ORIGINAL ARTICLE
Measurement of the Direct Oral Anticoagulants Apixaban, Dabigatran, Edoxaban, and Rivaroxaban in Human Plasma Using Turbulent Flow Liquid Chromatography With High-Resolution Mass Spectrometry Tracey Gous, MSc,* Lewis Couchman, MSc,*† Jignesh P. Patel, PhD,†‡ Chitongo Paradzai, MSc,‡ Roopen Arya, PhD,‡ and Robert J. Flanagan, PhD*
Background: Direct oral anticoagulants (DOACs) are prescribed
(2 weeks) and of 18% and 70% (1 day and 2 weeks, respectively) at room temperature.
for systemic anticoagulation. Fixed doses are recommended, but dose individualization may be warranted. Functional coagulation assays may be available, but their use requires knowledge of the drug taken. To provide alternative methodology for guiding dosage, we have developed and validated a liquid chromatography–mass spectrometric assay for apixaban, dabigatran, edoxaban, and rivaroxaban at the concentrations attained during therapy.
Conclusions: The method is suitable for high-throughput therapeutic drug monitoring of DOACs. The acquisition of full scan data allows for the retrospective identification of metabolites. The method can be used to identify a particular DOAC if information on the drug taken is lacking.
Methods: Samples, calibrators, and internal quality controls (100 mL) were mixed with internal standard solution (50 mg/L both dabigatran-13C6 and rivaroxaban-13C6 in acetonitrile) and, after centrifugation (16,400g, 4 minutes), supernatant (100 mL) was injected onto a Cyclone-C18-P-XL TurboFlow column. Analytes were focused onto an Accucore PhenylHexyl (2.1 · 100 mm, 2.6 mm) analytical column and eluted using a methanol + acetonitrile (1 + 1): aqueous ammonium acetate (10 mmol/L) gradient. Data were acquired using high-resolution mass spectrometry in full-scan mode (100–2000 m/z) with data-dependent fragmentation to confirm peak identity. Calibration was linear (1–500 mg/L all analytes).
Results: Total analysis time was 6 minutes. Intra-assay imprecision (% RSD) at 1 mg/L was 2.6%, 4.2%, 17.3%, and 9.5% for apixaban, dabigatran, edoxaban, and rivaroxaban, respectively. Mean recovery was 96%–101%. No signal suppression or enhancement was observed. Apixaban, dabigatran, and rivaroxaban were stable over 3 freeze–thaw cycles, after storage at room temperature, and at 2–88C for up to 2 weeks. Edoxaban was stable over 3 freeze–thaw cycles but showed a mean deterioration of 16% if stored at 2–88C
Received for publication November 14, 2013; accepted January 29, 2014. From the *Toxicology Unit, Department of Clinical Biochemistry, King’s College Hospital NHS Foundation Trust; †Institute of Pharmaceutical Science, King’s College London; and ‡King’s Thrombosis Centre, Department of Haematological Medicine, King’s College Hospital NHS Foundation Trust, London, United Kingdom. The authors declare no conflict of interest. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.drug-monitoring.com). Correspondence: Tracey Gous, MSc, Toxicology Unit, Department of Clinical Biochemistry, 3rd Floor, Bessemer Wing, King’s College Hospital NHS Foundation Trust, Denmark Hill, London SE5 9RS, United Kingdom (e-mail: [email protected]). Copyright © 2014 by Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
Key Words: liquid chromatography–high-resolution mass spectrometry, therapeutic drug monitoring, direct oral anticoagulants, dabigatran, rivaroxaban, apixaban, edoxaban (Ther Drug Monit 2014;36:597–605)
INTRODUCTION Oral anticoagulation is prescribed for millions of patients worldwide. For more than 50 years, vitamin K antagonists (VKAs) such as warfarin have been used for this purpose. However, use of VKAs is not straightforward and frequent testing of the international normalized ratio (INR) is needed to optimize dosage.1 Recently, direct oral anticoagulants (DOACs; Fig. 1) have become available. As of end-2013, 3 DOACs [apixaban (Eliquis; Bristol-Myers Squibb), dabigatran etexilate (Pradaxa; Boehringer Ingelheim), and rivaroxaban (Xarelto; Bayer)] were licensed in the United Kingdom for clinical use.2 [Note: Dabigatran etexilate is a pro-drug, the active moiety dabigatran (Fig. 1) being rapidly formed in vivo.] DOACs selectively inhibit coagulation serine proteases (Xa or IIa). They have a rapid onset of action, a relatively predictable pharmacokinetic profile, and a relatively short plasma half-life, making initiation, maintenance, and discontinuation of anticoagulant therapy considerably easier than with VKAs. However, at present, there is no effective antidote to immediately reverse anticoagulation should the need arise. With VKAs, the degree of anticoagulation is titrated to the patient’s requirement (usually an INR of 2–3), and thus, therapy is highly individualized, taking into account age, concomitant drug therapy, diet, alcohol intake, and body weight.3 However, with DOACs, a “one-size fits all” approach has been adopted, with dose adjustments suggested should the patient meet certain criteria, for example, dose reduction with dabigatran when renal function is impaired.
597
Gous et al
Ther Drug Monit Volume 36, Number 5, October 2014
plasma DOAC concentrations.9 For these reasons and for use in pharmacokinetic studies, methods have been developed and validated for the therapeutic drug monitoring of DOACs in plasma using liquid chromatography–tandem mass spectrometry (LC-MS/MS).17–23 The plasma concentrations of apixaban attained during therapy are reported to range up to 300 mg/L,16 whereas for dabigatran and rivaroxaban, the limits above which there is an increased risk bleeding are thought to be 200 and 240 mg/L, respectively.19,21 To monitor DOAC concentrations in patient samples and give the option of measuring metabolites of potential interest,18 we have developed a simple method to measure plasma apixaban, dabigatran, rivaroxaban, and a further DOAC edoxaban (Fig. 1), using TurboFlow sample preparation coupled with high-resolution LC-MS (LC-HR-MS). The method should prove able to incorporate further DOACs, such as betrixaban,24 as they are introduced clinically.
MATERIALS AND METHODS Chemicals and Reagents FIGURE 1. Structural formulae of apixaban, dabigratran, edoxaban, and rivaroxaban. The positions of the isotopically labelled 13C aromatic rings (for dabigatran-13C6 and rivaroxaban-13C6) are indicated (*).
However, the British Society of Haematology and the International Society of Thrombosis and Haemostasis have stipulated that measurement of DOAC activity might be warranted: (1) after overdose (accidental or otherwise), (2) in the elderly, especially those with comorbidities, (3) when patients switch from an existing oral anticoagulant, (4) in those at the extremes of body weight, (5) in patients with hepatic or renal impairment, (6) in patients co-prescribed other medications where there is a risk of drug–DOAC interactions, (7) to assess adherence, and (8) to assess anticoagulant activity before major surgery.4–11 There is at present no consensus on the best methodology for assessing DOAC activity in vivo and hence guiding dosage. Unlike with VKAs, traditional coagulation tests, such as the INR, cannot be used to monitor the effects of DOACs. This is because different DOACs have different sensitivity to coagulation assays, such as activated partial thromboplastin time, prothrombin time, thrombin time, dilute thrombin time, ecarin clotting time, and drug-specific anti-Xa assays.12,13 Assay sensitivity can also depend on the reagent/analyzer combination used.13–16 Despite this, different functional coagulation assays have been tested and calibrated to individual DOACs to estimate plasma DOAC concentrations in patient samples. As the number of DOACs available for clinical use is set to increase, it is possible that in some situations, for example, if the patient presents in coma, it will be not be known which DOAC the patient has been taking. Furthermore, because it has been suggested that plasma DOAC concentrations should be ,30 mg/L before major surgery, it is especially important to have good assay performance at lower
598
Dabigatran and rivaroxaban were from LGC Standards (Teddington, United Kingdom). Apixaban, edoxaban, dabigatran-13C6, and rivaroxaban-13C6 were from ALSA-CHIM (Illkirch, France). MS grade formic acid and ammonium acetate were from Fluka (Poole, United Kingdom). HPLC grade methanol, acetonitrile, and acetone were from Sigma–Aldrich (Poole, United Kingdom). AnalaR grade 2-propanol and concentrated hydrochloric acid were from VWR (Lutterworth, United Kingdom). Water was deionized (18 mV; Elga, Marlow, United Kingdom), and analyte-free pooled human plasma was from Sera Labs (Haywards Heath, United Kingdom).
Calibrators, Internal Quality Controls, and Internal Standard Solution Separate calibrator and internal quality control (IQC) stock solutions were prepared for dabigatran (200 mg/L in 0.1 mol/L aqueous hydrochloric acid), rivaroxaban (100 mg/L in acetonitrile), apixaban (400 mg/L in acetonitrile), and edoxaban (100 mg/L in acetonitrile). Calibrator working stock solutions were prepared at concentrations of 10 and 1 mg/L (all analytes) by dilution of individual stock solutions in acetonitrile. IQC stock and working solutions were prepared separately. To prepare calibrators and IQC solutions, appropriate volumes of working stock solutions were pipetted into volumetric glassware, evaporated to dryness under a gentle stream of nitrogen, and made up to volume with analyte-free human plasma. Calibrators were prepared containing each analyte at 1, 5, 10, 50, 100, and 500 mg/L, and IQC solutions were similarly prepared at 25, 75, and 400 mg/L. Prepared calibrators and IQC solutions were mixed by inversion and allowed to equilibrate (48C, 24 hours), before storage in approximately 1 mL portions at 2208C in 2-mL screw-top polypropylene tubes (Alpha Laboratories, Eastleigh, United Kingdom). To prepare the internal standard (IS) solution, individual stock solutions of dabigatran-13C6 (100 mg/L in 0.1 mol/L 2014 Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
aqueous hydrochloric acid) and rivaroxaban-13C6 (100 mg/L in acetonitrile) were prepared and stored at 2208C. Working IS solution (50 mg/L both dabigatran-13C6 and rivaroxaban-13C6) was prepared as necessary by appropriate dilution from the IS stock solutions with acetonitrile.
