Determination of edoxaban equivalent concentrations in human plasma by an automated anti-factor Xa chromogenic assay

Thrombosis Research 155 (2017) 121–127

Contents lists available at ScienceDirect

Thrombosis Research journal homepage: www.elsevier.com/locate/thromres

Full Length Article

Determination of edoxaban equivalent concentrations in human plasma by an automated anti-factor Xa chromogenic assay Ling He a,⁎, Jarema Kochan a, Min Lin a, Alexander Vandell a, Karen Brown b, Francois Depasse c a b c

Daiichi Sankyo Pharma Development, Edison, NJ, USA Formerly of Daiichi Sankyo Pharma Development, Edison, NJ, USA Diagnostica Stago, Asnières sur Seine, France

a r t i c l e

i n f o

Article history: Received 27 January 2017 Received in revised form 28 April 2017 Accepted 6 May 2017 Available online 07 May 2017 Keywords: Edoxaban Anti-FXa assay Coagulation assay Laboratory monitoring Direct oral anticoagulants Pharmacokinetics

a b s t r a c t Introduction: This phase I, open-label, multiple-dose, two-treatment study assessed the relationship between edoxaban equivalent concentration derived from an anti-FXa assay with the summed concentration of edoxaban and its active metabolite, M-4, as assessed by liquid chromatography coupled with tandem mass spectrometry (LC/MS/MS). This study also assessed the relationship between edoxaban plasma concentrations assessed by LC/MS/MS in sodium citrate and lithium heparin tubes. Materials and methods: Healthy volunteers were randomized to receive once-daily edoxaban 60 mg or 90 mg for 5 days (15 participants per treatment group). Serial blood samples were collected for analysis by LC/MS/MS and by the anti-FXa assay. Edoxaban equivalent levels were assessed using a commercially available anti-FXa activity assay with an edoxaban-specific setup. Results and conclusions: The day 5 concentration estimates were significantly correlated between the 2 assays (P b 0.0001 for both edoxaban doses). The geometric least squares mean (GLSM) ratio (90% confidence interval) for edoxaban equivalent concentrations vs edoxaban + M-4 concentrations was 114.3% (108.2–120.8) for edoxaban 60 mg (P b 0.0001) and 113.0% (107.1–119.2) for edoxaban 90 mg (P = 0.0002). The GLSM ratio for edoxaban concentrations in sodium citrate vs lithium heparin tubes for 60-mg and 90-mg edoxaban doses were 82.8% (78.5–87.3) and 83.9% (79.1–89.0), respectively. In this study, an anti-FXa chromogenic assay with edoxaban-specific calibrators and controls demonstrated good accuracy in estimating edoxaban concentrations across a wide range of concentrations relative to LC/MS/MS at steady state following the administration of once-daily edoxaban for 5 days. © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction Edoxaban is a direct, oral factor Xa (FXa) inhibitor with linear and predictable pharmacokinetics (PK) [1], indicated for the prevention of stroke in patients with nonvalvular atrial fibrillation (NVAF) and for the treatment of venous thromboembolism (VTE) [2]. Large phase III studies have demonstrated that direct oral anticoagulants (DOACs), including edoxaban, are at least as efficacious as warfarin and are associated with less major bleeding in patients with NVAF and VTE [3–10]. Unlike warfarin, DOACs typically do not require routine laboratory Abbreviations: AE, Adverse events; aPTT, activated partial thromboplastin time; ANOVA, Analysis of variance; CI, Confidence interval; DOACs, Direct oral anticoagulants; FDA, Food and Drug Administration; FXa, Factor Xa; GLSM, Geometric least squares means; LC/MS/MS, Liquid chromatography coupled with tandem mass spectrometry; LLOD, Lower limit of detection; PK, Pharmacokinetic; PT, prothrombin time; NVAF, Nonvalvular atrial fibrillation; VTE, Venous thromboembolism. ⁎ Corresponding author at: Senior Director, Clinical Bioanalysis, Daiichi Sankyo Pharma Development, 399 Thornall St, Edison, NJ 08837, USA. E-mail address: [email protected] (L. He).

monitoring [11]. Nevertheless, there are situations in which clinicians may want to be informed of DOAC drug concentrations, including before surgery or invasive procedures, when a patient is actively bleeding, following a DOAC overdose, on reoccurrence of thrombosis to determine if the drug is present in a therapeutic range, or if the patient develops renal failure [11,12]. Quantifying DOAC levels with liquid chromatography coupled with tandem mass spectrometry (LC/MS/MS) assays is specific, sensitive, and accurate [13,14]. However, LC/MS/MS assays are not readily available in routine clinical practice [15]. Clotting assays, such as prothrombin time (PT) and activated partial thromboplastin time (aPTT), are more widely available but are not quantitative and direct, oral FXa inhibitors have variable effects on these standard coagulation tests [11, 16–18]. Recent studies have suggested that some PT reagents could be useful to measure edoxaban, however, the impact of edoxaban on PT and aPTT depends on the reagents [19,20]. In a public workshop held by the US Food and Drug Administration (FDA) on in vitro diagnostic testing for DOACs, the FDA noted that appropriate assays should show accuracy and consistency at key medical

http://dx.doi.org/10.1016/j.thromres.2017.05.005 0049-3848/© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

