Evaluation of dabigatran, rivaroxaban and apixaban target-specific assays in a multicenter French study

Accepted Manuscript Evaluation of dabigatran, rivaroxaban and apixaban targetspecific assays in a multicenter French study

Isabelle Gouin-Thibault, Geneviève Freyburger, Emmanuel de Maistre, Sophie Susen, Xavier Delavenne, Jean-Louis Golmard, Yves Gruel, Pierre Sié, the GFHT study group on DOAC PII: DOI: Reference:

S0049-3848(17)30479-6 doi: 10.1016/j.thromres.2017.09.001 TR 6777

To appear in:

Thrombosis Research

Received date: Revised date: Accepted date:

25 April 2017 8 August 2017 1 September 2017

Please cite this article as: Isabelle Gouin-Thibault, Geneviève Freyburger, Emmanuel de Maistre, Sophie Susen, Xavier Delavenne, Jean-Louis Golmard, Yves Gruel, Pierre Sié, the GFHT study group on DOAC , Evaluation of dabigatran, rivaroxaban and apixaban target-specific assays in a multicenter French study, Thrombosis Research (2017), doi: 10.1016/j.thromres.2017.09.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT EVALUATION OF DABIGATRAN, RIVAROXABAN AND APIXABAN TARGET-SPECIFIC ASSAYS IN A MULTICENTER FRENCH STUDY. Isabelle Gouin-Thibault1,2, Geneviève Freyburger3, Emmanuel de Maistre4, Sophie Susen5,6, Xavier Delavenne7,8 Jean-Louis Golmard9, Yves Gruel10, Pierre Sié11 and the GFHT* study group on DOAC 1

INSERM UMR_S1140, Faculté de Pharmacie, Paris, France Laboratoire d’Hématologie, Centre Hospitalier Universitaire, Rennes, France 3 Laboratoire d’Hématologie, Centre Hospitalier Universitaire, Bordeaux, France 4 Unité d’Hémostase, Centre Hospitalier Universitaire, Dijon, France 5 Hémostase et transfusion, Centre Hospitalier Universitaire Lille 6 UMR_S1011, Faculté de Médecine Pôle Recherche, Université de Lille Nord de France, Lille, France 7 Laboratoire de Pharmacologie -Toxicologie, Centre Hospitalier Universitaire, Saint-Etienne, France 8 Groupe de Recherche sur la Thrombose, Université Jean Monnet, Saint-Etienne, France 9 Biostatistics, AP-HP, Pitié-Salpêtriere University Hospital, Paris, France 10 Hematologie-Hemostase, Centre Hospitalier Universitaire, Tours, France 11 Hematology Laboratory, Toulouse University Hospital, Toulouse, France

US

CR

IP

T

2

AN

*Groupe Français D’étude de l’Hémostase et la Thrombose

Corresponding author

CE

PT

ED

M

Participating investigators Blanc Jouvan F (Annecy ), Trichet C. (Argenteuil) Menard Deroure F (Bayonne ), : Mourey G (Besançon), Freyburger G (Bordeaux ) , : Perrot M, Galinat H (Brest), Brionne-François M, Le Querrec A (Caen), de Maistre E (Dijon ), Fontana P (Genève), Marlu R (Grenoble ), Pineau Vincent F (Le Mans), Bauters A (Lille), Donnard M (Limoges ), Berruyer M, Méraud Vialon G (Lyon), Aillaud MF, Trevisan E (Marseille ), Biron C, Sauguet P (Montpellier ), Toussaint Hacquard M, Eschwege V (Nancy) Fischer F, Appert Flory A (Nice), Mazoyer E, Abecassis L (Paris-Avicenne ), Huisse MG (Paris-Bichat ), Le Flem L (Paris-Biomnis ), Flaujac C (Paris-Cochin ), Alhenc M (Paris-HEGP), Martin Toutain I (Paris-La Pitié ), d’Audigier C (Paris-Saint Antoine), Hézard N (Reims), Le Cam Duchez V (Rouen), Meley R (Saint-Etienne), Voisin S (Toulouse), Delahousse B, Charles L (Tours)

AC

Dr Isabelle Gouin-Thibault Hématologie Biologique CHU Pontchaillou 2 rue Henri Le Guilloux 35033 Rennes Cedex 09 France

Work presentation: XXVth Congress of the International Society on Thrombosis and Hemostasis, Toronto, 2015

Word Count: 3347 1

ACCEPTED MANUSCRIPT ABSTRACT Dabigatran etexilate, rivaroxaban and apixaban (DOACs) are widely used and measurement of their concentration is desirable in certain clinical situations. Target-specific assays are available but limited information exists on their performance especially in their ability to accurately measure low and high

T

concentrations.

IP

Aims: To define, in a multicenter study, the precision and accuracy of DOAC measurements in daily practice.

CR

Methods: 15 plasma samples (kindly provided by Hyphen-Biomed) spiked with 5 blinded concentrations of dabigatran, rivaroxaban or apixaban (targeted 0-40-100-250-500 ng/mL, actual concentrations measured by

US

HPLC-MS/MS), were sent to 30 haemostasis laboratories. DOAC concentration, PT and aPTT were measured

AN

once in each sample using local reagents. Interlaboratory precision was determined by its coefficient of variation (CV) and accuracy by its bias.

M

Results: 464 DOAC measurements were performed in the 30 laboratories using 4 dabigatran and 5

ED

rivaroxaban/apixaban calibrated assays on 3 analysers. Inter-laboratory CVs were below 18% for concentrations ≥100 ng/mL, and higher for concentrations ~40 ng/mL; biases were below 8% for all drugs and concentrations.