Sample Preparation
Sample, calibrator, or IQC solution (100 mL) was added to a 60 · 7.5-mm i.d. glass tube (Dreyer tube; Esslab, Essex, United Kingdom) and working IS solution (200 mL) was added. The contents of the tube were vortex mixed (30 seconds) and centrifuged (16,400g, 4 minutes) and the supernatant was transferred to an autosampler vial (2-mL glass vial with 300-mL glass insert; Kinesis, St Neots, United Kingdom) and capped for analysis.
High-Resolution LC-MS An Aria Transcend TLX-II system (Thermo Fisher Scientific, San Jose, CA) consisting of 4 Accela 1250 highpressure quaternary pumps, valve interface module, and CTC PAL autosampler was used with a Q Exactive MS (Thermo Fisher Scientific, Bremen, Germany). The autosampler tray was maintained at +108C. Instrument control was performed using TraceFinder software (version 3.1; Thermo Fisher Scientific). System eluents were as follows: (A) 0.1% (vol/vol) formic acid in 10 mmol/L aqueous ammonium acetate, (B) 0.1% (vol/vol) formic acid in methanol:acetonitrile (1 + 1), and (C) acetonitrile:acetone:2-propanol (2 + 1 + 2). Prepared samples (100 mL) were injected onto a TurboFlow column (Cyclone C18-P-XL, 50 · 0.5-mm i.d., flow rate 2.0 mL/min). Retained analytes were eluted (reversed flow direction) from the TurboFlow column using elution solvent [eluent A:eluent B, 1 + 4 (vol/vol), 200 mL], stored in the holding loop, and focused through a T-piece onto an Accucore PhenylHexyl analytical column (2.7 mm average particle size, 100 · 2.1 mm i.d., Thermo Fisher Scientific, Runcorn, United Kingdom). The column was fitted with a 0.5-mm precolumn filter (Fisher Scientific, Loughborough, United Kingdom) and was maintained at 408C (HotPocket; Thermo Fisher Scientific). During gradient elution (total flow rate 0.30 mL/min) from the analytical column, the TurboFlow column was washed in turn (reversed flow direction) with
Measurement of DOACs Using LC-HR-MS
elution solvent and with wash solvent (eluent C). The whole system was then re-equilibrated before the next injection. The gradient elution and valve switching profiles are summarized in Table 1. The total analysis time was 6 minutes, including column re-equilibration. Eluent flow was diverted from the MS to waste for the first 3 minutes following each TurboFlow injection, and data were acquired for 3 minutes per analysis. MS detection was carried out in positive ionization mode using heated electrospray ionization [spray voltage 4 kV; temperatures: vaporizer 3008C; capillary 2508C; auxiliary, sheath, and sweep gases 55, 5, and 0 (arbitrary units), respectively, S-lens voltage 55 V]. Data were acquired in fullscan mode (70,000 resolution, range: 150–1000 m/z) with single data–dependent fragmentation (MS2) scans (17,500 resolution, stepped normalized collision energy 35% 6 50%, range: 150–1000 m/z) collected for each analyte during peak elution (based on a defined inclusion list) used to confirm peak identity. Data acquisition was performed using TraceFinder software.
Assay Calibration and Data Processing Quantitation was based on extracted accurate-mass full scan data [5 ppm window, external mass calibration (alternate days) using positive/negative ion calibration infusion solutions (Thermo Fisher Scientific)]. Calibration curves were generated by plotting the peak area ratio (analyte/IS) from the extracted ion chromatograms against analyte concentration. For apixaban and rivaroxaban, rivaroxaban-13C6 was used as the IS, and for dabigatran and edoxaban, dabigatran-13C6 was used. Lines were fitted by least squares regression (1/x weighting, not including or forced through the origin). Calibrators and matrix blanks were included at the beginning and end of each batch analysis, with IQCs included (1) after the first set of calibrators and immediately before the last set and (2) after every 10 injections throughout the sequence. Patient samples were analyzed in duplicate and the mean result taken. Samples with concentrations exceeding the calibration range for an analyte were diluted as appropriate with analyte-free pooled human plasma and re-assayed. Assay acceptance criteria were (1) linear (R2 . 0.98) calibration curves for each analyte and (2) IQC values within 615% nominal concentrations for all analytes.
TABLE 1. Summary of TurboFlow Method Loading (TurboFlow) Pump
Valves
Eluting (Analytical) Pump
Start Flow Flow Step Time (min) (mL/min) Gradient % Eluent A % Eluent B % Eluent C Tee Loop (mL/min) Gradient % Eluent A % Eluent B 1 2 3 4 5 6 7
0.00 0.50 1.50 2.00 3.00 3.50 4.50
2.00 0.10 1.00 1.00 1.00 1.50 2.00
Step Step Step Step Step Step Step
100 100 — — — 20 100
— — 100 100 — 80 —
— — — — 100 — —
Out In Out Out Out Out Out
Out In In In In In Out
0.50 0.50 0.30 0.30 0.30 0.30 0.50
Step Step Step Ramp Step Step Step
98 98 98 — — — 98
2 2 2 100 100 100 2
Step 1, sample loading; step 2, transfer of retained analytes to the analytical column; steps 3–5, gradient elution (analytical column) and column washing (TurboFlow column); step 6, elution loop filling (loading pump); step 7, column equilibration (1.5 minutes).
2014 Lippincott Williams & Wilkins
599
Gous et al
Method Validation The method was validated using the US Food and Drug Administration Center for Drug Evaluation and Research guidance for bioanalytical method validation.25 Intra-assay and interassay imprecision (% RSD) and inaccuracy were measured by replicate analysis of the IQC solutions and a solution prepared containing all analytes at the lower limit of quantification (LLoQ, 1 mg/L) on the same day and on different days, respectively. To investigate the overall process efficiency (PE) for the protein precipitation and TurboFlow steps,26 2 sets of solutions containing all analytes and ISs were prepared (for each set, 1 solution at a concentration of 200 mg/L, and 1 at 500 mg/L all analytes). One set of solutions was prepared in eluent A and the other set was prepared in analyte-free pooled human plasma. To assess absolute analyte recovery from the TurboFlow column, after diluting the aqueous solutions with eluent A (1 + 2, vol/vol) instead of working IS solution, these prepared solutions were (1) analyzed using the TurboFlow procedure and (2) analyzed with the TurboFlow system bypassed (ie, injection was performed directly onto the analytical column). The mean peak areas of each analyte analyzed using the complete TurboFlow procedure were compared with those observed from injection onto the analytical column only, which was assumed to represent 100% recovery. To assess the overall PE, mean peak areas for plasma solutions analyzed using the complete TurboFlow system (n = 3 at each concentration) were compared with those from aqueous solutions analyzed with the TurboFlow system bypassed. To further investigate ion suppression, analyte-free plasma from 10 independent sources was analyzed following treatment with acetonitrile instead of working IS solution. The detector response for each analyte was monitored while a solution containing all analytes (10 mg/L in acetonitrile, 10 mL/ min) was infused by syringe post-column.27 The LLoQ was taken as the lowest concentration (to the nearest 1 mg/L) where the intra-assay % RSD was less than 20. Carryover was assessed by running high IQCs (n = 3) followed immediately by low IQCs (n = 3) and comparing measured with nominal concentrations. Analyte stability in the IQC solutions was assessed (1) at room temperature over 14 days, (2) in a refrigerator (2–88C) over 14 days, and (3) through 3 freeze– thaw cycles, by comparison of mean peak areas (n = 3 at each concentration) against all the results obtained on analysis of IQC solutions that had been stored at 2208C before analysis (ie, had been subjected to a single freeze–thaw cycle). Autosampler tray stability was assessed using the IQC solutions and samples from patients treated with apixaban, dabigatran, or rivaroxaban. Prepared samples were analyzed at 0, 6, 12, and 72 hours, and mean peak areas (n = 2 at each time point) for all analytes and ISs were compared.
Patient Samples Dabigatran and Rivaroxaban Functional Coagulation Assays Patients requiring assessment of anticoagulant activity had blood drawn into 3.0-mL tubes containing 0.3 mL of 3.2% (wt/vol) aqueous trisodium citrate (Vacuette; Greiner
600
Ther Drug Monit Volume 36, Number 5, October 2014
Bio One, Stonehouse, United Kingdom). All tubes were filled to the draw-line indicator to ensure consistent dilution with citrate buffer and centrifuged [2 · 7 minutes, 1440g; 38R Rotina (Hettich, Tuttligen, Germany)] within 1 hour of sample collection. The diluted plasma was harvested and stored at 2408C before analysis, which was undertaken within 1 week of sample collection. Dabigatran concentrations were estimated using an ecarin chromogenic assay (HaemoSys, Jena, Germany) with dabigatran calibrators (50, 250, and 500 mg/L, Lot No. 22202) and controls (low and high, Lot No. 03001; Hyphen-Biomed, Neuville-Sur-Oise, France) supplied in lyophilized human plasma and reconstituted in deionized water according to the manufacturer’s instructions. Rivaroxaban concentrations were estimated using an STAliquid anti-Xa assay (Diagnostica Stago, Asnieres, France) with rivaroxaban calibrators (0, 100, 250, and 450 mg/L, Lot No. 281), and low and high controls (Diagnostica Stago, Lot No. 110838) were supplied in lyophilized human plasma and reconstituted in deionized water according to the manufacturer’s instructions. Analyses were conducted on the STA-R evolution analyzer (Diagnostica Stago).
LC-HR-MS Assay Excess, anonymized blood (EDTA-coated anticoagulant; Vacuette; Greiner) from patients prescribed dabigratran etexilate or rivaroxaban that had been submitted for hematological monitoring (full blood count) were analyzed after separation of plasma and storage at 2208C. Assays were performed in duplicate and the mean result taken. EDTA samples were collected within 2 hours of the samples used for functional assays. EDTA plasma samples from patients thought to have taken overdoses (times of ingestion and amounts taken not known) of dabigatran (67-year-old man) and apixaban (77-year-old woman) were also analyzed. Pharmacokinetic data were calculated using GraphPad Prism (Version 3.0).
