last QWBA time point analyzed, 23 of the 25 tissues still had detectable (though significantly reduced) radioactivity, as a result of metabolites retaining the radioactive label.

A total of 70 subjects providing 1818 samples (13 samples per subject per period) were included in the serum PK analysis. In the lowest dose cohort (5 mg), PK data were insufficient to estimate all ciraparantag PK parameters. In the other dose cohorts (15-300 mg), ciraparantag demonstrated dose-proportional PKs. Ciraparantag PK parameters, stratified by dose cohort (Table 1) and mean serum ciraparantag concentrations vs time across the dose cohorts (Figure 5A), are provided. The C_max for ciraparantag increased from 173 ng/mL at the 5-mg dose to 10 570 ng/mL at the 300-mg dose. Dose-proportional PK was further observed with AUC over 24 hours (AUC_0-24), which increased from an average of 151 to almost 3800 ng/mL per hour with 15- and 300-mg doses, respectively. Maximum serum ciraparantag concentrations occurred within 5 to 9 minutes (T_max), then rapidly declined, and were below the limit of quantification by 2 hours after drug administration. The t_1/2 of ciraparantag ranged from 12 to 19 minutes across cohorts. The median volume and total-body CL estimates ranged from 25.4 to 34.6 L/h and 65.7 to 116.0 L/h, respectively. The PK of ciraparantag after edoxaban administration (period 2) for T_max and other parameters (full data not shown) were similar to those of ciraparantag alone (Figure 5B).

BAP serum exposure (AUC_0-24 and AUC_0-inf) could not be estimated for the 2 cohorts with the lowest doses (5 mg and 15 mg) because of insufficient PK data. For the 25- to 300-mg dose cohorts, BAP AUC_0-24 ranged from 29% to 52% of total ciraparantag exposure (AUC_0-24; without molecular weight adjustment; Figure 5C) supporting rapid hydrolytic metabolism of ciraparantag via circulating peptidases.

The excretion of ciraparantag (Ae_0-24) in urine was unchanged. Main urine BAP PK results (Table 1; period 1, ciraparantag alone) showed that the median cumulative amount of BAP (A_e0-24) increased from 5.26 μg (5-mg dose) to 43 654.3 μg (300-mg dose). Up to 20% (F_m0-24) of the ciraparantag dose was recovered in urine as BAP. Median BAP CLr across the 25- to 300-mg doses, where data were available, ranged from 1.2 to 29.7 L/h.

Ciraparantag produced a dose-dependent reversal of edoxaban anticoagulation (10 mg/kg administered orally), as measured by blood loss, when ciraparantag was administered after edoxaban but before tail transection. After edoxaban anticoagulation, the ciraparantag doses of 1.25 and 2.5 mg/kg IV had no effect in reducing blood loss. Both the 5- and 10-mg/kg ciraparantag doses reduced blood loss volumes to levels similar to those of rats that did not receive edoxaban (Figure 6A). Figure 6B illustrates the blood loss at specific time intervals, measured individually. Reduced blood loss was also achieved when 30 mg/kg ciraparantag was given after the rat tail transection (ie, during active bleeding), which is different from other experiments in which ciraparantag was administered before the tail was transected (Figure 6C).

BAP administration, at doses equivalent to large doses of ciraparantag, also reversed the anticoagulant effects of 10 mg/kg oral edoxaban (data not shown). BAP (40 mg/kg IV) led to a significant reduction in blood loss that was equivalent to that of untreated animals (no edoxaban). This dose of BAP is equivalent to 170 mg/kg ciraparantag on a dry-powder-weight basis, assuming complete and immediate conversion of ciraparantag to BAP. Because the earlier experiment showed that ciraparantag 5 mg/kg IV fully reversed the anticoagulant effect of 10 mg/kg oral edoxaban, it is estimated that BAP has one-thirty-fourth of the potency of ciraparantag on a molar basis with respect to edoxaban reversal.

In both the rat tail transection model and a rat tail surface laceration model, ciraparantag reversed the anticoagulant effects of dabigatran, as measured by blood volume loss and bleeding time, respectively. In the tail transection model (Figure 6D), ciraparantag (31.25 mg/kg IV) reversed the anticoagulant effects of 37.5 mg/kg oral dabigatran, as measured by a reduction in the volume of blood loss similar to the levels observed in control animals. In the surface laceration model, ciraparantag (15 mg/kg IV) reversed the effects of IV dabigatran (1 mg/kg) to levels similar to those of control animals, as shown by reduced bleeding times (data not shown).

Rats receiving 3.125 mg/kg oral apixaban had a more than eightfold increase in blood loss compared with the control rats dosed with saline (Figure 6E). The administration of 12.5 mg/kg IV ciraparantag significantly reduced blood loss to the baseline equivalent, compared with no ciraparantag, indicating a full reversal of apixaban activity. Full reversal was also observed at a higher dose of ciraparantag.