122

L. He et al. / Thrombosis Research 155 (2017) 121–127

decision levels [21]. Chromogenic anti-FXa assays, which are currently in clinical use for measuring the levels of low molecular weight heparin [22], offer the possibility to adequately measure direct anti-FXa drugs, including edoxaban in plasma [23,24]. To that end, specific anti-FXa chromogenic assays, which measure the inhibition of the activity of exogenously added FXa, have recently been tailored for use with direct oral anti-FXa inhibitors [16,25]. In these tests, the residual activity from an excess of exogenously added FXa is inversely proportional to the amount of direct, oral FXa inhibitors and any relevant active metabolites present in the plasma sample. To date, there is no universal anti-FXa test for measuring all of the direct, oral FXa inhibitors because these assays need to be specifically calibrated for each drug [16]. When used with appropriate calibrators and controls, anti-FXa assays can accurately quantify plasma concentrations for rivaroxaban and apixaban over a wide concentration range with good intra- and inter-laboratory consistency [17,18,25–29]. For measuring edoxaban levels, several studies have assessed the performance of chromogenic assays designed to measure heparins or anti-FXa chromogenic assays with non-edoxaban–specific calibrators and controls [19,20]. However, the performances of anti-FXa chromogenic assays with edoxaban-specific calibrators and controls have not been reported. The aim of the present study was to assess the relationship between edoxaban equivalent concentration derived from an antiFXa assay, using edoxaban-specific calibrators and controls, with the summed concentration of edoxaban and its most abundant, active metabolite, M-4, as assessed by LC/MS/MS. The anti-FXa assay measures the activity of all active molecular moieties, therefore, in this context, edoxaban equivalent concentration refers to the level of edoxaban and all its active metabolites. The predominant active metabolite of edoxaban is M-4, and it reaches b10% of the exposure of edoxaban in healthy individuals [1,2,30]. Therefore, summing edoxaban and M-4 concentrations should closely approximate the edoxaban equivalent concentration. Blood samples for the anti-FXa assays are typically collected in sodium citrate tubes while blood samples for LC/MS/MS PK analysis in the clinical development of edoxaban were collected in lithium heparin tubes. Therefore, this study also assessed the relationship between plasma concentrations of edoxaban and M-4 as assessed by LC/MS/MS in 0.109 M sodium citrate and in lithium heparin tubes. The sodium citrate tubes contained pre-added buffer equal to 10% of the maximum blood collection volume, while the lithium heparin tubes had no buffer (spray-coated).

2.2. Study participants Healthy males and females were eligible to enroll if they were between 18 and 55 years of age with a body mass index between 18 and 30 kg/m2 at screening and check-in. Exclusion criteria included a history of any clinically significant disorder that might prevent the successful completion of the study or a clinically significant illness; stomach or intestinal surgery or resection that might alter absorption and/or excretion of orally administered drugs; a history of major bleeding, major trauma, or major surgical procedure of any type within 6 months of the first dose or a history of peptic ulcer, gastrointestinal bleeding, or dysfunctional uterine bleeding; a history of eye surgeries or trauma to the head or eye within 14 days of the first dose; a history or presence of an abnormal electrocardiogram or a Fridericia's corrected QT ≥ 450 msec for males and ≥470 msec for females at screening; and a previous edoxaban study within 6 months prior to the first dose. Female participants could not be pregnant or lactating. In addition, all participants who used an anticoagulant, coagulant, or antiplatelet medications within 30 days prior to the first dose or aspirin within 10 days prior to the first dose were excluded. 2.3. Sample collection Serial blood samples were collected for analysis by LC/MS/MS (in both lithium heparin and 0.109 M sodium citrate tubes) and by the anti-FXa assay (in 0.109 M sodium citrate tubes) on days 2, 3, and 4 prior to dosing (trough) and on day 5 predose and postdose (0.5, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 8, 10, 12, 14, 22, 24, 28, 32, 36, and 48 h). The lithium heparin sample tubes containing blood for plasma preparation were gently inverted 8 to 10 times to ensure thorough mixing of anticoagulants (spray-coated lithium heparin) and blood. Within 30 min of blood draw, the samples were centrifuged for approximately 10 min under 1500 ×g and 4 °C. The harvested plasma samples were stored in a −20 °C freezer. For blood samples collected in sodium citrate tubes, the volume of pre-existing 0.109 M sodium citrate buffer in the collection tube was equal to 10% of the specified total sample volume of the tube (eg, 0.3 mL citrate solution plus 2.7 mL of blood). The sample processing procedures were similar to those used for the lithium tubes, except that the centrifugation conditions were 2500 ×g at room temperature, and the resulting plasma were stored in a −70 °C or below freezer. The blood samples for the anti-FXa assay were collected and processed according to the Clinical and Laboratory Standards Institute guidelines [31]. 2.4. Assay methods

2. Material and methods 2.1. Study design This phase I, open-label, multiple-dose, two-treatment study was conducted in accordance with the International Conference on Harmonisation Harmonised Tripartite Guideline on Good Clinical Practice. The study was conducted in healthy adult male and female participants at one site in the US (Medpace Clinical Pharmacology Unit; Cincinnati, Ohio). All participants gave written informed consent prior to participating in the study. At check-in (day −1), participants were screened for inclusion/exclusion criteria, underwent a physical examination, provided blood and urine samples for clinical laboratory testing, and had vital signs checked. Participants remained in the unit through day 7. Participants were randomly assigned to receive once-daily edoxaban 60 mg or 90 mg on days 1 to 5. All treatments were administered with 240 mL of water in the morning following an overnight fast of 10 h. On days 1 to 4, participants continued to fast for 2 h after dosing; on day 5 participants continued to fast for 4 h postdose. Adverse events (AEs) were continuously monitored throughout.