RESULTS Method Performance Typical chromatograms are shown in Figure 2. The major product ions shown in the MS2 spectrum (see Figure, Supplemental Digital Content 1, http://links.lww.com/TDM/A80, which shows an example of a data-dependent MS2 spectrum generated for dabigatran) were in agreement with the product ions used in published methods using selected reaction monitoring–based LC-MS/MS instrumentation for all analytes assay.17–21,23 Intra-assay and interassay imprecision and inaccuracy are summarized in Table 2. Calibration was linear (R2 . 0.99) for all analytes across the range 1–500 mg/L. For all analytes and ISs, although the m/z extraction window was set to 65 ppm for data processing and quantitation, the measured mass accuracy (using daily external mass calibration) was typically less than 2 ppm (see Table, Supplemental Digital Content 2, http://links.lww.com/TDM/A81, which provides typical mass accuracy data for all analytes). No carryover was observed. Mean absolute recovery (n = 3 at each of the 2 concentrations 2014 Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
Measurement of DOACs Using LC-HR-MS
FIGURE 2. Typical extracted ion chromatogram. Extraction window 6 5 ppm based on theoretical m/z. Chromatograms: calibrator (10 mg/L each analyte), IQC solution (25 mg/L each analyte), and plasma sample from a patient prescribed rivaroxaban (20 mg/d, rivaroxaban concentration 199 mg/L). (i) Dabigatran, (ii) dabigatran-13C6, (iii) rivaroxaban, (iv) rivaroxaban-13C6, (v) edoxaban, and (vi) apixaban.
studied, solutions prepared in eluent A) for all analytes on the TurboFlow column was 96%–101%. The mean calculated PE values were 75%, 88%, 87%, 100%, and 84% for apixaban, dabigatran, dabigratran-13C6, edoxaban, rivaroxaban, and rivaroxaban-13C6, respectively. In the qualitative postcolumn infusion experiments, for the blank matrices tested (10 independent sources), no significant regions of either signal suppression or enhancement were observed near to the retention times of the analytes of interest (see Figure, Supplemental Digital Content 3, http://links.lww.com/TDM/A82, a typical post-column infusion chromatogram for an analyte-free plasma sample). Apixaban, dabigatran, and rivaroxaban all showed good stability over 3 freeze–thaw cycles and after storage at room temperature and at 2–88C for up to 2 weeks (mean values 91%–105%). Edoxaban was also stable over 3 freeze–thaw cycles (91%–105%), but showed a mean deterioration of 16% in 2 weeks at 2–88C, and marked deterioration when stored at room temperature for 24 hours (18%) and 2 weeks (70%). All analytes were stable (mean peak areas 72 hours postextraction .86% of initial mean peak areas) in the post-protein precipitation solutions on storage in vials in the autosampler tray (48C, at least 72 hours). In this evaluation, we achieved .600 injections from a single TurboFlow column without noticeable deterioration in performance.
Clinical Samples The median (range) dabigatran concentration was 132 (67–252) mg/L in samples from 8 patients [median (range) age = 76 (66–89) years] prescribed dabigatran etexilate [median (range) dose = 260 (220–300) mg/d]. The median (range) rivaroxaban concentration was 168 (3–594) mg/L in 2014 Lippincott Williams & Wilkins
samples from 14 patients [median (range) age = 57 (33–75) years], all of whom were prescribed rivaroxaban (20 mg/d). The median (range) time since last dose varied from 15 (2– 64) hours for dabigatran and 12 (2–23) hours for rivaroxaban. A sample from a patient taking apixaban (84-year-old woman, prescribed dose 5 mg/d, sample taken 2 hours postdose) was found to contain 254 mg/L of apixaban. No samples from patients prescribed edoxaban were available. There was general agreement between the results given by the functional assays and by LC-HR-MS (Fig. 3). However, large differences were observed for some samples, and the LC-HR-MS results were generally higher than those given by the functional assays (bias of 211 and 229 mg/L for dabigatran and rivaroxaban, respectively). For rivaroxaban concentrations below 50 mg/L (LC-HR-MS assay), there was a poor correlation with the functional assay results (y = 0.430x + 7.8, R2 = 0.56). Retrospective data interrogation to identify plasma dabigatran metabolites, based on a previous in vivo study by Blech et al,18 showed the presence of 5 dabigatran metabolites in samples from patients prescribed this drug (Fig. 4). It is not known if any of these metabolites are pharmacologically active. The results from the apixaban and the dabigatran overdose patients are shown in Figure 5. For apixaban, data were fitted to a 1-compartment model that yielded rate constants (6SE) of either 0.024 6 0.003 hour21 or 0.021 6 0.001 hour21, giving plasma half-lives of either 31.5 or 33.0 hours, depending on whether the first point was included, a half-life slightly longer than that reported after administration of apixaban tablets in a volunteer study.28 For dabigatran, the LC-HR-MS data were fitted to a 2-compartment model using least squares regression analysis and weighted
601
Ther Drug Monit Volume 36, Number 5, October 2014
Gous et al
TABLE 2. Summary of Intra-assay and Interassay Imprecision and Inaccuracy Data Obtained on Analysis of the Internal Quality Control Solutions (N = 10 at all Concentrations) Analyte Intra-assay Apixaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Dabigatran Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Edoxaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Rivaroxaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Interassay Apixaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Dabigatran Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Edoxaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Rivaroxaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal)
LLoQ (Nominal 1 mg/L)
IQC A (Nominal 25 mg/L)
IQC B (Nominal 75 mg/L)
IQC C (Nominal 400 mg/L)
1.1 2.6 106.1
23.2 2.8 92.8
73.1 2.0 97.5
405.7 3.3 101.4
1.0 4.2 101.9
24.5 5.2 98.0
77.1 5.5 102.8
416.5 3.3 104.1
0.9 17.3 91.1
22.4 9.0 89.6
79.9 5.7 106.5
452.2 3.4 113.1
0.9 9.5 93.5
24.8 3.2 99.2
78.1 2.9 104.1
416.3 3.4 104.1
1.0 15.6 96.1
23.1 3.7 92.4
73.6 2.6 98.2
404.9 4.3 101.2
1.0 15.1 98.6
24.4 6.6 97.5
77.3 5.0 103.1
419.0 5.3 104.8
1.2 21.2 116.0
24.3 7.9 97.1
77.5 4.1 103.3
416.4 2.9 104.1
1.0 11.2 96.0
24.5 3.1 98.1
76.6 7.1 102.1
403.6 3.1 100.9
1/C2. The rate constants were 0.043 6 0.005 hour21 and 0.0117 6 0.0008 hour21, respectively, and the corresponding plasma half-lives were 16.1 and 59.2 hours. There was no evidence of an absorption phase, and because tmax is said to be 1–3 hours after oral dosage,10 the first sample was probably taken several hours postingestion. The initial half-life (5 samples collected over 48 hours) corresponds to the literature value of 12–17 hours,10 but the final elimination phase is obviously much longer.
DISCUSSION Although the analytical validation has shown this method to be suitable for the detection and quantitation of edoxaban in synthetic samples over the calibration range 1–500 mg/L,22 analysis of samples from patients prescribed edoxaban is needed to fully validate the assay. The in vitro stability data presented here suggest that samples for edoxaban analysis should ideally be analyzed as soon as possible
602
after collection, or plasma should be harvested and stored at 2208C before analysis. The overall correlation between the results obtained with LC-HR-MS and with the functional assays (Fig. 5) was not as good as results reported by others when comparing LCMS/MS with different functional assays for dabigatran19,20 and rivaroxaban21 but showed similar results with regards to the poor correlation of LC-MS/MS and functional assay results for rivaroxaban at concentrations below 50 mg/L.21 There were no samples available to evaluate the correlation with LC-HR-MS at low concentrations of dabigatran, although previous studies have emphasized the benefits of using LC-MS for such samples.19,20 Factors that may have contributed to the relatively poor correlations observed in our study include (1) the relatively small number of samples assayed, (2) differences between EDTA plasma and citrated plasma (previous LC-MS/MS comparison studies19–21 used citrated samples for both assays), (3) slight (,2 hours) differences in sample timing between the citrated and EDTA 2014 Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
Measurement of DOACs Using LC-HR-MS
FIGURE 3. Comparison of patient sample results for dabigatran (n = 8) and rivaroxaban (n = 14). A, Scatter plot. Dashed gray line indicates line of identity. Trend lines of dabigatran and rivaroxaban are indicated as dashed black line and solid black line, respectively. B, Bland–Altman plot. Dashed gray line indicates zero bias. Mean bias for dabigatran and rivaroxaban are indicated as dashed black line and solid black line, respectively.
samples, and (4) other factors that may influence the functional assays, such as the influence of sample timing with respect to the last dose for rivaroxaban29 and the possible contribution of active metabolites to the activity of dabigatran.18,30 For dabigatran and rivaroxaban, the LLoQs of the LCHR-MS method (1 mg/L using a 100-mL sample) far exceed those of the functional assays (sensitivity thresholds according to the manufacturers were 20 and 50 mg/L for dabigatran and rivaroxaban, respectively). This may be of particular use when monitoring the clearance of drug following overdose, when assessing adherence, in cases of recurrent thrombosis, before major surgery, or before fibrinolytic therapy of acute ischemic stroke. The LC-HR-MS method can be used for patients prescribed any of the DOACs assayed and does not require prior knowledge as to how each DOAC affects a given functional assay. Moreover, for carrying out the functional assays, strict sampling procedures must be followed to ensure accurate results (citrated sample, plasma harvested within 1 hour of collection, and stored at 2408C before batch analysis). In contrast, the LC-HR-MS method utilizes EDTA 2014 Lippincott Williams & Wilkins
plasma, which means in principle no additional samples are required from patients having full blood count measurements.