PT

In DOAC-free samples, concentrations were all below the lower limit of quantification except for one value

CE

(dabigatran: 35 ng/mL). Depending on the concentrations, significant differences were found between reagents in rivaroxaban and apixaban concentration values. PT and aPTT ratios displayed a low sensitivity to apixaban.

AC

Conclusion: Our results suggest that calibrated DOAC assays allow the reliable measurement of a wide range of drug concentrations, even though improvement of their performances is necessary, especially for measuring low concentrations.

Keywords: DOAC measurement, PT, aPTT, direct factor Xa inhibitor, direct thrombin inhibitor, anti-Xa activity, ecarin assay, dilute thrombin time

2

ACCEPTED MANUSCRIPT INTRODUCTION The advent of direct oral anticoagulants (DOACs) is a major breakthrough in the management of patients with venous thromboembolic diseases and atrial fibrillation. DOACs are approved in many countries in these indications and a large number of individuals are receiving DOAC therapy. Clinical trials with DOACs have

T

been performed without laboratory assessment, owing to their predictable anticoagulant response and wide

IP

therapeutic index. However, measuring the anticoagulant effect of the drug and/or its concentration might be

CR

helpful for patients’ management in instances such as emergency surgery or invasive procedures, bleeding, or patients presenting with an acute ischemic stroke [1]. In such situations, the common laboratory tests

US

Prothrombin Time (PT), and Activated Partial Thromboplastin time (aPTT) can be used to suspect the presence

AN

in plasma of pharmacological levels of rivaroxaban, dabigatran [1,2] but not apixaban [3–5]. However, these tests are unreliable for quantifying the drug concentrations and a normal PT or APTT, depending on the reagent,

M

do not exclude on-therapy levels of dabigatran and rivaroxaban [1,2,6,7].

ED

Dedicated commercial tests based on anti-Xa activity for rivaroxaban and apixaban and anti-IIa activity for dabigatran concentration measurements, have been designed for this purpose and are available in a growing

PT

number of non-specialized laboratories, but limited information exists on their performance in routine practice.

CE

This multicenter study was aimed to evaluate the precision and accuracy of rivaroxaban, apixaban and dabigatran plasma concentration measurements in different laboratories using different reagent/calibrating

AC

systems and instruments. Special attention was given to low (residual) and high (overdose) concentrations, relevant to the management of emergency invasive procedures or bleeding events respectively. We also assessed the sensitivity of PT and aPTT reagents, used in daily practice, to DOAC different concentrations. The results indicate a relatively good between-laboratory agreement in the determination of drug concentrations and confirm the variability of the anticoagulant effect assessed using PT or APTT.

3

ACCEPTED MANUSCRIPT MATERIALS AND METHODS Preparation of test-samples Lyophilized aliquots of plasma samples spiked with 5 blinded concentrations of the 3 DOAC, dabigatran, rivaroxaban and apixaban (xabans) (expected values: 0,  40; 100 to 150;  250 and  500 ng/mL) were kindly

T

provided by Hyphen Biomed. The actual concentrations were measured with a validated tandem liquid

IP

chromatography mass spectrometry (HPLC-MS/MS) as previously described [4,8,9]. This technique displays a

CR

large quantification range from 1.00 to 1000 ng/mL, with inter-day and intra-day precisions of the method

US

below 15% and accuracy below 10%. A set of 15 samples identified with the name of the DOAC, but blinded

AN

for the drug levels, were shipped to 30 hemostasis laboratories listed at the end of the manuscript.

DOAC measurements and routine laboratory coagulation tests

M

The lyophilized plasma samples were reconstituted according to the manufacturer’s recommendations. The

ED

centers were asked to perform on the 15 samples, in simplicate on the same day, DOAC concentration measurement, PT and aPTT, with their usual methods and reagents. As shown in Table 1, 4 reagents for

PT

dabigatran and 5 for xabans were used for DOAC concentration measurement, but Hemoclot® (Hyphen) and

CE

STA Liquid anti-Xa® (Stago) represented a large majority for measuring dabigatran and xabans, respectively. Calibrators and controls were from two companies, but again Hyphen represented the very large majority for

AC

dabigatran assays, whereas Hyphen and Stago were equally balanced for xaban assays. The reagent/calibrator combinations (Table 1) were homogenous (i.e provided by the same company, Hyphen or Stago) in all but one laboratory where STA-liquid anti-Xa® reagent from Stago and calibrators from Hyphen were used. The range of the commercial calibrators was comprised between < 50 to about 500 -700 ng/mL. A few laboratories (n=5), also used Biophen® dabigatran, rivaroxaban or apixaban calibrator low, focused on the lowest concentrations (3 levels : 0,  50 ng/mL,  100 ng/mL).

4

ACCEPTED MANUSCRIPT A lower plasma dilution for measurements of low concentrations or a sample dilution in normal pool plasma for measurement of high concentration were performed in 2 and 3 laboratories, respectively. The coagulation analysers were from 3 manufacturers: STAR®, Stago (n=19), ACL-TOP®, Werfen (n=11) and BCS®, Siemens (n=1). In one laboratory, dabigatran concentrations were measured using 4 reagent/analyser

T

combinations and rivaroxaban and apixaban concentrations using 3 reagent/analyser combinations. Most

IP

laboratories reported lower and upper limits of detection of 30 and 500 ng/mL, respectively.

CR

PT and aPTT were measured using 5 and 7 reagents, respectively (Table 1). Results of PT and aPTT and TT were expressed as ratio [plasma spiked with DOAC (sec) / normal pool plasma (sec)] using the local normal

US

pool plasma.