2.4.1. Anti-FXa Edoxaban equivalent levels were assessed by Medpace Research Laboratory (Cincinnati, OH) using a commercially available anti-FXa activity assay (STA®-Liquid Anti-Xa; Diagnostica Stago, Asnières sur Seine, France) with an edoxaban-specific setup using the STA®Edoxaban Calibrator and STA®-Edoxaban Control on the STA®-R analyzer (Diagnostica Stago, Asnières sur Seine, France). The assay comprises 4 levels of calibrators and 2 levels of controls. The nominal concentrations for calibrators are 0 (blank) and approximately 30, 90, 140 ng/mL and for controls approximately 37 and 100 ng/mL, with lot-specific concentration values assigned based on independent LC/ MS/MS measurements. The edoxaban concentration levels of each lot of calibrators and controls were determined by high-performance liquid chromatography-mass spectrometry. All calibrator and control reagents were reconstituted in distilled water per the kit instructions. The assay was performed according to a detailed standardized protocol provided by the manufacturer. The lower limit of detection (LLOD) for this assay was 15 ng/mL. The upper limit of quantification for this assay was 450 ng/mL. The assay includes an automated redilution of plasma samples containing an edoxaban equivalent concentration higher than

L. He et al. / Thrombosis Research 155 (2017) 121–127

the value of the top calibrator (approximately 140 ng/mL). Although not required by the manufacturer, a daily calibration curve was constructed for the anti-FXa assay using the edoxaban calibrators; this allowed to check the consistency of results calculated against the original calibration curve vs the calibration curve of the day. The edoxaban controls were measured within each analytical run for acceptance of sample measurements.

2.4.2. LC/MS/MS Plasma samples with lithium heparin and sodium citrate as anticoagulants were collected and analyzed with validated LC/MS/MS assays that are specific for the respective matrix types. Plasma samples were stored at a nominal temperature of − 20 °C for lithium heparin plasma samples and − 80 °C for sodium citrate plasma samples, at which edoxaban and M-4 metabolite in the plasma matrix were demonstrated to be stable during multiple freeze/thaw cycles. Sample preparation was conducted in a 96-well microtiter plate using a Quadra 96 Sample Processor (Tomtec, Hamden, CT). The analytes and deuterium-labeled internal standards (d6-edoxaban and d3- M-4) were isolated from plasma sample (200 μL for lithium heparin plasma assay and 20 μL for sodium citrate plasma assay) using Oasis MCX™ (mixed-mode cation exchange resin) 96-well solid-phase extraction plate (Waters, Milford, MA). Extracted samples were subjected to gradient chromatography using 5 mM ammonium acetate (pH 7.0) as mobile phase A and methanol as mobile phase B, at a flow rate of 0.3 mL/min. The HPLC column used was Zorbax Eclipse XDB Phenyl (50 mm length, 2.1 mm internal diameter, 5 μm particle size) (Agilent, Wilmington, DE). The analytes and internal standards were detected and quantified using quadruple mass spectrometers (API 4000 for lithium heparin plasma assay, API 5000 for sodium citrate plasma assay) with TurboIonSpray source in the positive ion mode (SCIEX, Framingham, MA). The LLOD for edoxaban and M-4 were 0.764 ng/mL and 0.0792 ng/mL, respectively.

2.5. Planned sample size and statistical analysis Fifteen subjects were enrolled per treatment group, with the expectation that 12 subjects per treatment would complete the study. The sample size was not based on statistical considerations and was considered sufficient to achieve the study objectives. All concentration summaries and PK parameters were calculated using the PK analysis data set. The PK analysis data set consisted of all participants who received at least 1 dose of edoxaban and had a corresponding measurable edoxaban equivalent concentration by anti-FXa or edoxaban concentration by LC/MS/MS. For comparison to the antiFXa edoxaban equivalent concentration, the LC/MS/MS concentrations of edoxaban and its main, active metabolite, M-4, were summed. The average concentrations of edoxaban + M-4 across scheduled time points by dose (ie, 60-mg and 90-mg edoxaban) and collection method (eg, LC/MS/MS in sodium citrate, and LC/MS/MS in heparin) were analyzed using descriptive statistics. Similarly, the edoxaban equivalent concentrations assessed by anti-FXa assay were analyzed across time points by dose using descriptive statistics. Pearson's correlation coefficients were calculated by dose and scheduled time points, as well as overall for the comparison of each method. For comparison of the assays, an analysis of variance (ANOVA) was performed by dose and time point on the ln-transformed concentrations with method of collection as a fixed effect. Geometric least squares means (GLSM) and 90% confidence intervals (CI) from the models were back transformed to the original scale. The concentration estimates between the assays were considered statistically different when the 90% CIs did not cross 100. In addition, Bland-Altman plots were constructed to visualize the limits of agreement between the assays.