Rats dosed with 5 mg/kg oral rivaroxaban showed more than a twofold increase in blood loss compared with the control rats dosed with saline (Figure 6F). The administration of ciraparantag at 6.25 mg/kg IV led to a significant reduction in blood loss compared with that of the saline control. When the dose of ciraparantag was increased to 31.25 mg/kg, the blood loss was reduced to baseline levels, which indicates that ciraparantag can fully reverse the anticoagulant effect of rivaroxaban at this dose.

Ciraparantag was compared with protamine sulfate for its ability to reverse the anticoagulant effects of UFH (Figure 6G). After a dose of 1 mg/kg IV UFH, administration of 20 mg/kg IV ciraparantag significantly reduced blood loss, whereas 10 mg/kg IV protamine sulfate did not lead to a significant change in blood loss, compared with saline controls. Interestingly, protamine significantly reduced the effective UFH concentration to baseline (based on an anti-FXa activity assay) and significantly reduced aPTT to baseline, whereas ciraparantag did not significantly decrease aPTT levels. Thus, these coagulation assays (anti-FXa activity assay and aPTT) suggest restoration of coagulation after protamine, but not after ciraparantag, despite the clinical evidence showing that protamine did not, but ciraparantag did, reduce blood loss. These and other data indicate that some assays may not be appropriate for monitoring the effectiveness of ciraparantag in this species with this model. The effective reversal dose of ciraparantag increased from 10 to 20 mg/kg when the dose of UFH was increased from 1 to 2.5 mg/kg IV, demonstrating dose proportionality.

Ciraparantag did not interact with lidocaine when administered together to anesthetized rats. In the presence of 1 mg/kg lidocaine, 20 mg/kg IV ciraparantag fully reversed the anticoagulant effects of 1 mg/kg IV UFH and reduced blood loss to levels similar to the control.

Ciraparantag was compared with protamine sulfate for its ability to reverse the anticoagulant effects of enoxaparin. Ciraparantag (30 mg/kg IV) fully reversed the anticoagulant activity of enoxaparin (10 mg/kg IV), as shown by the restoration of blood loss compared with that of untreated animals (Figure 6H). In comparison, protamine sulfate, at the recommended dose of 10 mg/kg IV, did not lead to a significant change in blood loss after enoxaparin, when compared with that in saline-treated animals. This experiment showed that aPTT returned to control values after protamine (despite having no apparent effect on increased bleeding), whereas aPTT was not altered by ciraparantag (despite reversing the increase in bleeding after enoxaparin).

The ability of ciraparantag to reverse edoxaban anticoagulation was assessed by measuring bleeding time after a standardized laceration of the median liver lobe in rats. After obtaining a baseline bleeding time, 1 mg/kg IV edoxaban was administered and then, 35 minutes later, 0 (control), 10, 20, or 30 mg/kg IV ciraparantag was administered. Ten minutes after ciraparantag (45 minutes after edoxaban) was administered, a second laceration was made, and bleeding time was evaluated again. The average percentage change from baseline in bleeding time was 166% in the saline control group vs 66%, 20%, and 14%, respectively, for the 10, 20, and 30 mg/kg ciraparantag groups (Figure 7A). The bleeding times of both the 20- and 30-mg/kg doses of ciraparantag were significantly reduced compared with edoxaban-treated control rats, indicating that ciraparantag reversed the anticoagulant effect of edoxaban, whereas the 10-mg/kg dose of ciraparantag yielded a similar trend but was not statistically significant. In another experiment, WBCTs were also performed, and the blood pellets from these WBCTs were examined by scanning electron microscopy. Figure 7B illustrates the progressive reestablishment of fibrin strands at 60 minutes, with increasing doses of ciraparantag approximating the baseline fibrin structure.

Ciraparantag is a small, synthetic, water-soluble, cationic molecule that is designed to bind specifically to UFH and LMWHs. DLS was used to study the association of ciraparantag with various DOACs and control molecules in aqueous solution at physiological pH. The data support the hypothesis that ciraparantag directly binds to DOACs and LMWH through noncovalent hydrogen bonds and charge-charge interactions. By doing so, ciraparantag prevents the DOACs from associating with the relevant coagulation factors (eg, FXa), thereby allowing for normal coagulation factor activity to be restored. The binding specificity of ciraparantag to the new oral FXa and FIIa anticoagulants, as well as older anticoagulants such as UFH and LMWH (enoxaparin), is supported by several different types of data. These include (1) intentional molecular design: ciraparantag was specifically designed to bind to heparins with hydrogen bonding and charge interactions (not by covalent bonding). This design effect was demonstrated by specific binding to a heparin affinity chromatography column. (2) DLS: data show concentration-dependent increases in the apparent size of ciraparantag, UFH, and enoxaparin. Subsequent to this finding, ciraparantag was found to bind to DOACs through noncovalent bonding. These data show evidence of a physical association between ciraparantag and FXa/FIIa-inhibiting anticoagulants through ionic and/or hydrophobic, noncovalent bonding. (3) Reversal of anticoagulant activity: experiments in animal models of bleeding demonstrated ciraparantag’s ability to block the anticoagulant activity of UFH and enoxaparin, in addition to reversing the anticoagulant effects of dabigatran, apixaban, edoxaban, and rivaroxaban and reducing bleeding in animal models. (4) Protein binding: evaluations of ciraparantag in human plasma indicated that there is no off-target protein binding, such as to FXa or FIIa or to other plasma proteins studied. The correlation between a lack of a physical association as measured by DLS and in vivo binding further serves to validate DLS as a screening tool for binding specificity. Taken together, this information supports the purposeful design of ciraparantag as a potential antidote for UFH, enoxaparin, and specific oral direct FXa/FIIa inhibitors, as opposed to wider nonspecific binding of proteins or other, nonheparin or non-FXa/FIIa anticoagulants and/or drugs. To date, off-target binding with the drugs tested has not been seen, but we cannot exclude the possibility of such binding if a more extensive list were tested.