AN

In plasma containing dabigatran, undiluted Thrombin Time (TT) was measured. In plasma samples containing rivaroxaban or apixaban, anti-Xa assay calibrated with Low Molecular Weight Heparin (LMWH) and expressed

M

as IU/ mL was also performed: STA Liquid anti-Xa® (Stago) was used in 18 laboratories, Biophen heparin®,

ED

Hyphen in 8 laboratories and IL liquid anti-Xa®, Werfen in 2 laboratories.

PT

Statistical analysis

CE

The median, minimal and maximal, values were calculated for each sample. Accuracy of DOAC concentration measurements was calculated by its bias, by dividing the difference between the measured DOAC concentration

AC

and the respective HPLC-MS/MS value by the HPLC-MS/MS value and multiplying the result by 100, and was expressed as mean ± standard deviation. The inter-laboratory variability was evaluated by determining the coefficient of variation (CV-%) for inter-laboratory precision. Accuracy and precision were not assessed for samples containing no drug. The sign test was used to test whether the values were systematically greater or less than the actual concentration (HPLC-MS/MS) in each laboratory for each drug. Since 30 sign tests were performed, a Bonferroni correction was applied: a nominal p-value value of 0.005/30 = 0.0016 was considered

5

ACCEPTED MANUSCRIPT significant after correction. When available, the results were compared between concentrations measured with the “standard” calibrators or the “low” calibrators, using a paired Student t-test. For each sample, an ANOVA followed by a Tukey-Kramer test was used for pairwise comparisons of reagents and analysers. Correlations between PT or aPTT ratios and DOAC concentrations for each anticoagulant were assessed using

IP

T

Spearman's rank correlation coefficients.

CR

RESULTS

DOAC concentrations measured in spiked plasmas are all close to actual values (HPLC-MS/MS) (Table 2). In

US

three laboratories, significant systematic biases were found but they were no longer present after Bonferroni

AN

correction.

A strong correlation between aPTT, PT values and DOAC concentrations was evidenced with the three drugs

ED

M

(p<0.0001, not shown). Individual PT and aPTT values according to the reagents used are reported Figure 1.

Dabigatran

PT

As expected, the distribution of reagents among laboratories is largely uneven, with the use of Hemoclot ®/

CE

Dabigatran plasma calibrators combination in most laboratories (n=28) either on a STAR® or an ACL-TOP®, with no difference between these 2 analysers (Figure 2). Concentrations measured with other reagents or the

AC

other analyser were not compared due to the low number of values (Figure 2). Inter-laboratory CVs were ≤10% for concentrations ≥100 ng/mL, and 29% for the lowest concentrations i.e 38 ng/mL (Table 2). Biases were below 7 % for all concentrations (Table 2). In dabigatran-free samples, concentrations were all below the local lower limits of quantification, excepted in one lab with a dabigatran value of 35 ng/mL.

6

ACCEPTED MANUSCRIPT When the Biophen® dabigatran low calibrators were used to measure the lowest concentration in 5 laboratories, no significant difference in the values was shown with those measured with the “standard” calibrators, in the other laboratories. aPTT was more sensitive to dabigatran than PT with some differences in the magnitude of the effect from one

T

reagent to another (Figure 1).

IP

In all dabigatran-spiked samples, thrombin time values (n=96) were above the upper limit of the reference

CR

range, defined as a ratio of 1.25, and most of them were above the upper limit of measure.

US

Rivaroxaban

AN

More than half laboratories (n=17) used STA-Liquid anti-Xa® for rivaroxaban measurement. Inter-laboratory CVs were ≤18% for concentrations ≥100 ng/mL, and 25% for the lowest concentrations i.e 40

M

ng/mL (Table 2). Biases were below 8% for all concentrations (Table 2). In rivaroxaban-free samples,

ED

concentrations were all below the local lower limits of quantification. Significant differences were found between STA Liquid anti-Xa® and Biophen Heparin® assays, for the lowest

PT

concentrations (Figure 3). As well, differences were found between STAR® and ACL-TOP® analysers which

on ACL-TOP.

CE

are likely related to the reagent, since STA liquid anti-Xa® tests were run on STAR and Biophen Heparin® tests

AC

When the Biophen® rivaroxaban low calibrators were used to measure the lowest concentration in 5 laboratories, no significant difference in the values was shown with those measured with the “standard” calibrators, in the other laboratories. As expected, PT was more sensitive to rivaroxaban than aPTT (Figure 1). Depending on the reagent used, LMWH anti-Xa activities varied from 0.23 to 1.74 IU/mL (median: 0.72) in samples spiked with 40 ng/mL and from 0.57 IU/mL to above the upper limit of quantification in samples

7

ACCEPTED MANUSCRIPT spiked with 101 ng/mL of rivaroxaban. LMWH anti-Xa activities in other rivaroxaban-spiked samples were all above the upper limit of quantification.

Apixaban

T

As for rivaroxaban, apixaban concentrations were measured with STA Liquid anti-Xa®, Stago in more than half

IP

the laboratories (n=14).

CR

Inter-laboratory CVs were ≤10% for concentrations ≥100 ng/mL, and 16% for the lowest concentrations i.e 37 ng/mL (Table 2). Biases were below 4% for all concentrations (Table 2). In apixaban-free samples,

US

concentrations were all below the local lower limits of quantification.

AN

Significant difference was found between STA liquid anti-Xa® on STAR® analyser and Biophen heparin® assays on ACL-TOP® analyser, for one concentration (Figure 4), likely related to the reagent, as for

M

rivaroxaban.