123

3. Results 3.1. Baseline demographics and clinical characteristics All randomized participants (N = 30) received ≥ 1 dose of study drug (edoxaban 60 mg, n = 15; edoxaban 90 mg, n = 15). The PK analysis data set included 28 participants (60-mg edoxaban, n = 13; 90-mg edoxaban, n = 15). One participant withdrew consent for personal reasons and one participant withdrew due to hematochezia. The majority of participants (70% [21/30]) were male, 50% (15/30) were White, and the mean ± standard deviation (SD) age was 39.4 ± 10.04. The demographics and baseline clinical characteristics were similar in both treatment groups (Table 1). 3.2. Anti-FXa assay The calibration curves for the anti-FXa assay showed a high correlation coefficient (r N 0.998). The average coefficient of variation for the two levels of controls was 2.78%. All of the edoxaban control measurements in each analytical run for both levels of controls were within the prespecified acceptance range (ie, 29–45 ng/mL for control 1, and 85–115 for control 2). The mean (SD) across analytical runs for edoxaban control 1 was 38.1 (1.2), and 103.0 (2.4) ng/mL for control 2. The day 5 mean ± SD edoxaban equivalent concentrations, as measured by the anti-FXa assay, and the edoxaban + M-4 concentration, as measured with LC/MS/MS in sodium citrate, across postdose time points for both edoxaban doses are shown in Fig. 1A and B. The concentration estimates derived from the two assays across all postdose time points were significantly correlated (Fig. 1C; edoxaban 60 mg: r = 0.987; edoxaban 90 mg: r = 0.989; both doses: r = 0.988; P b 0.0001 for all). For predose and at all time points between 0.5 h and 32 h postdose, the correlations were also significant between the assays (P b 0.0001 for both doses combined). Bland-Altman plots for the ratio of day 5 postdose anti-FXa edoxaban equivalent concentrations vs edoxaban + M-4 concentrations (LC/MS/MS in sodium citrate) are shown in Fig. 2. For the 60-mg edoxaban dose, the 95% limit of agreement was 0.85 to 1.45 for the ratio. For the 90-mg edoxaban dose, the 95% limit of agreement was 0.82 to 1.47 for the ratio. The Bland-Altman plots for the difference between postdose anti-FXa edoxaban equivalent concentrations vs edoxaban + M-4 concentrations are shown in Supplementary Fig. 1. Table 2 shows the difference scores and ratios for anti-FXa edoxaban equivalent concentrations vs LC/MS/MS assays; edoxaban + M-4 stratified by edoxaban concentration ranges. For concentration ranges in a clinically relevant range (between N30 and 50 ng/mL), the mean concentration estimate by anti-FXa was 12% higher than by LC/MS/MS or approximately 4 ng/mL greater. Fig. 3 shows the GLSM ratios (90% CI) derived from the ANOVA for anti-FXa edoxaban equivalent concentrations vs edoxaban + M-4 concentrations (LC/MS/MS in sodium citrate). The observed GLSM ratios were relatively consistent in the range of 4 to 12 h postdose, when the edoxaban equivalent concentration was above the LLOD for the anti-FXa assay (15 ng/mL). Across all time points, the GLSM ratio for edoxaban equivalent concentrations vs edoxaban + M-4 concentrations was Table 1 Baseline demographics and clinical characteristics.

Age (y), mean (SD) Male, n (%) Race, n (%) White Black or African American BMI (kg/m2), mean (SD)

Edoxaban 60 mg (n = 15)

Edoxaban 90 mg (n = 15)

Overall

36.3 (9.60) 9 (60.0)

42.5 (9.80) 12 (80.0)

39.4 (10.0) 21 (70.0)

5 (33.3) 10 (66.7) 25.8 (2.8)

10 (66.7) 5 (33.3) 25.9 (2.7)

15 (50.0) 15 (50.0) 25.9 (2.7)

BMI, body mass index; SD, standard deviation.

124

L. He et al. / Thrombosis Research 155 (2017) 121–127

Fig. 1. Edoxaban equivalent mean ± SD concentrations as measured by the anti-FXa assay and edoxaban + M-4 mean ± SD concentrations as measured by LC/MS/MS in sodium citrate tubes following administration of (A) 60-mg and (B) 90-mg edoxaban doses; dashed lines show the concentration range 20–400 ng/mL (C) correlation between anti-FXa edoxaban equivalent concentration and LC/MS/MS edoxaban + M-4 concentration in sodium citrate tubes across all postdose time points. FXa, factor Xa; LC/MS/MS, liquid chromatography coupled with tandem mass spectrometry; SD, standard deviation.

114.3% (108.2–120.8) for edoxaban 60 mg (P b 0.0001) and 113.0% (107.1, 119.2) for edoxaban 90 mg (P = 0.0002). Overall, the GLSM observed with anti-FXa assay were slightly larger than those observed with LC/MS/MS. This slight positive bias was more pronounced around the limit of detection. 3.3. Edoxaban plasma concentration in sodium citrate vs lithium heparin tubes The edoxaban concentration estimates from plasma samples in the sodium citrate and lithium heparin tubes as assessed by LC/MS/MS were significantly correlated (60 mg edoxaban: r = 0.988; 90 mg edoxaban: r = 0.987; both doses: r = 0.987; P b 0.001 for all). At every time point, the correlations were also significant (P ≤ 0.0025 at all postdose time points for both doses and overall). There was a consistent negative bias observable in the Bland-Altman analysis for the difference between the sodium citrate tube and the lithium heparin tubes (Fig. 4A and B). This negative bias is consistent with the dilution effect in the sodium citrate tubes. There was a consistent ratio of GLSM (90% CI) observed across a wide range of concentrations (Fig. 4C and D). At every postdose time point, the GLSM ratios from the sodium citrate tubes were smaller than those from the lithium heparin tubes. Across all time points the GLSM ratio for edoxaban concentrations in sodium citrate tubes vs lithium heparin tubes was 82.8% (78.5–87.3) for edoxaban 60 mg and 83.9% (79.1–89.0) for edoxaban 90 mg (both P b 0.0001).