PK experiments indicate that ciraparantag fits a 1-compartment model with a rapid distribution phase (in some cases, complete by the time of first blood sampling at 3 minutes), does not accumulate, and has minimal plasma protein binding. Ciraparantag undergoes hydrolysis of its peptide bonds after administration to form a primary metabolite (BAP) and a second, less relevant metabolite, monoarginine piperazine. In sum, ciraparantag rapidly achieves peak plasma concentrations, then undergoes rapid widespread distribution throughout the body. Some ciraparantag (∼30%) is metabolized via hydrolysis to BAP by circulating plasma peptidases, both of which are primarily (>90%) renally eliminated and are below measurable levels by 2 hours after a dose.

In various animal models, ciraparantag reversed bleeding definitively with each of the oral FXa inhibitors (apixaban, edoxaban, and rivaroxaban), the oral FIIa inhibitor (dabigatran), and the parenteral IIa/Xa inhibitors (enoxaparin and UFH). This effect was illustrated when the anticoagulant and ciraparantag were given before the bleeding injury, as well as when ciraparantag was given after the bleeding injury (with edoxaban given initially). As noted in these experiments, ciraparantag reversed bleeding in a dose-dependent manner with a higher dose required for some DOACs. This difference was also seen in a recently reported phase 2 trial¹¹  in volunteer subjects where a higher dose was needed to fully reverse anticoagulation with rivaroxaban than with apixaban. These findings will be incorporated in dose selection for future trials. In recently published human trials, ciraparantag was also shown to reverse anticoagulation in healthy volunteers treated with edoxaban,⁸  as well as with enoxaparin.¹²  In these clinical trials reversal of anticoagulant activity was measured with a manual WBCT. As noted, plasma-based coagulation assays are unsuitable because of binding of ciraparantag to reagents used for anticoagulation or activation of blood in vitro (ie, in the test tube). Currently, an automated point-of-care coagulometer that replicates the WBCT is being developed that provides sensitive, accurate, and precise measurements of whole-blood clotting.^13,14  In these human trials, there was no evidence of procoagulant activity with ciraparantag, as measured by levels of D-dimer, prothrombin fragment 1.2, and tissue factor pathway inhibitor levels through 24 hours after the dose.

In summary, ciraparantag has several unique attributes including its broad-spectrum activity in reversing oral and parenteral Xa inhibitors and oral IIa inhibitors; its ability to be rapidly and easily administered; and its long, functional PD half-life. The preclinical trials reported herein, as well as initial clinical trials,^8,9,11,12  have supported these beneficial characteristics. Additional phase 2 trials are ongoing to assess the ability of ciraparantag to reverse anticoagulation with other DOACs in healthy volunteers before phase 3 clinical trials are initiated in patients with major bleeding or in need of urgent surgery.

Paula Moon-Massat, who is an employee of AMAG Pharmaceuticals Inc, provided editorial assistance to the authors during preparation of the manuscript.

This work was supported in its entirety by Perosphere Pharmaceuticals Inc (Danbury, CT).

Institutions where work was performed: Perosphere Inc, Bedford, NY (DLS, heparin chromatography, and in vivo rat experiments); MPI Research Inc, Mattawan, MI (quantitative whole-body autoradiography); and Duke Clinical Research Unit, Durham, NC (ciraparantag PK study in humans).

Contribution: X.J., L.C., and C.B. performed the experiments; S.S., B.E.L., S.H.B., X.J., L.C., C.B., J.A., and A.B. analyzed the results; S.S., B.E.L., S.H.B., X.J., L.C., and C.B. designed the experiments; J.A., B.E.L., S.H.B., A.B., S.V., and S.S. wrote and edited the manuscript; and all authors reviewed the manuscript and made critical revisions to the intellectual content, approved the final version of the manuscript to be submitted to the journal, and agreed to be accountable for all aspects of the work.

Conflict-of-interest disclosure: S.H.B., X.J., and L.C. are employees of Perosphere Technologies Inc, which is developing a point-of-care device for measurement of WBCT. S.S. sits on the Board of Directors of Perosphere Technologies Inc. J.A. is a paid advisor to and S.V. is a paid consultant for AMAG Pharmaceuticals Inc, which is developing ciraparantag as a reversal agent for DOACs. The remaining authors declare no competing financial interests.

Correspondence: Jack Ansell, 15 Waterview Way, Long Branch, NJ 07740; e-mail: ansellje@gmail.com.