ED

When the Biophen® apixaban low calibrators were used to measure the lowest concentration in 5 laboratories, no significant difference in the values was shown with those measured with the “standard” calibrators, in the

PT

other laboratories.

CE

Even though a strong correlation between aPTT, PT values and apixaban concentrations was evidenced, both PT and aPTT exhibited very low sensitivity to apixaban, regardless of the reagents.

AC

Depending on the reagent used, LMWH anti-Xa activities varied from 0.17 to 0.55 IU/mL (median: 0.50) in samples spiked with 37 ng/mL and from 0.81 IU/mL to above the upper limit of quantification in samples spiked with 141 ng/mL. LMWH anti-Xa activities in other apixaban-spiked samples were all above the upper limit of quantification.

8

ACCEPTED MANUSCRIPT

DISCUSSION The main result of this multicentre study is that, in spite of a variety of methods representative of those used in our country, commercial reagents with dedicated calibrators and controls allow reliable measurements over the

T

usual dynamic range of DOAC concentrations reported in clinical trials, with good accuracy and acceptable

IP

precision under routine laboratory conditions. Steady-state peak and through levels in patients treated with

CR

DOACs are highly variable but with the median peak and through values around 200 ng/mL and 100 ng/mL, respectively, for twice-daily DOAC (dabigatran and apixaban) and around 200 ng/mL and 30 ng/mL,

US

respectively, for once-daily rivaroxaban [10]. Therefore, we chose to test these “on-treatment” concentrations.

AN

Moreover, the 30-50 ng/mL concentration is in line with the 30 ng/mL concentration proposed as the safety haemostatic threshold before an invasive procedure with high risk of bleeding, by the French Working Group

M

on perioperative haemostasis and the Scientific and Standardization Committee of the International Society on

ED

Thrombosis and Hemostasis and with the 50 ng/mL which is the proposed threshold to warrant antidote administration in case of bleeding [11,12]. Finally, the highest concentration we tested (around 500 ng/mL)

PT

corresponds to high concentrations that may reflect an accumulation.

CE

Since calibrated anti-Xa or anti-IIa assays for DOAC measurement were not readily available until recently in unspecialized laboratories, most studies designed for the evaluation of these assays published so far are mono-

AC

or bi-centric and very few multi-centric [13,14], devoted to only one DOAC. To the best of our knowledge, our study is the first designed on a relatively large multi-centric basis, testing a panel of calibrated assays for the measurement of the three most used DOACs over a range of plasma concentrations relevant to treatment (Table 3). This allows an evaluation of commercial target-specific assays in “real world” clinical laboratory workup. As far as dabigatran is concerned, calibrated diluted thrombin time (dTT) and ecarin chromogenic assay (ECA) are two sensitive quantitative methods for the measurement of direct thrombin inhibitors [2,15,16]. In a

9

ACCEPTED MANUSCRIPT multicenter study, good correlation between laboratory methods (ECA and dTT) and LC-MS/MS measurements for dabigatran was reported, but with significant differences in measured concentrations between laboratories and method [17]. In our study, most laboratories (28/30) used the same reagent for dabigatran measurement (Hemoclot®, Hyphen), which may account for the good performances, regardless of the analyser used, excepted

T

with the sample containing the lowest dabigatran level, as discussed below.

IP

Several standardized easy to run, sensitive and specific calibrated chromogenic anti-Xa assays have been

CR

developed to quantify rivaroxaban and apixaban [2,10,16,18,19]. The performances of these assays in spiked plasmas, evaluated in multicentre studies, vary according to the reagents, the drug concentrations and the design

US

of the studies [3,4,13,14,20,21]. Inter-laboratory CVs ranging from < 5% when using one reagent (9 centers) for

AN

measuring a concentration of 71 ng/mL, to 44% when using 4 reagents (12 centers) for measuring a concentration of 750 ng/mL, have been reported with rivaroxaban (Table 3) [13,14,20,21]. For apixaban

Similar between-reagent differences were reported in the measurement of

ED

with rivaroxaban [3,4,22].

M

measurement, the performances reported in the literature as well as in our hands are slightly better than those

of drug concentrations [16,23].

PT

unfractionated heparin, but they were considered not clinically relevant with respect to the on-treatment range

CE

Overall, for DOAC concentrations ≥50 ng/ml, our results are comparable to, or better than, those obtained in the UK Nequas and the ECAT Foundation surveys that included a large number of centers [24,25]: for instance, in

AC

the ECAT survey, the CVs vary from 11% to 17% for concentrations of dabigatran of 62 and 114 ng/mL, whereas the CVs are around 12% and 10% for 199 and 145 ng/ml of rivaroxaban and apixaban, respectively. We find significant differences in the concentrations of rivaroxaban and apixaban measured with STA-Liquid anti-Xa® and Biophen Heparin® (figures 3 and 4). The magnitude and the direction of these differences which were non-systematic and varied with the drug level and the DOAC, do not allow us to recommend the use of one reagent rather than another.

10

ACCEPTED MANUSCRIPT Due to the high inter-individual variability in the response to DOACs the delay since the last intake does not allow to predict DOAC concentrations, which can only be accurately quantified using reliable tests [7,9,26]. In the emergency peri-operative setting, measurement of concentrations ≤ 50 ng/mL is therefore a major concern. In our study, in samples containing the lowest DOAC concentration, the mean value of dabigatran concentration

T

is 36 ng/mL (actual value 38 ng/mL), and that of rivaroxaban is 40 ng/mL (actual value 40 ng/mL), with inter-

IP

laboratory CVs of 29 % and 25%, leading to acceptable ranges of 28 to 48 ng/mL and 30 to 50 ng/mL,

CR

respectively. In case of rivaroxaban, a significant difference between reagents was found (figure 3A) with a trend toward a better accuracy when using STA Liquid anti-Xa® compared to Biophen Heparin® assays.