3.4. Safety and tolerability Overall 26.7% (8/30) participants experienced ≥ 1 AE during the course of the study (60 mg: 33.3% [5/15]; 90 mg: 20.0% [3/15]); all were mild in intensity. There were no serious AEs and no deaths; one AE led to discontinuation (hematochezia). The only AEs occurring in ≥ 2 participants were headache (n = 3) and hematochezia (n = 2). All AEs were of mild severity. Of the reported AEs, only hematochezia was considered by investigators to be edoxaban-related.

4. Discussion This phase I, open-label, multiple-dose, two-treatment PK study compared the edoxaban equivalent concentrations estimated by an anti-FXa assay with appropriate edoxaban-specific calibrators and controls with the concentrations of edoxaban + M-4 measured by a LC/MS/ MS assay. The predominant active metabolite of edoxaban is M-4, and it reaches b10% of the exposure of edoxaban in healthy individuals [1,2, 30]. The exposure to the other metabolites is b5% of exposure to edoxaban [1,2]. In addition, M-4 has similar anti-FXa activity and molecular weight as its parent drug [1,2,32]; therefore, summing edoxaban and M-4 concentrations should closely approximate the edoxaban equivalent concentration. In this study, there was a robust linear correlation and good agreement between the concentration estimates of the assays. This

L. He et al. / Thrombosis Research 155 (2017) 121–127

Fig. 2. Bland–Altman plots showing ratio scores for anti-FXa edoxaban equivalent concentrations vs LC/MS/MS edoxaban + M-4 concentrations vs the average concentration values across both assays for the (A) 60 mg and (B) 90 mg (edoxaban doses, respectively. FXa, factor Xa; LC/MS/MS, liquid chromatography coupled with tandem mass spectrometry; SD, standard deviation.

correlation was found across all postdose time points. In the Bland-Altman analysis, the agreement between edoxaban equivalent concentrations (anti-FXa assay) vs edoxaban + M-4 concentrations (LC/MS/MS) was relatively consistent across measured concentrations ranging Table 2 Difference scores and ratios for anti-FXa edoxaban equivalent concentrations vs LC/MS/ MS assays edoxaban + M-4 stratified by edoxaban concentration ranges. Edoxaban concentration ranges, ng/mL b15

N50–100 N100–300

N300–450

N450

Difference (anti-FXa - LC/MS/MS) Mean 4.98 3.80 4.25 (SD) (1.53) (1.89) (2.72) Median 4.42 3.65 4.09 95% CI 4.43, 3.37, 3.40, 5.53 4.22 5.10

9.65 (7.43) 8.14 7.93, 11.37

13.85 (20.15) 15.13 11.04, 16.63

23.04 (28.21) 23.12 12.10, 33.98

n

74

202

28

−41.57 (6.48) −41.57 − 99.76, 16.62 2

1.14 (0.10) 1.13 1.12, 1.16 74

1.09 (0.12) 1.09 1.07, 1.10

1.07 (0.09) 1.06 1.04, 1.11

202

28

32

15–30

77

N30–50

42

Ratio (anti-FXa vs LC/MS/MS) Mean 1.47 1.22 1.12 (SD) (0.23) (0.13) (0.07) Median 1.39 1.22 1.11 95% CI 1.39, 1.19, 1.09, 1.55 1.25 1.14 n 32 77 42

0.91 (0.01) 0.91 0.82, 1.01 2

CI, confidence interval; FXa, factor Xa; LC/MS/MS, liquid chromatography coupled with tandem mass spectrometry; SD, standard deviation.

125

Fig. 3. Ratio of GLSM concentrations (90% CI) across postdose time points following administration of (A) 60-mg and (B) 90-mg doses of edoxaban for anti-FXa edoxaban equivalent concentrations vs LC/MS/MS edoxaban + M-4 concentrations in sodium citrate tubes plotted on the left y-axis. Edoxaban equivalent concentration (anti-FXa assay) values are shown in gray and plotted on the right y-axis. CI, confidence interval; FXa, factor Xa; GLSM, geometric least squares mean; LC/MS/MS, liquid chromatography coupled with tandem mass spectrometry *P b 0.05.

from 15 ng/mL to 450 ng/mL. In addition, the observed GLSM ratios were relatively consistent at edoxaban concentrations above the LLOD for the anti-FXa assay (15 ng/mL). However, there was a positive bias noted that was more pronounced around the limit of detection. The reason for this positive bias is not known, but it may be related to batch-tobatch variance. It is important to note that, between 100 ng/mL and 150 ng/mL, there was a slight change in slope observed in the correlation between edoxaban concentrations measured by the two assays. This effect is likely due to the redilution of the plasma samples that occurred in this concentration range. Anti-FXa assays require samples to be collected in sodium citrate tubes, but samples were collected in lithium heparin tubes during the clinical development of edoxaban for use in LC/MS/MS assays. The sodium citrate tubes contain a liquid additive equal to 10% of the maximum fill volume of the tube, while the lithium heparin tubes contain no liquid additive and thus do not dilute the sample. Therefore, this study also assessed the relationship between edoxaban concentration assessed by LC/MS/MS in samples collected by both methods. As would be expected, there was a dilution effect observed in samples collected in sodium citrate tubes. The GLSM concentrations for samples collected in sodium citrate tubes were approximately 16% to 17% less than the GLSM concentrations for samples collected in lithium heparin tubes. This conversion factor may be applicable when comparing LC/MS/MS concentrations that were collected using the standard lithium heparin tubes with anti-FXa assay edoxaban equivalent concentration estimates. Several in vitro studies have assessed the performance of chromogenic assays designed to measure heparin and anti-FXa chromogenic assays with non-edoxaban–specific calibrators and controls for measuring edoxaban concentrations [19,20]. In a previous in vitro study,