Practical Considerations Use of the Aria Transcend TLX-II (duplex) system (Thermo Fisher Scientific) can double sample throughput (6 minutes total TurboFlow analysis time with a 3-minute MS data collection window). The generic applicability of TurboFlow to a range of analytes31 means that it may be possible to add new DOACs, such as betrixaban,24 or metabolites of existing drugs (Fig. 4), to the method if required as reference compounds become available. Furthermore, the approach using LC-HR-MS, rather than the more typical approach to LC-MS/MS analysis, means that no additional tuning of these new compounds, or modifications to the MS method settings, will be necessary. The ability to retrospectively interrogate full-scan MS data may be useful for the identification of drug metabolites, for example. The sample preparation described involves the addition of an IS solution in acetonitrile and a simple protein precipitation. The ISs used were only soluble and stable in organic solvent, and so the protein precipitation
603
Gous et al
Ther Drug Monit Volume 36, Number 5, October 2014
FIGURE 4. Identification of dabigatran metabolites in a sample from a 66-year-old man prescribed 220 mg/d of dabigatran etexilate (LC-HR-MS plasma dabigatran concentration 218 mg/L). Metabolite m/z values and structural formulae (as suggested by Blech et al 16) are shown on the chromatograms, which were filtered by m/z (5 ppm window) from fullscan data.
step was necessary following addition of the ISs. However, despite the fact that direct injection of protein-containing material is possible using TurboFlow,32 this protein precipitation step serves to increase column life and improves the reliability of the assay.33 As regards turnaround time and the availability of the assay on an emergency basis, traditionally, the use of LC-MS in therapeutic drug monitoring has been by batch processing, which is not ideal in the emergency situation. However, use of isotopic internal calibration34 offers the opportunity to deliver results not only with the accuracy,
604
selectivity, and sensitivity of MS but also with the turnaround time and assay availability offered by random-access clinical analyzers.
CONCLUSIONS
The method developed is simple and robust. Only 200 mL of sample is required for a duplicate analysis, and the method is suitable for the measurement of the analytes over the range of concentrations typically encountered in therapy (1–500 mg/L), 2014 Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
FIGURE 5. Plasma apixaban (A) and dabigatran (B) concentrations against time in overdose patients (amounts ingested and times taken unknown). For dabigatran, 2 samples (18 and 64 hours postingestion, respectively) were also analyzed by functional assay (LC-HR-MS: 243 and 67 mg/L, respectively; functional assay: 252 and 62 mg/L, respectively).
a wider range than can be routinely achieved with functional coagulation assays. Moreover, the assay can be used to identify a particular drug if information as to the precise DOAC taken is lacking. ACKNOWLEDGMENTS The authors thank Thermo Fisher Scientific for provision of LC-HR-MS instrumentation and Dr R Whelpton for pharmacokinetic calculations. REFERENCES 1. Kuruvilla M, Gurk-Turner C. A review of warfarin dosing and monitoring. Proc (Bayl Univ Med Cent). 2001;14:305–306. 2. Royal Pharmaceutical Society of Great Britain and British Medical Association. British National Formulary. 66th ed. London: Pharmaceutical Press; 2013. 3. Holford NH. Clinical pharmacokinetics and pharmacodynamics of warfarin. Clin Pharmacokinet. 1986;11:483–504. 4. Baglin T, Keeling D, Kitchen S. Effects on routine coagulation screens and assessment of anticoagulant intensity in patients taking oral dabigatran or rivaroxaban: guidance from the British Committee for Standards in Haematology. Br J Haematol. 2012;159:427–429. 5. Schulman S. Advantages and limitations of the new anticoagulants. J Intern Med. 2014;275:1–11. 6. Ten Cate H. New oral anticoagulants: discussion on monitoring and adherence should start now! Thromb J. 2013;11:8–12. 7. MHRA. Drug safety update–dabigatran. 2011;5:A2. Available at: http:// www.mhra.gov.uk/Safetyinformation/DrugSafetyUpdate/CON137771. Accessed June 24, 2013. 8. Harper P, Young L, Merriman E. Bleeding risk with dabigatran in the frail elderly. N Engl J Med. 2012;366:864–866. 9. Pernod G, Albaladejo P, Godier A, et al. Management of major bleeding complications and emergency surgery in patients on long-term treatment with direct oral anticoagulants, thrombin or factor-Xa inhibitors. Proposals of the Working Group on Perioperative Haemostasis (GIHP). Ann Fr Anesth Reanim. 2013;32:691–700. 10. Gong IY, Kim RB. Importance of pharmacokinetic profile and variability as determinants of dose and response to dabigatran, rivaroxaban, and apixaban. Can J Cardiol. 2013;29(7 suppl):S24–S33. 11. Mendell J, Zahir H, Matsushima N, et al. Drug-drug interaction studies of cardiovascular drugs involving P-glycoprotein, an efflux transporter, on the pharmacokinetics of edoxaban, an oral factor Xa inhibitor. Am J Cardiovasc Drugs. 2013;13:331–342.
2014 Lippincott Williams & Wilkins
Measurement of DOACs Using LC-HR-MS
12. van Ryn J, Strangler J, Haertter S, et al. Dabigatran etexilate—a novel, reversible, oral direct thrombin inhibitor: interpretation of coagulation assays and reversal of anticoagulant activity. Thromb Haemostas. 2010;103:1116–1127. 13. Douxfils J, Mullier F, Loosen C, et al. Assessment of the impact of rivaroxaban on coagulation assays: laboratory recommendations for the monitoring of rivaroxaban and review of the literature. Thromb Res. 2012;130:956–966. 14. Douxfils J, Muller F, Robert S, et al. Impact of dabigatran on a large panel of routine or specific coagulation assays—laboratory recommendations for monitoring of dabigatran etexilate. Thromb Haemostas. 2012;107:985–997. 15. Samama MM, Amiral J, Guinet C, et al. Monitoring plasma levels of factor Xa inhibitors: how, why and when? Expert Rev Hematol. 2013;6:155–164. 16. Douxfils J, Chatelain C, Chatelain B, et al. Impact of apixaban on routine and specific coagulation assays: a practical laboratory guide. Thromb Haemost. 2013;110:283–294. 17. Rohde G. Determination of rivaroxaban—a novel, oral, direct factor Xa inhibitor—in human plasma by high performance liquid chromatography–tandem mass spectrometry. J Chromatogr B. 2008;872:43–50. 18. Blech S, Ebner T, Ludwig-Schwellinger E, et al. The metabolism and disposition of the oral direct thrombin inhibitor, dabigatran, in humans. Drug Metab Dispos. 2008;36:386–399. 19. Antovic JP, Skeppholm M, Eintrei J, et al. Evaluation of coagulation assays versus LC-MS/MS for determinations of dabigatran concentrations in plasma. Eur J Clin Pharmacol. 2013;69:1875–1881. 20. Douxfils J, Dogne JM, Mullier F, et al. Comparison of calibrated dilute thrombin time and aPTT tests with LC-MS/MS for the therapeutic drug monitoring of patients treated with dabigatran etexilate. Thromb Haemost. 2013;110:543–549. 21. Douxfils J, Tamigniau A, Chatelain B, et al. Comparison of calibrated chromogenic anti-Xa assay and PT tests with LC-MS/MS for the therapeutic drug monitoring of patients treated with rivaroxaban. Thromb Haemost. 2013;110:723–731. 22. Bathala MS, Masumoto H, Oguma T, et al. Pharmacokinetics, biotransformation, and mass balance of edoxaban, a selective, direct factor Xa inhibitor, in humans. Drug Metab Dispos. 2012;40:2250–2255. 23. Delavenne X, Mismetti P, Basset T. Rapid determination of apixaban concentration in human plasma by liquid chromatography/tandem mass spectrometry: application to pharmacokinetic study. J Pharm Biomed Anal. 2013;78–79:150–153. 24. Palladino M, Merli G, Thomson L. Evaluation of the oral direct factor Xa inhibitor—betrixaban. Expert Opin Investig Drugs. 2013;22:1465–1472. 25. Food and Drug Administration/Center for Drug Evaluation and Research (FDA/CDER). Guidance for industry. Bioanal Method Validation. 2001. Available at: http://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/UCM070107. pdf. Accessed September 1, 2013. 26. Polettini A. Applications of LC-MS in Toxicology. London: Pharmaceutical Press; 2006. 27. Bonfiglio R, King RC, Olah TV, et al. The effects of sample preparation on the variability of electrospray ionization response for model drug compounds. Rapid Commun Mass Spectrom. 1999;13:1175–1185. 28. Frost C, Wang J, Nepal S, et al. Apixaban, an oral, direct factor Xa inhibitor: single dose safety, pharmacokinetics, pharmacodynamics and food effect in healthy subjects. Br J Clin Pharmacol. 2013;75:476–487. 29. Mani H, Hesse C, Stratmann G, et al. Rivaroxaban differentially influences ex vivo coagulation assays based on the administration time. Thromb Haemost. 2011;106:156–164. 30. Ebner T, Wagner K, Wienen W. Dabigatran acylglucuronide, the major human metabolite of dabigatran: in vitro formation, stability, and pharmacological activity. Drug Metab Dispos. 2010;38:1567–1575. 31. Herman JL. Generic approach to high throughput ADME screening for lead candidate optimization. Int J Mass Spectrom. 2004;238:107–117. 32. Couchman L. Turbulent flow chromatography in bioanalysis: a review. Biomed Chromatogr. 2012;26:892–905. 33. He X, Kozak M. Development of a liquid chromatography-tandem mass spectrometry method for plasma-free metanephrines with ion-pairing turbulent flow online extraction. Anal Bioanal Chem. 2012;403:3003–3010. 34. Couchman L, Belsey SL, Handley SA, et al. A novel approach to quantitative LC-MS/MS: therapeutic drug monitoring of clozapine and norclozapine using isotopic internal calibration. Anal Bioanal Chem. 2013; 405:9455–9466.