US

However, since the completion of the study, Hyphen has proposed some improvements in their assay (see

AN

below). Finally, the mean value of apixaban is 38 ng/mL (actual value 37 ng/mL), with an inter-laboratory CV of 16 %, without apparent effect of the reagent used (figure 4A). These results confirms those obtained in

M

single-centre studies.

ED

Our results suggest that the use of “low” calibrators do not improve the performances of commercial assays to mesure dabigatran, rivaroxaban or apixaban concentrations. However, these results have to be interpreted with

PT

caution since the concentrations measured with a “low” or a “standard” calibration curve were performed in

CE

different laboratories and, only 5 values are available with “low” calibrators. Similarly sample dilution was carried out in some occurrences to improve the performances for the measurement of low or high

AC

concentrations, but too seldom data were available in our study to draw any conclusion. The question on whether the use of adapted calibrators (“low” calibration curve) or methods (lower sample dilution) would be more suitable for measurement of low concentrations of rivaroxaban or dabigatran has been raised by several groups. A “low calibrator set” assures a precise determination of rivaroxaban concentrations < 25 ng/mL with a lower limit of quantification (LLQ) of 10 ng/mL while the LLQ with the “standard calibrator set” is about 25 ng/mL [27]. In patients with rivaroxaban concentrations below 102 ng/ml, the use of “low” calibrators for the

11

ACCEPTED MANUSCRIPT anti-Xa assays, direct Xa inhibitor (DiXa-I®) and Heparin LRT®, slightly improved the correlation, the linear regression, and the inter-rater reliability with the LC-MS/MS measurement [28]. Finally, the Biophen DiXaI LOW® (lower plasma dilution and use of “low” standards) and STA LAX® showed better correlation with LCMS/MS measurements than Biophen DiXaI® in patients not bridged with LMWH [29]. When considering

T

dabigatran, the Hemoclot Thrombin Inhibitor LOW® (Hyphen) and the ECA-II® (Stago) assays, which are

IP

specifically adapted to evaluate low dabigatran concentrations performed well with dabigatran concentrations in

CR

plasma < 50 ng/mL and were more accurate than standard Hemoclot Thrombin Inhibitor® assay [30,31]. Despite these apparently better performances, adapting calibrations and/or methods is a cumbersome process to

US

implement in routine practice. Moreover the clinical relevance of the small gain in performances is not

AN

warranted.

Overall, depending on the drug level, significant differences in rivaroxaban or apixaban concentrations were

M

noted according to the reagents. The magnitude and the direction of these differences which were non-

ED

systematic and varied with the drug level and the DOAC, do not allow us to recommend the use of one reagent rather than another.

PT

It is widely accepted that routine laboratory tests such as PT and aPTT are poorly sensitive to DOACs and

CE

unsuitable to detect apixaban. In spite of a strong correlation between PT or aPTT ratio with DOAC concentrations, our results confirm the lack of performances of PT and aPTT to detect low concentrations of

AC

rivaroxaban and dabigatran and "on-therapy" concentrations of apixaban. Interestingly, in our study, it appears that the variability between laboratories is as least as important as difference between reagents (figure 1). Therefore, the combination of variabilities due to reagent sensitivity, instrumentation, and, not evaluated in our study, preanalytical variables and baseline coagulation values in patients, explains that routine coagulation tests are not reliable for quantifying DOACs, and unable to detect concentrations <50 ng/mL.

12

ACCEPTED MANUSCRIPT We also confirmed that thrombin time is highly sensitive to low concentrations of dabigatran. The anti-Xa assays calibrated with LMWH used in this study were able to detect the lowest concentrations of rivaroxaban and apixaban. Our study has several limitations. We did not use plasmas samples collected in subjects treated with DOACs.

T

This would be impracticable for testing the performances of reagents over a wide range of drug plasma levels in

IP

a large multicentre study. For this purpose, the use of spiked plasma samples is a well-established methodology

CR

to test the performances of reagents in a large multicentre study, but it does not allow to study the influence of the patient plasma. This is likely to be important for non-specific assays such as PT or aPTT, but an higher

US

inter-laboratory variability using ex-vivo samples compared to spiked plasmas has been reported for the

AN

quantification of drug plasma levels as well [14]. Another limitation is the uneven distribution of reagents among laboratories, with Hemoclot® for dabigatran and STA-Liquid anti-Xa® for xabans used in 28 and 17 out

M

of the 30 laboratories, respectively. This distribution was that existing at the time of the study, before the

ED

introduction of new players in the field. Especially, we have not evaluated the performances of the modified tests proposed by the in-vitro diagnostic companies since the completion of the study i.e a modified colometric

PT

assay for dabigatran measurement, the STA-ECA II® from Stago and, a modification of the regression mode of

CE

the calibration curve for the Biophen Heparin® assay. In conclusion, our results obtained with spiked lyophilized plasmas show that commercial specific DOAC

AC

assays performed according to manufacturer recommendations, display good performances in routine laboratory practice and would allow a reliable measurement of dabigatran, rivaroxaban and apixaban with acceptable interlaboratory variability. Improvements are necessary for measuring low concentrations. An external quality control exercise is now required before spreading this conclusion to the growing number of non-specialized laboratories performing these tests. The use of PT or aPTT should be discouraged especially to detect “ontherapy” concentrations of apixaban.

13

ACCEPTED MANUSCRIPT CONFLICTS OF INTEREST Authors: none.