126

L. He et al. / Thrombosis Research 155 (2017) 121–127

Fig. 4. Bland–Altman plots showing ratio scores for LC/MS/MS edoxaban concentration from samples collected in sodium citrate tubes vs lithium heparin tubes relative to the average concentration values across both collection methods following administration of (A) 60-mg and (B) 90-mg edoxaban doses. Ratio of GLSM concentration (90% CI) across postdose time points following administration of (C) 60-mg and (D) 90-mg doses of edoxaban for LC/MS/MS in sodium citrate tubes vs lithium heparin tubes. CI, confidence interval; FXa, factor Xa; GLSM, geometric least squares mean; Hep, heparin; LC/MS/MS, liquid chromatography coupled with tandem mass spectrometry; SD, standard deviation *P b 0.05.

chromogenic assays without edoxaban-specific calibrators and controls showed good sensitivity to measuring edoxaban levels in edoxabanspiked plasma [19]. However, to our knowledge, this is the first study to assess the applicability of an anti-FXa assay to assess edoxaban concentrations in human plasma samples using edoxaban-specific calibrators and controls. Previous publications have assessed the suitability of anti-FXa assays for the measurement of rivaroxaban and apixaban concentrations [17, 18,25–29]. As in this study, those publications reported that anti-FXa chromogenic assays can accurately measure a wide range of rivaroxaban and apixaban concentrations after the administration of therapeutic doses [17,18,25–29]. In a multicenter field trial, an antiFXa chromogenic assay accurately measured rivaroxaban concentrations in plasma in the range of 20 ng/mL to 660 ng/mL [18]. Similarly, in patients with acute coronary syndrome enrolled in a phase III study, apixaban concentrations as measured by anti-FXa assay strongly correlated with concentration levels assessed by mass spectrometry/high performance liquid chromatography (r = 0.967) [28]. These results, together with our results, support that anti-FXa chromogenic assays, when used with appropriate calibrators and controls, are suitable for measuring levels of direct, oral FXa inhibitors. There are some limitations of the present study that could affect the generalizability of the results. This study was conducted in only a small number of healthy participants and the focus of the study was on comparing different assay methodologies and plasma matrices. It will be important to assess the performance of anti-FXa assays for measuring edoxaban concentrations in patients in whom anticoagulants are indicated. The FDA notes that, for assays used for routine monitoring of DOAC levels, the relationship should be established between the measured therapeutic range and clinical outcomes in relevant patient populations [21]. Only two doses of edoxaban were investigated (60 mg and 90 mg once-daily). In addition, this study was performed at a single

study center and therefore measures of inter-laboratory consistency could not be assessed. 5. Conclusions In this study, an anti-FXa chromogenic assay with edoxaban-specific calibrators and controls demonstrated good correlation and accuracy in estimating edoxaban concentrations across a wide range of concentrations relative to LC/MS/MS at steady state following the administration of once-daily edoxaban for 5 days. Although edoxaban typically does not require routine laboratory monitoring, there are situations when clinical management may require specific quantification of anticoagulation, such as before surgery or when a patient is actively bleeding [11]. In these situations, anti-FXa assays may be useful in quickly determining edoxaban concentrations; however, further clinical research is needed to document which concentration thresholds or ranges are clinically relevant. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.thromres.2017.05.005.

Conflict of interests Ling He, Jarema Kochan, Min Lin, and Alexander Vandell are employees of Daiichi Sankyo, Inc. Karen Brown was an employee at Daiichi Sankyo, Inc. at the time the study was conducted. Francois Depasse is an employee of Diagnostica Stago.

Funding The study was funded by Daiichi Sankyo, Inc. (Parsippany, NJ).