605
Measurement of the Direct Oral Anticoagulants Apixaban, Dabigatran, Edoxaban, and Rivaroxaban in Human Plasma Using Turbulent Flow Liquid Chromatography With High-Resolution Mass Spectrometry Tracey Gous, MSc,* Lewis Couchman, MSc,*† Jignesh P. Patel, PhD,†‡ Chitongo Paradzai, MSc,‡ Roopen Arya, PhD,‡ and Robert J. Flanagan, PhD*
Background: Direct oral anticoagulants (DOACs) are prescribed
(2 weeks) and of 18% and 70% (1 day and 2 weeks, respectively) at room temperature.
for systemic anticoagulation. Fixed doses are recommended, but dose individualization may be warranted. Functional coagulation assays may be available, but their use requires knowledge of the drug taken. To provide alternative methodology for guiding dosage, we have developed and validated a liquid chromatography–mass spectrometric assay for apixaban, dabigatran, edoxaban, and rivaroxaban at the concentrations attained during therapy.
Conclusions: The method is suitable for high-throughput therapeutic drug monitoring of DOACs. The acquisition of full scan data allows for the retrospective identification of metabolites. The method can be used to identify a particular DOAC if information on the drug taken is lacking.
Methods: Samples, calibrators, and internal quality controls (100 mL) were mixed with internal standard solution (50 mg/L both dabigatran-13C6 and rivaroxaban-13C6 in acetonitrile) and, after centrifugation (16,400g, 4 minutes), supernatant (100 mL) was injected onto a Cyclone-C18-P-XL TurboFlow column. Analytes were focused onto an Accucore PhenylHexyl (2.1 · 100 mm, 2.6 mm) analytical column and eluted using a methanol + acetonitrile (1 + 1): aqueous ammonium acetate (10 mmol/L) gradient. Data were acquired using high-resolution mass spectrometry in full-scan mode (100–2000 m/z) with data-dependent fragmentation to confirm peak identity. Calibration was linear (1–500 mg/L all analytes).
Results: Total analysis time was 6 minutes. Intra-assay imprecision (% RSD) at 1 mg/L was 2.6%, 4.2%, 17.3%, and 9.5% for apixaban, dabigatran, edoxaban, and rivaroxaban, respectively. Mean recovery was 96%–101%. No signal suppression or enhancement was observed. Apixaban, dabigatran, and rivaroxaban were stable over 3 freeze–thaw cycles, after storage at room temperature, and at 2–88C for up to 2 weeks. Edoxaban was stable over 3 freeze–thaw cycles but showed a mean deterioration of 16% if stored at 2–88C
Received for publication November 14, 2013; accepted January 29, 2014. From the *Toxicology Unit, Department of Clinical Biochemistry, King’s College Hospital NHS Foundation Trust; †Institute of Pharmaceutical Science, King’s College London; and ‡King’s Thrombosis Centre, Department of Haematological Medicine, King’s College Hospital NHS Foundation Trust, London, United Kingdom. The authors declare no conflict of interest. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.drug-monitoring.com). Correspondence: Tracey Gous, MSc, Toxicology Unit, Department of Clinical Biochemistry, 3rd Floor, Bessemer Wing, King’s College Hospital NHS Foundation Trust, Denmark Hill, London SE5 9RS, United Kingdom (e-mail: [email protected]). Copyright © 2014 by Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
Key Words: liquid chromatography–high-resolution mass spectrometry, therapeutic drug monitoring, direct oral anticoagulants, dabigatran, rivaroxaban, apixaban, edoxaban (Ther Drug Monit 2014;36:597–605)
INTRODUCTION Oral anticoagulation is prescribed for millions of patients worldwide. For more than 50 years, vitamin K antagonists (VKAs) such as warfarin have been used for this purpose. However, use of VKAs is not straightforward and frequent testing of the international normalized ratio (INR) is needed to optimize dosage.1 Recently, direct oral anticoagulants (DOACs; Fig. 1) have become available. As of end-2013, 3 DOACs [apixaban (Eliquis; Bristol-Myers Squibb), dabigatran etexilate (Pradaxa; Boehringer Ingelheim), and rivaroxaban (Xarelto; Bayer)] were licensed in the United Kingdom for clinical use.2 [Note: Dabigatran etexilate is a pro-drug, the active moiety dabigatran (Fig. 1) being rapidly formed in vivo.] DOACs selectively inhibit coagulation serine proteases (Xa or IIa). They have a rapid onset of action, a relatively predictable pharmacokinetic profile, and a relatively short plasma half-life, making initiation, maintenance, and discontinuation of anticoagulant therapy considerably easier than with VKAs. However, at present, there is no effective antidote to immediately reverse anticoagulation should the need arise. With VKAs, the degree of anticoagulation is titrated to the patient’s requirement (usually an INR of 2–3), and thus, therapy is highly individualized, taking into account age, concomitant drug therapy, diet, alcohol intake, and body weight.3 However, with DOACs, a “one-size fits all” approach has been adopted, with dose adjustments suggested should the patient meet certain criteria, for example, dose reduction with dabigatran when renal function is impaired.
597
Gous et al
Ther Drug Monit Volume 36, Number 5, October 2014
plasma DOAC concentrations.9 For these reasons and for use in pharmacokinetic studies, methods have been developed and validated for the therapeutic drug monitoring of DOACs in plasma using liquid chromatography–tandem mass spectrometry (LC-MS/MS).17–23 The plasma concentrations of apixaban attained during therapy are reported to range up to 300 mg/L,16 whereas for dabigatran and rivaroxaban, the limits above which there is an increased risk bleeding are thought to be 200 and 240 mg/L, respectively.19,21 To monitor DOAC concentrations in patient samples and give the option of measuring metabolites of potential interest,18 we have developed a simple method to measure plasma apixaban, dabigatran, rivaroxaban, and a further DOAC edoxaban (Fig. 1), using TurboFlow sample preparation coupled with high-resolution LC-MS (LC-HR-MS). The method should prove able to incorporate further DOACs, such as betrixaban,24 as they are introduced clinically.
MATERIALS AND METHODS Chemicals and Reagents FIGURE 1. Structural formulae of apixaban, dabigratran, edoxaban, and rivaroxaban. The positions of the isotopically labelled 13C aromatic rings (for dabigatran-13C6 and rivaroxaban-13C6) are indicated (*).
However, the British Society of Haematology and the International Society of Thrombosis and Haemostasis have stipulated that measurement of DOAC activity might be warranted: (1) after overdose (accidental or otherwise), (2) in the elderly, especially those with comorbidities, (3) when patients switch from an existing oral anticoagulant, (4) in those at the extremes of body weight, (5) in patients with hepatic or renal impairment, (6) in patients co-prescribed other medications where there is a risk of drug–DOAC interactions, (7) to assess adherence, and (8) to assess anticoagulant activity before major surgery.4–11 There is at present no consensus on the best methodology for assessing DOAC activity in vivo and hence guiding dosage. Unlike with VKAs, traditional coagulation tests, such as the INR, cannot be used to monitor the effects of DOACs. This is because different DOACs have different sensitivity to coagulation assays, such as activated partial thromboplastin time, prothrombin time, thrombin time, dilute thrombin time, ecarin clotting time, and drug-specific anti-Xa assays.12,13 Assay sensitivity can also depend on the reagent/analyzer combination used.13–16 Despite this, different functional coagulation assays have been tested and calibrated to individual DOACs to estimate plasma DOAC concentrations in patient samples. As the number of DOACs available for clinical use is set to increase, it is possible that in some situations, for example, if the patient presents in coma, it will be not be known which DOAC the patient has been taking. Furthermore, because it has been suggested that plasma DOAC concentrations should be ,30 mg/L before major surgery, it is especially important to have good assay performance at lower
598
Dabigatran and rivaroxaban were from LGC Standards (Teddington, United Kingdom). Apixaban, edoxaban, dabigatran-13C6, and rivaroxaban-13C6 were from ALSA-CHIM (Illkirch, France). MS grade formic acid and ammonium acetate were from Fluka (Poole, United Kingdom). HPLC grade methanol, acetonitrile, and acetone were from Sigma–Aldrich (Poole, United Kingdom). AnalaR grade 2-propanol and concentrated hydrochloric acid were from VWR (Lutterworth, United Kingdom). Water was deionized (18 mV; Elga, Marlow, United Kingdom), and analyte-free pooled human plasma was from Sera Labs (Haywards Heath, United Kingdom).
Calibrators, Internal Quality Controls, and Internal Standard Solution Separate calibrator and internal quality control (IQC) stock solutions were prepared for dabigatran (200 mg/L in 0.1 mol/L aqueous hydrochloric acid), rivaroxaban (100 mg/L in acetonitrile), apixaban (400 mg/L in acetonitrile), and edoxaban (100 mg/L in acetonitrile). Calibrator working stock solutions were prepared at concentrations of 10 and 1 mg/L (all analytes) by dilution of individual stock solutions in acetonitrile. IQC stock and working solutions were prepared separately. To prepare calibrators and IQC solutions, appropriate volumes of working stock solutions were pipetted into volumetric glassware, evaporated to dryness under a gentle stream of nitrogen, and made up to volume with analyte-free human plasma. Calibrators were prepared containing each analyte at 1, 5, 10, 50, 100, and 500 mg/L, and IQC solutions were similarly prepared at 25, 75, and 400 mg/L. Prepared calibrators and IQC solutions were mixed by inversion and allowed to equilibrate (48C, 24 hours), before storage in approximately 1 mL portions at 2208C in 2-mL screw-top polypropylene tubes (Alpha Laboratories, Eastleigh, United Kingdom). To prepare the internal standard (IS) solution, individual stock solutions of dabigatran-13C6 (100 mg/L in 0.1 mol/L 2014 Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
aqueous hydrochloric acid) and rivaroxaban-13C6 (100 mg/L in acetonitrile) were prepared and stored at 2208C. Working IS solution (50 mg/L both dabigatran-13C6 and rivaroxaban-13C6) was prepared as necessary by appropriate dilution from the IS stock solutions with acetonitrile.