AC

CE

PT

ED

M

AN

US

CR

IP

T

Lyophilized plasma samples were kindly provided by Hyphen Biomed.

14

ACCEPTED MANUSCRIPT REFERENCES

AC

CE

PT

ED

M

AN

US

CR

IP

T

[1] T. Baglin, A. Hillarp, A. Tripodi, I. Elalamy, H. Buller, W. Ageno, Measuring Oral Direct Inhibitors (ODIs) of thrombin and factor Xa: A recommendation from the Subcommittee on Control of Anticoagulation of the Scientific and Standardisation Committee of the International Society on Thrombosis and Haemostasis, J. Thromb. Haemost. 11 (2013) 756–60. [2] B.T. Samuelson, A. Cuker, D.M. Siegal, M. Crowther, D.A. Garcia, Laboratory Assessment of the Anticoagulant Activity of Direct Oral Anticoagulants: A Systematic Review, Chest. 151 (2017) 127–138. [3] J. Harenberg, S. Du, C. Weiss, R. Krämer, D. Hoppensteadt, J. Walenga, working party: methods to determine apixaban of the Subcommittee on Control of Anticoagulation of the International Society of Thrombosis and Haemostasis, Report of the Subcommittee on Control of Anticoagulation on the determination of the anticoagulant effects of apixaban: communication from the SSC of the ISTH, J. Thromb. Haemost.. 12 (2014) 801–804. [4] 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 (2014) 240–248. [5] J. Douxfils, C. Chatelain, B. Chatelain, J.-M. Dogné, F. Mullier, Impact of apixaban on routine and specific coagulation assays: a practical laboratory guide, Thromb. Haemost. 110 (2013). [6] S. Testa, C. Legnani, A. Tripodi, O. Paoletti, V. Pengo, R. Abbate, L. Bassi, P. Carraro, M. Cini, R. Paniccia, D. Poli, G. Palareti, Poor comparability of coagulation screening test with specific measurement in patients on direct oral anticoagulants: results from a multicenter/multiplatform study, J. Thromb. Haemost. 14 (2016) 2194-2201 [7] G. Freyburger, G. Macouillard, S. Labrouche, F. Sztark, Coagulation parameters in patients receiving dabigatran etexilate or rivaroxaban: two observational studies in patients undergoing total hip or total knee replacement, Thromb. Res. 127 (2011) 457–465. [8] X. Delavenne, J. Moracchini, S. Laporte, P. Mismetti, T. Basset, UPLC MS/MS assay for routine quantification of dabigatran - a direct thrombin inhibitor - in human plasma, J. Pharm. Biomed. Anal. 58 (2012) 152–156. [9] I. Gouin-Thibault, X. Delavenne, A. Blanchard, V. Siguret, J.E. Salem, C. Narjoz, P. Gaussem, P. Beaune, C. Funck-Brentano, M. Azizi, P. Mismetti, M.A. Loriot, Inter-individual variability in dabigatran and rivaroxaban exposure: contribution of ABCB1 genetic polymorphisms and interaction with clarithromycin, J. Thromb. Haemost. 15 (2017) 273-283 [10] A. Cuker, Laboratory measurement of the non-vitamin K antagonist oral anticoagulants: selecting the optimal assay based on drug, assay availability, and clinical indication, J. Thromb. Thrombolysis. 41 (2016) 241–247. [11] J.H. Levy, W. Ageno, N.C. Chan, M. Crowther, P. Verhamme, J.I. Weitz, Subcommittee on Control of Anticoagulation, When and how to use antidotes for the reversal of direct oral anticoagulants: guidance from the SSC of the ISTH, J. Thromb. Haemost. 14 (2016) 623–627. [12] G. Pernod, P. Albaladejo, A. Godier, C.M. Samama, S. Susen, Y. Gruel, N. Blais, P. Fontana, A. Cohen, J.V. Llau, N. Rosencher, J.-F. Schved, E. de Maistre, M.M. Samama, P. Mismetti, P. Sié, 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) - March 2013, Arch. Cardiovasc. Dis. 106 (2013) 382-393 [13] T.A. Helin, A. Pakkanen, R. Lassila, L. Joutsi-Korhonen, Laboratory Assessment of Novel Oral Anticoagulants: Method Suitability and Variability between Coagulation Laboratories, Clin. Chem. 59 (2013) 807–814. 15