L. He et al. / Thrombosis Research 155 (2017) 121–127

Acknowledgements Medical writing and editorial support was provided by Stefan Kolata, PhD, of AlphaBioCom, LLC (King of Prussia, PA), and funded by Daiichi Sankyo, Inc. (Parsippany, NJ). The authors would also like to acknowledge Madhuri Desai, previously of Daiichi Sankyo Pharma Development (Edison, NJ), Raj Mangaraj of Q2 Solutions (Morrisville, NC), and Lukasz Biernat and Joseph Hagood of Medpace (Cincinnati, Ohio). References [1] D.A. Parasrampuria, K.E. Truitt, Pharmacokinetics and pharmacodynamics of edoxaban, a non-vitamin K antagonist oral anticoagulant that inhibits clotting factor Xa, Clin. Pharmacokinet. 55 (6) (2016) 641–655. [2] SAVAYSA™, (Edoxaban) tablets for oral use, Full Prescribing Information, Daiichi Sankyo Inc., Parsippany, NJ, USA, 2015. [3] S.J. Connolly, M.D. Ezekowitz, S. Yusuf, J. Eikelboom, J. Oldgren, A. Parekh, J. Pogue, P.A. Reilly, E. Themeles, J. Varrone, S. Wang, M. Alings, D. Xavier, J. Zhu, R. Diaz, B.S. Lewis, H. Darius, H.C. Diener, C.D. Joyner, L. Wallentin, RE-LY Steering Committee and Investigators, Dabigatran versus warfarin in patients with atrial fibrillation, N. Engl. J. Med. 361 (12) (2009) 1139–1151. [4] R.P. Giugliano, C.T. Ruff, E. Braunwald, S.A. Murphy, S.D. Wiviott, J.L. Halperin, A.L. Waldo, M.D. Ezekowitz, J.I. Weitz, J. Spinar, W. Ruzyllo, M. Ruda, Y. Koretsune, J. Betcher, M. Shi, L.T. Grip, S.P. Patel, I. Patel, J.J. Hanyok, M. Mercuri, E.M. Antman, ENGAGE AF-TIMI 48 Investigators, Edoxaban versus warfarin in patients with atrial fibrillation, N. Engl. J. Med. 369 (22) (2013) 2093–2104. [5] C.B. Granger, J.H. Alexander, J.J. McMurray, R.D. Lopes, E.M. Hylek, M. Hanna, H.R. AlKhalidi, J. Ansell, D. Atar, A. Avezum, M.C. Bahit, R. Diaz, J.D. Easton, J.A. Ezekowitz, G. Flaker, D. Garcia, M. Geraldes, B.J. Gersh, S. Golitsyn, S. Goto, A.G. Hermosillo, S.H. Hohnloser, J. Horowitz, P. Mohan, P. Jansky, B.S. Lewis, J.L. Lopez-Sendon, P. Pais, A. Parkhomenko, F.W. Verheugt, J. Zhu, L. Wallentin, ARISTOTLE Committees and Investigators, Apixaban versus warfarin in patients with atrial fibrillation, N. Engl. J. Med. 365 (11) (2011) 981–992. [6] Hokusai VTE Investigators, H.R. Buller, H. Decousus, M.A. Grosso, M. Mercuri, S. Middeldorp, M.H. Prins, G.E. Raskob, S.M. Schellong, L. Schwocho, A. Segers, M. Shi, P. Verhamme, P. Wells, Edoxaban versus warfarin for the treatment of symptomatic venous thromboembolism, N. Engl. J. Med. 369 (15) (2013) 1406–1415. [7] Einstein Investigators, R. Bauersachs, S.D. Berkowitz, B. Brenner, H.R. Buller, H. Decousus, A.S. Gallus, A.W. Lensing, F. Misselwitz, M.H. Prins, G.E. Raskob, A. Segers, P. Verhamme, P. Wells, G. Agnelli, H. Bounameaux, A. Cohen, B.L. Davidson, F. Piovella, S. Schellong, Oral rivaroxaban for symptomatic venous thromboembolism, N. Engl. J. Med. 363 (26) (2010) 2499–2510. [8] Einstein-Pe Investigators, H.R. Buller, M.H. Prins, A.W. Lensin, H. Decousus, B.F. Jacobson, E. Minar, J. Chlumsky, P. Verhamme, P. Wells, G. Agnelli, A. Cohen, S.D. Berkowitz, H. Bounameaux, B.L. Davidson, F. Misselwitz, A.S. Gallus, G.E. Raskob, S. Schellong, A. Segers, Oral rivaroxaban for the treatment of symptomatic pulmonary embolism, N. Engl. J. Med. 366 (14) (2012) 1287–1297. [9] M.R. Patel, K.W. Mahaffey, J. Garg, G. Pan, D.E. Singer, W. Hacke, G. Breithardt, J.L. Halperin, G.J. Hankey, J.P. Piccini, R.C. Becker, C.C. Nessel, J.F. Paolini, S.D. Berkowitz, K.A. Fox, R.M. Califf, ROCKET AF Investigators, Rivaroxaban versus warfarin in nonvalvular atrial fibrillation, N. Engl. J. Med. 365 (10) (2011) 883–891. [10] S. Schulman, C. Kearon, A.K. Kakkar, P. Mismetti, S. Schellong, H. Eriksson, D. Baanstra, J. Schnee, S.Z. Goldhaber, RE-COVER Study Group, Dabigatran versus warfarin in the treatment of acute venous thromboembolism, N. Engl. J. Med. 361 (24) (2009) 2342–2352. [11] T. Baglin, The role of the laboratory in treatment with new oral anticoagulants, J. Thromb. Haemost. 11 (Suppl. 1) (2013) 122–128. [12] A. Cuker, H. Husseinzadeh, Laboratory measurement of the anticoagulant activity of edoxaban: a systematic review, J. Thromb. Thrombolysis 39 (3) (2015) 288–294. [13] X. Delavenne, P. Mismetti, T. Basset, Rapid determination of apixaban concentration in human plasma by liquid chromatography/tandem mass spectrometry: application to pharmacokinetic study, J. Pharm. Biomed. Anal. 78-79 (2013) 150–153.