Sample Preparation
Sample, calibrator, or IQC solution (100 mL) was added to a 60 · 7.5-mm i.d. glass tube (Dreyer tube; Esslab, Essex, United Kingdom) and working IS solution (200 mL) was added. The contents of the tube were vortex mixed (30 seconds) and centrifuged (16,400g, 4 minutes) and the supernatant was transferred to an autosampler vial (2-mL glass vial with 300-mL glass insert; Kinesis, St Neots, United Kingdom) and capped for analysis.
High-Resolution LC-MS An Aria Transcend TLX-II system (Thermo Fisher Scientific, San Jose, CA) consisting of 4 Accela 1250 highpressure quaternary pumps, valve interface module, and CTC PAL autosampler was used with a Q Exactive MS (Thermo Fisher Scientific, Bremen, Germany). The autosampler tray was maintained at +108C. Instrument control was performed using TraceFinder software (version 3.1; Thermo Fisher Scientific). System eluents were as follows: (A) 0.1% (vol/vol) formic acid in 10 mmol/L aqueous ammonium acetate, (B) 0.1% (vol/vol) formic acid in methanol:acetonitrile (1 + 1), and (C) acetonitrile:acetone:2-propanol (2 + 1 + 2). Prepared samples (100 mL) were injected onto a TurboFlow column (Cyclone C18-P-XL, 50 · 0.5-mm i.d., flow rate 2.0 mL/min). Retained analytes were eluted (reversed flow direction) from the TurboFlow column using elution solvent [eluent A:eluent B, 1 + 4 (vol/vol), 200 mL], stored in the holding loop, and focused through a T-piece onto an Accucore PhenylHexyl analytical column (2.7 mm average particle size, 100 · 2.1 mm i.d., Thermo Fisher Scientific, Runcorn, United Kingdom). The column was fitted with a 0.5-mm precolumn filter (Fisher Scientific, Loughborough, United Kingdom) and was maintained at 408C (HotPocket; Thermo Fisher Scientific). During gradient elution (total flow rate 0.30 mL/min) from the analytical column, the TurboFlow column was washed in turn (reversed flow direction) with
Measurement of DOACs Using LC-HR-MS
elution solvent and with wash solvent (eluent C). The whole system was then re-equilibrated before the next injection. The gradient elution and valve switching profiles are summarized in Table 1. The total analysis time was 6 minutes, including column re-equilibration. Eluent flow was diverted from the MS to waste for the first 3 minutes following each TurboFlow injection, and data were acquired for 3 minutes per analysis. MS detection was carried out in positive ionization mode using heated electrospray ionization [spray voltage 4 kV; temperatures: vaporizer 3008C; capillary 2508C; auxiliary, sheath, and sweep gases 55, 5, and 0 (arbitrary units), respectively, S-lens voltage 55 V]. Data were acquired in fullscan mode (70,000 resolution, range: 150–1000 m/z) with single data–dependent fragmentation (MS2) scans (17,500 resolution, stepped normalized collision energy 35% 6 50%, range: 150–1000 m/z) collected for each analyte during peak elution (based on a defined inclusion list) used to confirm peak identity. Data acquisition was performed using TraceFinder software.
Assay Calibration and Data Processing Quantitation was based on extracted accurate-mass full scan data [5 ppm window, external mass calibration (alternate days) using positive/negative ion calibration infusion solutions (Thermo Fisher Scientific)]. Calibration curves were generated by plotting the peak area ratio (analyte/IS) from the extracted ion chromatograms against analyte concentration. For apixaban and rivaroxaban, rivaroxaban-13C6 was used as the IS, and for dabigatran and edoxaban, dabigatran-13C6 was used. Lines were fitted by least squares regression (1/x weighting, not including or forced through the origin). Calibrators and matrix blanks were included at the beginning and end of each batch analysis, with IQCs included (1) after the first set of calibrators and immediately before the last set and (2) after every 10 injections throughout the sequence. Patient samples were analyzed in duplicate and the mean result taken. Samples with concentrations exceeding the calibration range for an analyte were diluted as appropriate with analyte-free pooled human plasma and re-assayed. Assay acceptance criteria were (1) linear (R2 . 0.98) calibration curves for each analyte and (2) IQC values within 615% nominal concentrations for all analytes.
TABLE 1. Summary of TurboFlow Method Loading (TurboFlow) Pump
Valves
Eluting (Analytical) Pump
Start Flow Flow Step Time (min) (mL/min) Gradient % Eluent A % Eluent B % Eluent C Tee Loop (mL/min) Gradient % Eluent A % Eluent B 1 2 3 4 5 6 7
0.00 0.50 1.50 2.00 3.00 3.50 4.50
2.00 0.10 1.00 1.00 1.00 1.50 2.00
Step Step Step Step Step Step Step
100 100 — — — 20 100
— — 100 100 — 80 —
— — — — 100 — —
Out In Out Out Out Out Out
Out In In In In In Out
0.50 0.50 0.30 0.30 0.30 0.30 0.50
Step Step Step Ramp Step Step Step
98 98 98 — — — 98
2 2 2 100 100 100 2
Step 1, sample loading; step 2, transfer of retained analytes to the analytical column; steps 3–5, gradient elution (analytical column) and column washing (TurboFlow column); step 6, elution loop filling (loading pump); step 7, column equilibration (1.5 minutes).
2014 Lippincott Williams & Wilkins
599
Gous et al
Method Validation The method was validated using the US Food and Drug Administration Center for Drug Evaluation and Research guidance for bioanalytical method validation.25 Intra-assay and interassay imprecision (% RSD) and inaccuracy were measured by replicate analysis of the IQC solutions and a solution prepared containing all analytes at the lower limit of quantification (LLoQ, 1 mg/L) on the same day and on different days, respectively. To investigate the overall process efficiency (PE) for the protein precipitation and TurboFlow steps,26 2 sets of solutions containing all analytes and ISs were prepared (for each set, 1 solution at a concentration of 200 mg/L, and 1 at 500 mg/L all analytes). One set of solutions was prepared in eluent A and the other set was prepared in analyte-free pooled human plasma. To assess absolute analyte recovery from the TurboFlow column, after diluting the aqueous solutions with eluent A (1 + 2, vol/vol) instead of working IS solution, these prepared solutions were (1) analyzed using the TurboFlow procedure and (2) analyzed with the TurboFlow system bypassed (ie, injection was performed directly onto the analytical column). The mean peak areas of each analyte analyzed using the complete TurboFlow procedure were compared with those observed from injection onto the analytical column only, which was assumed to represent 100% recovery. To assess the overall PE, mean peak areas for plasma solutions analyzed using the complete TurboFlow system (n = 3 at each concentration) were compared with those from aqueous solutions analyzed with the TurboFlow system bypassed. To further investigate ion suppression, analyte-free plasma from 10 independent sources was analyzed following treatment with acetonitrile instead of working IS solution. The detector response for each analyte was monitored while a solution containing all analytes (10 mg/L in acetonitrile, 10 mL/ min) was infused by syringe post-column.27 The LLoQ was taken as the lowest concentration (to the nearest 1 mg/L) where the intra-assay % RSD was less than 20. Carryover was assessed by running high IQCs (n = 3) followed immediately by low IQCs (n = 3) and comparing measured with nominal concentrations. Analyte stability in the IQC solutions was assessed (1) at room temperature over 14 days, (2) in a refrigerator (2–88C) over 14 days, and (3) through 3 freeze– thaw cycles, by comparison of mean peak areas (n = 3 at each concentration) against all the results obtained on analysis of IQC solutions that had been stored at 2208C before analysis (ie, had been subjected to a single freeze–thaw cycle). Autosampler tray stability was assessed using the IQC solutions and samples from patients treated with apixaban, dabigatran, or rivaroxaban. Prepared samples were analyzed at 0, 6, 12, and 72 hours, and mean peak areas (n = 2 at each time point) for all analytes and ISs were compared.
Patient Samples Dabigatran and Rivaroxaban Functional Coagulation Assays Patients requiring assessment of anticoagulant activity had blood drawn into 3.0-mL tubes containing 0.3 mL of 3.2% (wt/vol) aqueous trisodium citrate (Vacuette; Greiner
600
Ther Drug Monit Volume 36, Number 5, October 2014
Bio One, Stonehouse, United Kingdom). All tubes were filled to the draw-line indicator to ensure consistent dilution with citrate buffer and centrifuged [2 · 7 minutes, 1440g; 38R Rotina (Hettich, Tuttligen, Germany)] within 1 hour of sample collection. The diluted plasma was harvested and stored at 2408C before analysis, which was undertaken within 1 week of sample collection. Dabigatran concentrations were estimated using an ecarin chromogenic assay (HaemoSys, Jena, Germany) with dabigatran calibrators (50, 250, and 500 mg/L, Lot No. 22202) and controls (low and high, Lot No. 03001; Hyphen-Biomed, Neuville-Sur-Oise, France) supplied in lyophilized human plasma and reconstituted in deionized water according to the manufacturer’s instructions. Rivaroxaban concentrations were estimated using an STAliquid anti-Xa assay (Diagnostica Stago, Asnieres, France) with rivaroxaban calibrators (0, 100, 250, and 450 mg/L, Lot No. 281), and low and high controls (Diagnostica Stago, Lot No. 110838) were supplied in lyophilized human plasma and reconstituted in deionized water according to the manufacturer’s instructions. Analyses were conducted on the STA-R evolution analyzer (Diagnostica Stago).
LC-HR-MS Assay Excess, anonymized blood (EDTA-coated anticoagulant; Vacuette; Greiner) from patients prescribed dabigratran etexilate or rivaroxaban that had been submitted for hematological monitoring (full blood count) were analyzed after separation of plasma and storage at 2208C. Assays were performed in duplicate and the mean result taken. EDTA samples were collected within 2 hours of the samples used for functional assays. EDTA plasma samples from patients thought to have taken overdoses (times of ingestion and amounts taken not known) of dabigatran (67-year-old man) and apixaban (77-year-old woman) were also analyzed. Pharmacokinetic data were calculated using GraphPad Prism (Version 3.0).