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

[14] 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 (2012) 492–498. [15] J. Douxfils, F. Mullier, S. Robert, C. Chatelain, B. Chatelain, J.-M. Dogné, Impact of dabigatran on a large panel of routine or specific coagulation assays. Laboratory recommendations for monitoring of dabigatran etexilate, Thromb. Haemost. 107 (2012) 985–997. [16] B.J. Dale, N.C. Chan, J.W. Eikelboom, Laboratory measurement of the direct oral anticoagulants, Br. J. Haematol. 172 (2016) 315–336. [17] R. Gosselin, E. Hawes, S. Moll, D. Adcock, Performance of various laboratory assays in the measurement of dabigatran in patients receiving therapeutic doses: a prospective study based on peak and trough plasma levels, Am. J. Clin. Pathol. 141 (2014) 262–267. [18] J. Douxfils, H. Mani, V. Minet, B. Devalet, B. Chatelain, J.-M. Dogné, F. Mullier, Non-VKA Oral Anticoagulants: Accurate Measurement of Plasma Drug Concentrations, BioMed Res. Int. 2015 (2015) 345138. [19] R.C. Gosselin, D.M. Adcock, The laboratory’s 2015 perspective on direct oral anticoagulant testing, J. Thromb. Haemost. 14 (2016) 886–893. [20] M.M. Samama, G. Contant, T.E. Spiro, E. Perzborn, C. Guinet, Y. Gourmelin, L. Le Flem, G. Rohde, J.L. Martinoli, Evaluation of the anti-factor Xa chromogenic assay for the measurement of rivaroxaban plasma concentrations using calibrators and controls, Thromb. Haemost. 107 (2012) 379–387. [21] J. Harenberg, S. Marx, C. Weiss, R. Krämer, M. Samama, S. Schulman, Subcommittee on Control of Anticoagulation of the ISTH, Report of the Subcommittee of Control of Anticoagulation on the determination of the anticoagulant effects of rivaroxaban, J. Thromb. Haemost. JTH. 10 (2012) 1433– 1436. [22] A. Hillarp, K.M. Gustafsson, L. Faxälv, 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 (2014) 1545–1553. [23] R. Gosselin, R.P. Grant, D.M. Adcock, Comparison of the effect of the anti-Xa direct oral anticoagulants apixaban, edoxaban, and rivaroxaban on coagulation assays, Int. J. Lab. Hematol. 38 (2016) 505–513. [24] UK NEQAS for Blood Coagulation, (n.d.). www.ukneqasbc.org. [25] ECAT Foudation. External quality Control for Assays and Tests with focus on Thrombosis and Hemostasis., (2016). [26] I.Y. Gong, R.B. Kim, Importance of pharmacokinetic profile and variability as determinants of dose and response to dabigatran, rivaroxaban, and apixaban, Can. J. Cardiol. 29 (2013) S24-33. [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 (2012) 191–198. [28] O. Königsbrügge, P. Quehenberger, S. Belik, G. Weigel, C. Seger, A. Griesmacher, I. Pabinger, C. Ay, Anti-coagulation assessment with prothrombin time and anti-Xa assays in real-world patients on treatment with rivaroxaban, Ann. Hematol. 94 (2015) 1463–1471. [29] S. Lessire, J. Douxfils, L. Pochet, A.-S. Dincq, A.-S. Larock, M. Gourdin, J.-M. Dogné, B. Chatelain, F. Mullier, Estimation of Rivaroxaban Plasma Concentrations in the Perioperative Setting in Patients With or Without Heparin Bridging, Clin. Appl. Thromb. Off. J. Int. Acad. Clin. Appl. Thromb. (2016) [30] J. Douxfils, S. Lessire, A.-S. Dincq, P. Hjemdahl, Y. Rönquist-Nii, A. Pohanka, M. Gourdin, B. Chatelain, J.-M. Dogné, F. Mullier, Estimation of dabigatran plasma concentrations in the perioperative setting. An ex vivo study using dedicated coagulation assays, Thromb. Haemost. 113 (2015) 862–869.

16

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

[31] K. Boonen, E. Schmitz, F. Rozestraten, D. van den Heuvel, L. Brunsveld, P. van der Voort, D. van de Kerkhof, Real life dabigatran and metabolite concentrations, focused on inter-patient variability and assay differences in patients with atrial fibrillation, Clin. Chem. Lab. Med. (2017).

17

ACCEPTED MANUSCRIPT Table 1. Reagent/calibrator combinations and control plasmas used to measure dabigatran, rivaroxaban and apixaban concentrations and PT and aPTT. Numbers in brackets correspond to the number of users for each component.

Reagent/calibrator combinations

Dabigatran

Rivaroxaban

Apixaban

- Hemoclot®,Hyphen/ Dabigatran Plasma Calibrator®, Hyphen (28)

- STA liquid anti-Xa®, Stago/ STA®-Rivaroxaban Calibrator, Stago (16)

- STA liquid anti-Xa®, Stago/ STA®-apixaban Calibrator, Stago (14)

- STA liquid anti-Xa®, Stago/ Rivaroxaban Plasma Calibrator®, Hyphen (1)

- Biophen heparin®, Hyphen/ STA®-apixaban Calibrator, Stago (2)

- Biophen heparin®, Hyphen/ Rivaroxaban Plasma Calibrator®, Hyphen (8)

- Biophen heparin®, Hyphen/ apixaban Plasma Calibrator®, Hyphen (5)

- Diluted TT®, Werfen/ Dabigatran Plasma Calibrator®, Hyphen (1)

IP

T

- ECAT®, Stago/ ECA-T standard®, Stago (1)

CR

- Ecarin "home"/ Dabigatran Plasma Calibrator®, Hyphen (1)

- IL liquid anti-Xa®, Werfen/ Rivaroxaban Plasma Calibrator®, Hyphen (2)

US

- DiXa®, Hyphen/ Rivaroxaban Plasma Calibrator®, Hyphen (3)

- DiXa®, Hyphen/ apixaban Plasma Calibrator®, Hyphen (2) - Anti-Xa Chromogenix®, Werfen/ STA®-apixaban Calibrator, Stago (2)

- Dabigatran Control Plasma®, Hyphen (30) Stago (1)

ED

- ECA-T control®,

AC

aPTT

CE

PT

PT

- STA®-Rivaroxaban Stago (16)

Control,

M

Controls

AN

- Anti-Xa Chromogenix®, Werfen/ Rivaroxaban Plasma Calibrator®, Hyphen (1)

- Biophen Rivaroxaban Control Plasma®, Hyphen (15)

Neoplastin® CI+, Stago (16) HemosIL®RecombiPlasTin 2G, Werfen (4) Neoplastin® CI, Stago (3) Thromborel ® S, Siemens (2) Dade® Innovin®, Siemens (3)