127

[14] E.M. Schmitz, K. Boonen, D.J. van den Heuvel, J.L. van Dongen, M.W. Schellings, J.M. Emmen, F. van der Graaf, L. Brunsveld, D. van de Kerkhof, Determination of dabigatran, rivaroxaban and apixaban by ultra-performance liquid chromatography - tandem mass spectrometry (UPLC-MS/MS) and coagulation assays for therapy monitoring of novel direct oral anticoagulants, J. Thromb. Haemost. 12 (10) (2014) 1636–1646. [15] D.M. Siegal, M.A. Crowther, Acute management of bleeding in patients on novel oral anticoagulants, Eur. Heart J. 34 (7) (2013) 489–498b. [16] J.A. Reiffel, J.I. Weitz, P. Reilly, E. Kaminskas, T. Sarich, P. Sager, J. Seltzer, Cardiac Safety Research Consortium presenters and participants, NOAC monitoring, reversal agents, and post-approval safety and effectiveness evaluation: A cardiac safety research consortium think tank, Am. Heart J. 177 (2016) 74–86. [17] A. Hillarp, K.M. Gustafsson, L. Faxalv, K. Strandberg, F. Baghaei, I. Fagerberg Blixter, M. Berndtsson, T.L. Lindahl, Effects of the oral, direct factor Xa inhibitor apixaban on routine coagulation assays and anti-FXa assays, J. Thromb. Haemost. 12 (9) (2014) 1545–1553. [18] M.M. Samama, G. Contant, T.E. Spiro, E. Perzborn, C. Guinet, Y. Gourmelin, L. Le Flem, G. Rohde, J.L. Martinoli, L. Rivaroxaban Anti-Factor Xa Chromogenic Assay Field Trial Laboratories, Evaluation of the anti-factor Xa chromogenic assay for the measurement of rivaroxaban plasma concentrations using calibrators and controls, Thromb. Haemost. 107 (2) (2012) 379–387. [19] J. Douxfils, B. Chatelain, C. Chatelain, J.M. Dogne, F. Mullier, Edoxaban: impact on routine and specific coagulation assays. A practical laboratory guide, Thromb. Haemost. 115 (2) (2016) 368–381. [20] T. Iba, M. Emmi, M. Hiki, M. Nagayama, K. Aihara, Y. Tabe, M. Yuri, A. Ohsaka, Comparison of prothrombin time tests used in the monitoring of edoxaban and their evaluation as indicators of the reversal effect, Int. J. Hematol. 103 (6) (2016) 665–672. [21] US Food and Drug Administration, Public Workshop - In Vitro Diagnostic Testing for Direct Oral Anticoagulants. bhttp://www.fda.gov/MedicalDevices/NewsEvents/ WorkshopsConferences/ucm459448.htmN, 2015 (accessed 07/11/2016). [22] S.M. Bates, J.I. Weitz, Coagulation assays, Circulation 112 (4) (2005) e53–e60. [23] M.M. Samama, S. Kunitada, A. Oursin, F. Depasse, S. Heptinstall, Comparison of a direct factor Xa inhibitor, edoxaban, with dalteparin and ximelagatran: a randomised controlled trial in healthy elderly adults, Thromb. Res. 126 (4) (2010) e286–e293. [24] M.M. Samama, J.L. Martinoli, L. LeFlem, C. Guinet, G. Plu-Bureau, F. Depasse, E. Perzborn, Assessment of laboratory assays to measure rivaroxaban—an oral, direct factor Xa inhibitor, Thromb. Haemost. 103 (4) (2010) 815–825. [25] M.M. Samama, G. Contant, T.E. Spiro, E. Perzborn, L. Le Flem, C. Guinet, Y. Gourmelin, G. Rohde, J.L. Martinoli, Laboratory assessment of rivaroxaban: a review, Thromb. J. 11 (1) (2013) 11. [26] I. Gouin-Thibault, C. Flaujac, X. Delavenne, S. Quenet, M.H. Horellou, S. Laporte, V. Siguret, T. Lecompte, Assessment of apixaban plasma levels by laboratory tests: suitability of three anti-Xa assays. A multicentre French GEHT study, Thromb. Haemost. 111 (2) (2014) 240–248. [27] H. Mani, G. Rohde, G. Stratmann, C. Hesse, N. Herth, S. Schwers, E. Perzborn, E. Lindhoff-Last, Accurate determination of rivaroxaban levels requires different calibrator sets but not addition of antithrombin, Thromb. Haemost. 108 (1) (2012) 191–198. [28] R.C. Becker, H. Yang, Y. Barrett, P. Mohan, J. Wang, L. Wallentin, J.H. Alexander, Chromogenic laboratory assays to measure the factor Xa-inhibiting properties of apixaban—an oral, direct and selective factor Xa inhibitor, J. Thromb. Thrombolysis 32 (2) (2011) 183–187. [29] L.M. Asmis, L. Alberio, A. Angelillo-Scherrer, W. Korte, A. Mendez, G. Reber, B. Seifert, H. Stricker, D.A. Tsakiris, W.A. Wuillemin, Rivaroxaban: quantification by anti-FXa assay and influence on coagulation tests: a study in 9 Swiss laboratories, Thromb. Res. 129 (4) (2012) 492–498. [30] M.S. Bathala, H. Masumoto, T. Oguma, L. He, C. Lowrie, J. Mendell, Pharmacokinetics, biotransformation, and mass balance of edoxaban, a selective, direct factor Xa inhibitor, in humans, Drug Metab. Dispos. 40 (12) (2012) 2250–2255. [31] D.M. Adcock, Collection, Transport, and Processing of Blood Specimes for Testing Plasma-Based Coagulation Assays and Molecular Hemostasis Assays: Approved Guideline, fifth ed. Clinical and Laboratory Standards Institute, Wayne, PA, 2008. [32] Toxicology Data Network, Substance Name: D21-2393.bhttps://chem.nlm.nih.gov/ chemidplus/rn/834919-19-6N, (accessed December 16, 2016).