RESULTS Method Performance Typical chromatograms are shown in Figure 2. The major product ions shown in the MS2 spectrum (see Figure, Supplemental Digital Content 1, http://links.lww.com/TDM/A80, which shows an example of a data-dependent MS2 spectrum generated for dabigatran) were in agreement with the product ions used in published methods using selected reaction monitoring–based LC-MS/MS instrumentation for all analytes assay.17–21,23 Intra-assay and interassay imprecision and inaccuracy are summarized in Table 2. Calibration was linear (R2 . 0.99) for all analytes across the range 1–500 mg/L. For all analytes and ISs, although the m/z extraction window was set to 65 ppm for data processing and quantitation, the measured mass accuracy (using daily external mass calibration) was typically less than 2 ppm (see Table, Supplemental Digital Content 2, http://links.lww.com/TDM/A81, which provides typical mass accuracy data for all analytes). No carryover was observed. Mean absolute recovery (n = 3 at each of the 2 concentrations 2014 Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
Measurement of DOACs Using LC-HR-MS
FIGURE 2. Typical extracted ion chromatogram. Extraction window 6 5 ppm based on theoretical m/z. Chromatograms: calibrator (10 mg/L each analyte), IQC solution (25 mg/L each analyte), and plasma sample from a patient prescribed rivaroxaban (20 mg/d, rivaroxaban concentration 199 mg/L). (i) Dabigatran, (ii) dabigatran-13C6, (iii) rivaroxaban, (iv) rivaroxaban-13C6, (v) edoxaban, and (vi) apixaban.
studied, solutions prepared in eluent A) for all analytes on the TurboFlow column was 96%–101%. The mean calculated PE values were 75%, 88%, 87%, 100%, and 84% for apixaban, dabigatran, dabigratran-13C6, edoxaban, rivaroxaban, and rivaroxaban-13C6, respectively. In the qualitative postcolumn infusion experiments, for the blank matrices tested (10 independent sources), no significant regions of either signal suppression or enhancement were observed near to the retention times of the analytes of interest (see Figure, Supplemental Digital Content 3, http://links.lww.com/TDM/A82, a typical post-column infusion chromatogram for an analyte-free plasma sample). Apixaban, dabigatran, and rivaroxaban all showed good stability over 3 freeze–thaw cycles and after storage at room temperature and at 2–88C for up to 2 weeks (mean values 91%–105%). Edoxaban was also stable over 3 freeze–thaw cycles (91%–105%), but showed a mean deterioration of 16% in 2 weeks at 2–88C, and marked deterioration when stored at room temperature for 24 hours (18%) and 2 weeks (70%). All analytes were stable (mean peak areas 72 hours postextraction .86% of initial mean peak areas) in the post-protein precipitation solutions on storage in vials in the autosampler tray (48C, at least 72 hours). In this evaluation, we achieved .600 injections from a single TurboFlow column without noticeable deterioration in performance.
Clinical Samples The median (range) dabigatran concentration was 132 (67–252) mg/L in samples from 8 patients [median (range) age = 76 (66–89) years] prescribed dabigatran etexilate [median (range) dose = 260 (220–300) mg/d]. The median (range) rivaroxaban concentration was 168 (3–594) mg/L in 2014 Lippincott Williams & Wilkins
samples from 14 patients [median (range) age = 57 (33–75) years], all of whom were prescribed rivaroxaban (20 mg/d). The median (range) time since last dose varied from 15 (2– 64) hours for dabigatran and 12 (2–23) hours for rivaroxaban. A sample from a patient taking apixaban (84-year-old woman, prescribed dose 5 mg/d, sample taken 2 hours postdose) was found to contain 254 mg/L of apixaban. No samples from patients prescribed edoxaban were available. There was general agreement between the results given by the functional assays and by LC-HR-MS (Fig. 3). However, large differences were observed for some samples, and the LC-HR-MS results were generally higher than those given by the functional assays (bias of 211 and 229 mg/L for dabigatran and rivaroxaban, respectively). For rivaroxaban concentrations below 50 mg/L (LC-HR-MS assay), there was a poor correlation with the functional assay results (y = 0.430x + 7.8, R2 = 0.56). Retrospective data interrogation to identify plasma dabigatran metabolites, based on a previous in vivo study by Blech et al,18 showed the presence of 5 dabigatran metabolites in samples from patients prescribed this drug (Fig. 4). It is not known if any of these metabolites are pharmacologically active. The results from the apixaban and the dabigatran overdose patients are shown in Figure 5. For apixaban, data were fitted to a 1-compartment model that yielded rate constants (6SE) of either 0.024 6 0.003 hour21 or 0.021 6 0.001 hour21, giving plasma half-lives of either 31.5 or 33.0 hours, depending on whether the first point was included, a half-life slightly longer than that reported after administration of apixaban tablets in a volunteer study.28 For dabigatran, the LC-HR-MS data were fitted to a 2-compartment model using least squares regression analysis and weighted
601
Ther Drug Monit Volume 36, Number 5, October 2014
Gous et al
TABLE 2. Summary of Intra-assay and Interassay Imprecision and Inaccuracy Data Obtained on Analysis of the Internal Quality Control Solutions (N = 10 at all Concentrations) Analyte Intra-assay Apixaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Dabigatran Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Edoxaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Rivaroxaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Interassay Apixaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Dabigatran Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Edoxaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal) Rivaroxaban Mean measured (mg/L) RSD (%) Inaccuracy (% nominal)
LLoQ (Nominal 1 mg/L)
IQC A (Nominal 25 mg/L)
IQC B (Nominal 75 mg/L)
IQC C (Nominal 400 mg/L)
1.1 2.6 106.1
23.2 2.8 92.8
73.1 2.0 97.5
405.7 3.3 101.4
1.0 4.2 101.9
24.5 5.2 98.0
77.1 5.5 102.8
416.5 3.3 104.1
0.9 17.3 91.1
22.4 9.0 89.6
79.9 5.7 106.5
452.2 3.4 113.1
0.9 9.5 93.5
24.8 3.2 99.2
78.1 2.9 104.1
416.3 3.4 104.1
1.0 15.6 96.1
23.1 3.7 92.4
73.6 2.6 98.2
404.9 4.3 101.2
1.0 15.1 98.6
24.4 6.6 97.5
77.3 5.0 103.1
419.0 5.3 104.8
1.2 21.2 116.0
24.3 7.9 97.1
77.5 4.1 103.3
416.4 2.9 104.1
1.0 11.2 96.0
24.5 3.1 98.1
76.6 7.1 102.1
403.6 3.1 100.9
1/C2. The rate constants were 0.043 6 0.005 hour21 and 0.0117 6 0.0008 hour21, respectively, and the corresponding plasma half-lives were 16.1 and 59.2 hours. There was no evidence of an absorption phase, and because tmax is said to be 1–3 hours after oral dosage,10 the first sample was probably taken several hours postingestion. The initial half-life (5 samples collected over 48 hours) corresponds to the literature value of 12–17 hours,10 but the final elimination phase is obviously much longer.
DISCUSSION Although the analytical validation has shown this method to be suitable for the detection and quantitation of edoxaban in synthetic samples over the calibration range 1–500 mg/L,22 analysis of samples from patients prescribed edoxaban is needed to fully validate the assay. The in vitro stability data presented here suggest that samples for edoxaban analysis should ideally be analyzed as soon as possible
602
after collection, or plasma should be harvested and stored at 2208C before analysis. The overall correlation between the results obtained with LC-HR-MS and with the functional assays (Fig. 5) was not as good as results reported by others when comparing LCMS/MS with different functional assays for dabigatran19,20 and rivaroxaban21 but showed similar results with regards to the poor correlation of LC-MS/MS and functional assay results for rivaroxaban at concentrations below 50 mg/L.21 There were no samples available to evaluate the correlation with LC-HR-MS at low concentrations of dabigatran, although previous studies have emphasized the benefits of using LC-MS for such samples.19,20 Factors that may have contributed to the relatively poor correlations observed in our study include (1) the relatively small number of samples assayed, (2) differences between EDTA plasma and citrated plasma (previous LC-MS/MS comparison studies19–21 used citrated samples for both assays), (3) slight (,2 hours) differences in sample timing between the citrated and EDTA 2014 Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
Measurement of DOACs Using LC-HR-MS
FIGURE 3. Comparison of patient sample results for dabigatran (n = 8) and rivaroxaban (n = 14). A, Scatter plot. Dashed gray line indicates line of identity. Trend lines of dabigatran and rivaroxaban are indicated as dashed black line and solid black line, respectively. B, Bland–Altman plot. Dashed gray line indicates zero bias. Mean bias for dabigatran and rivaroxaban are indicated as dashed black line and solid black line, respectively.
samples, and (4) other factors that may influence the functional assays, such as the influence of sample timing with respect to the last dose for rivaroxaban29 and the possible contribution of active metabolites to the activity of dabigatran.18,30 For dabigatran and rivaroxaban, the LLoQs of the LCHR-MS method (1 mg/L using a 100-mL sample) far exceed those of the functional assays (sensitivity thresholds according to the manufacturers were 20 and 50 mg/L for dabigatran and rivaroxaban, respectively). This may be of particular use when monitoring the clearance of drug following overdose, when assessing adherence, in cases of recurrent thrombosis, before major surgery, or before fibrinolytic therapy of acute ischemic stroke. The LC-HR-MS method can be used for patients prescribed any of the DOACs assayed and does not require prior knowledge as to how each DOAC affects a given functional assay. Moreover, for carrying out the functional assays, strict sampling procedures must be followed to ensure accurate results (citrated sample, plasma harvested within 1 hour of collection, and stored at 2408C before batch analysis). In contrast, the LC-HR-MS method utilizes EDTA 2014 Lippincott Williams & Wilkins
plasma, which means in principle no additional samples are required from patients having full blood count measurements.
Practical Considerations Use of the Aria Transcend TLX-II (duplex) system (Thermo Fisher Scientific) can double sample throughput (6 minutes total TurboFlow analysis time with a 3-minute MS data collection window). The generic applicability of TurboFlow to a range of analytes31 means that it may be possible to add new DOACs, such as betrixaban,24 or metabolites of existing drugs (Fig. 4), to the method if required as reference compounds become available. Furthermore, the approach using LC-HR-MS, rather than the more typical approach to LC-MS/MS analysis, means that no additional tuning of these new compounds, or modifications to the MS method settings, will be necessary. The ability to retrospectively interrogate full-scan MS data may be useful for the identification of drug metabolites, for example. The sample preparation described involves the addition of an IS solution in acetonitrile and a simple protein precipitation. The ISs used were only soluble and stable in organic solvent, and so the protein precipitation
603
Gous et al
Ther Drug Monit Volume 36, Number 5, October 2014
FIGURE 4. Identification of dabigatran metabolites in a sample from a 66-year-old man prescribed 220 mg/d of dabigatran etexilate (LC-HR-MS plasma dabigatran concentration 218 mg/L). Metabolite m/z values and structural formulae (as suggested by Blech et al 16) are shown on the chromatograms, which were filtered by m/z (5 ppm window) from fullscan data.
step was necessary following addition of the ISs. However, despite the fact that direct injection of protein-containing material is possible using TurboFlow,32 this protein precipitation step serves to increase column life and improves the reliability of the assay.33 As regards turnaround time and the availability of the assay on an emergency basis, traditionally, the use of LC-MS in therapeutic drug monitoring has been by batch processing, which is not ideal in the emergency situation. However, use of isotopic internal calibration34 offers the opportunity to deliver results not only with the accuracy,
604
selectivity, and sensitivity of MS but also with the turnaround time and assay availability offered by random-access clinical analyzers.
CONCLUSIONS
The method developed is simple and robust. Only 200 mL of sample is required for a duplicate analysis, and the method is suitable for the measurement of the analytes over the range of concentrations typically encountered in therapy (1–500 mg/L), 2014 Lippincott Williams & Wilkins
Ther Drug Monit Volume 36, Number 5, October 2014
FIGURE 5. Plasma apixaban (A) and dabigatran (B) concentrations against time in overdose patients (amounts ingested and times taken unknown). For dabigatran, 2 samples (18 and 64 hours postingestion, respectively) were also analyzed by functional assay (LC-HR-MS: 243 and 67 mg/L, respectively; functional assay: 252 and 62 mg/L, respectively).
a wider range than can be routinely achieved with functional coagulation assays. Moreover, the assay can be used to identify a particular drug if information as to the precise DOAC taken is lacking. ACKNOWLEDGMENTS The authors thank Thermo Fisher Scientific for provision of LC-HR-MS instrumentation and Dr R Whelpton for pharmacokinetic calculations. REFERENCES 1. Kuruvilla M, Gurk-Turner C. A review of warfarin dosing and monitoring. Proc (Bayl Univ Med Cent). 2001;14:305–306. 2. Royal Pharmaceutical Society of Great Britain and British Medical Association. British National Formulary. 66th ed. London: Pharmaceutical Press; 2013. 3. Holford NH. Clinical pharmacokinetics and pharmacodynamics of warfarin. Clin Pharmacokinet. 1986;11:483–504. 4. Baglin T, Keeling D, Kitchen S. Effects on routine coagulation screens and assessment of anticoagulant intensity in patients taking oral dabigatran or rivaroxaban: guidance from the British Committee for Standards in Haematology. Br J Haematol. 2012;159:427–429. 5. Schulman S. Advantages and limitations of the new anticoagulants. J Intern Med. 2014;275:1–11. 6. Ten Cate H. New oral anticoagulants: discussion on monitoring and adherence should start now! Thromb J. 2013;11:8–12. 7. MHRA. Drug safety update–dabigatran. 2011;5:A2. Available at: http:// www.mhra.gov.uk/Safetyinformation/DrugSafetyUpdate/CON137771. Accessed June 24, 2013. 8. Harper P, Young L, Merriman E. Bleeding risk with dabigatran in the frail elderly. N Engl J Med. 2012;366:864–866. 9. Pernod G, Albaladejo P, Godier A, et al. Management of major bleeding complications and emergency surgery in patients on long-term treatment with direct oral anticoagulants, thrombin or factor-Xa inhibitors. Proposals of the Working Group on Perioperative Haemostasis (GIHP). Ann Fr Anesth Reanim. 2013;32:691–700. 10. Gong IY, Kim RB. Importance of pharmacokinetic profile and variability as determinants of dose and response to dabigatran, rivaroxaban, and apixaban. Can J Cardiol. 2013;29(7 suppl):S24–S33. 11. Mendell J, Zahir H, Matsushima N, et al. Drug-drug interaction studies of cardiovascular drugs involving P-glycoprotein, an efflux transporter, on the pharmacokinetics of edoxaban, an oral factor Xa inhibitor. Am J Cardiovasc Drugs. 2013;13:331–342.
2014 Lippincott Williams & Wilkins
Measurement of DOACs Using LC-HR-MS
12. van Ryn J, Strangler J, Haertter S, et al. Dabigatran etexilate—a novel, reversible, oral direct thrombin inhibitor: interpretation of coagulation assays and reversal of anticoagulant activity. Thromb Haemostas. 2010;103:1116–1127. 13. Douxfils J, Mullier F, Loosen C, et al. Assessment of the impact of rivaroxaban on coagulation assays: laboratory recommendations for the monitoring of rivaroxaban and review of the literature. Thromb Res. 2012;130:956–966. 14. Douxfils J, Muller F, Robert S, et al. Impact of dabigatran on a large panel of routine or specific coagulation assays—laboratory recommendations for monitoring of dabigatran etexilate. Thromb Haemostas. 2012;107:985–997. 15. Samama MM, Amiral J, Guinet C, et al. Monitoring plasma levels of factor Xa inhibitors: how, why and when? Expert Rev Hematol. 2013;6:155–164. 16. Douxfils J, Chatelain C, Chatelain B, et al. Impact of apixaban on routine and specific coagulation assays: a practical laboratory guide. Thromb Haemost. 2013;110:283–294. 17. Rohde G. Determination of rivaroxaban—a novel, oral, direct factor Xa inhibitor—in human plasma by high performance liquid chromatography–tandem mass spectrometry. J Chromatogr B. 2008;872:43–50. 18. Blech S, Ebner T, Ludwig-Schwellinger E, et al. The metabolism and disposition of the oral direct thrombin inhibitor, dabigatran, in humans. Drug Metab Dispos. 2008;36:386–399. 19. Antovic JP, Skeppholm M, Eintrei J, et al. Evaluation of coagulation assays versus LC-MS/MS for determinations of dabigatran concentrations in plasma. Eur J Clin Pharmacol. 2013;69:1875–1881. 20. Douxfils J, Dogne JM, Mullier F, et al. Comparison of calibrated dilute thrombin time and aPTT tests with LC-MS/MS for the therapeutic drug monitoring of patients treated with dabigatran etexilate. Thromb Haemost. 2013;110:543–549. 21. Douxfils J, Tamigniau A, Chatelain B, et al. Comparison of calibrated chromogenic anti-Xa assay and PT tests with LC-MS/MS for the therapeutic drug monitoring of patients treated with rivaroxaban. Thromb Haemost. 2013;110:723–731. 22. Bathala MS, Masumoto H, Oguma T, et al. Pharmacokinetics, biotransformation, and mass balance of edoxaban, a selective, direct factor Xa inhibitor, in humans. Drug Metab Dispos. 2012;40:2250–2255. 23. Delavenne X, Mismetti P, Basset T. Rapid determination of apixaban concentration in human plasma by liquid chromatography/tandem mass spectrometry: application to pharmacokinetic study. J Pharm Biomed Anal. 2013;78–79:150–153. 24. Palladino M, Merli G, Thomson L. Evaluation of the oral direct factor Xa inhibitor—betrixaban. Expert Opin Investig Drugs. 2013;22:1465–1472. 25. Food and Drug Administration/Center for Drug Evaluation and Research (FDA/CDER). Guidance for industry. Bioanal Method Validation. 2001. Available at: http://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/UCM070107. pdf. Accessed September 1, 2013. 26. Polettini A. Applications of LC-MS in Toxicology. London: Pharmaceutical Press; 2006. 27. Bonfiglio R, King RC, Olah TV, et al. The effects of sample preparation on the variability of electrospray ionization response for model drug compounds. Rapid Commun Mass Spectrom. 1999;13:1175–1185. 28. Frost C, Wang J, Nepal S, et al. Apixaban, an oral, direct factor Xa inhibitor: single dose safety, pharmacokinetics, pharmacodynamics and food effect in healthy subjects. Br J Clin Pharmacol. 2013;75:476–487. 29. Mani H, Hesse C, Stratmann G, et al. Rivaroxaban differentially influences ex vivo coagulation assays based on the administration time. Thromb Haemost. 2011;106:156–164. 30. Ebner T, Wagner K, Wienen W. Dabigatran acylglucuronide, the major human metabolite of dabigatran: in vitro formation, stability, and pharmacological activity. Drug Metab Dispos. 2010;38:1567–1575. 31. Herman JL. Generic approach to high throughput ADME screening for lead candidate optimization. Int J Mass Spectrom. 2004;238:107–117. 32. Couchman L. Turbulent flow chromatography in bioanalysis: a review. Biomed Chromatogr. 2012;26:892–905. 33. He X, Kozak M. Development of a liquid chromatography-tandem mass spectrometry method for plasma-free metanephrines with ion-pairing turbulent flow online extraction. Anal Bioanal Chem. 2012;403:3003–3010. 34. Couchman L, Belsey SL, Handley SA, et al. A novel approach to quantitative LC-MS/MS: therapeutic drug monitoring of clozapine and norclozapine using isotopic internal calibration. Anal Bioanal Chem. 2013; 405:9455–9466.
605