PTT-A, Stago (10) HemosIL® SynthASil Werfen (3) CK. Prest®, Stago (2) TriniCLOT®aPTT, Tcoag (10) Pathromtin®S, Siemens (1) HemosIL® APTT-SP, Werfen(1) Cephascreen®, Stago (1)

18

- STA®-Apixaban Control, Stago (18) - Biophen Apixaban Plasma®, Hyphen (7)

Control

ACCEPTED MANUSCRIPT

40

CE

37

141

AC

Apixaban (n=27)

PT

499

246 492

25.3

7.4±27.2

18.3

3.7±17.6

9.4

1.5±9.6

5.9

-2.4±5.8

38 [20-50] 137 [99-153] 239 [205-268] 478 [ 420-516]

16.3

0.4±16.4

10.4

-4.4±9.9

7.3

-3.5±7.1

4.8

-3.1±4.6

ED

101 252

40 [29-77] 90 [66-159] 257 [212-334] 490 [420-586]

M

Rivaroxaban (n=32)

AN

US

CR

IP

T

Table 2. Dabigatran, rivaroxaban and apixaban concentrations in ng/mL: inter-laboratory CV and accuracy. Actual values correspond to the concentrations measured using the reference method (LC-MS/MS) in ng/mL. Measured concentrations correspond to the concentrations measured in the participating centers using their local reagent. Actual Measured Inter-laboratory Accuracy concentration concentration precision Bias, % CV median % mean ± sd [min-max] Dabigatran (n=33) 38 36 29.1 -1.5±28.7 [16-62] 114 121 9.5 6.2±10.1 [91-140] 248 261 8.2 4.8±8.6 [206-316] 494 477 8.0 -2.4±7.8 [381-603]

19

ACCEPTED MANUSCRIPT Table 3: Multicenter studies evaluating the performances of specific assays for measurements of DOACs. Study

DrugConcentration ng/mL

Reagents

Mean conc ± sd Median (range) Accuracy / LC-MS/MS

Rivaroxaban 20, 199, 662

6 local reagents + Rotachrom Heparin®

Mean ≠/true value: 1% to 15% (local) 0 to 10% (rotachrom)

Harenberg J, 2012 n=12 centers [21]

Rivaroxaban 100, 350, 750

4 provided reagents

Mean conc. ± sd : 100: 112 ± 16 350: 378 ± 101 750: 759 ± 346

Asmis LM, 2012 [14] n=9 centers

Rivaroxaban 71 to 372

Biophen Heparin 6®

Helin TA, 2013[13] n=8 centers

Rivaroxaban

4 local reagents Calibrators: 0 to 150 ng/mL

Gouin-Thibault I, 2014 [4] n=13 centers

Apixaban 96, 209, 393, 828

3 provided reagents

Harenberg J, 2014 [3] n=12 centers

Apixaban 30, 60, 100, 330

5 provided reagents

Hillarp A , 2014 [22] n=4 centers Gouin-Thibault I, GFHT n=30 centers

Apixaban 25 to 1000

IP

CR

AN

ED PT

Accuracy / LC-MS/MS Mean bias ± sd (%) 2.6±2.4 to 27.1± 1.6 Mean conc. ± sd 30: 26 ± 24 60: 62 ± 26 100: 115 ± 28 330: 345 ± 46 Recovery (%) : 88% to 119%

Dabigatran 38, 114, 248, 494

4 local reagents

Rivaroxabn 40, 101, 252, 499

5 local reagents

Accuracy / LC-MS/MS : Mean bias ± sd (%) -2.4±7.8 to 6.2±10.1 Accuracy / LC-MS/MS : Mean bias ± sd (%) -2.4±5.8 to 7;4±27.2

Apixaban 37, 141, 246, 492

5 local reagents

CE

2 local reagents

AC

14 to 37% (local) 5 to 17% (rotachrom) 7 to 25% 37 to 44% (750 ng/mL) 2.6 to 10.5%

US

Accuracy / LC-MS/MS: < 0.1% to 14.6% Mean conc. (range) 60: 67 (59-74) 146: 134 (111-154) 305: 207 (138-313)

M

60, 146, 305

T

Samama MM, 2012 [20] n=24 centers

Inter-lab. CV%

Accuracy / LC-MS/MS: Mean bias ± sd (%) -4.4±9.9 to 0.4± 16.4

20

8 to 34% (305 ng/mL) < 12%

13% (330 ng/mL) to 93% (30 ng/mL)

8 (494 ng/mL) to 29% (38 ng/mL)

6 (499 ng/mL) to 25% (40 ng/mL) 5 (478 ng/mL) to 16% (38 ng/mL)

ACCEPTED MANUSCRIPT FIGURE LEGENDS

IP

T

Figure 1: PT and aPTT ratios in dabigatran-, rivaroxaban- and apixaban-spiked samples according to the reagents PT reagents 1= Neoplastin® CI+, Stago ; 2=HemosIL®RecombiPlasTin 2G, Werfen; 3= Neoplastin® CI, Stago; 4=Thromborel ® S, Siemens; 5=Dade® Innovin®, Siemens.

AN

US

CR

aPTT reagents 1=PTT-A, Stago; 2=HemosIL® SynthASil Werfen; 3=CK. Prest®, Stago; 4=TriniCLOT®aPTT, Tcoag; 5=Pathromtin®S, Siemens; 6=HemosIL® APTT-SP, Werfen; 7=Cephascreen®, Stago

PT

ED

M

Figure 2: Dabigatran concentrations measured in dabigatran-spiked samples according to the reagents/analysers, for each drug level. A: 38 ng/mL; B: 114 ng/mL; C: 248 ng/mL; D: 494 ng/mL. 1= Hemoclot® Hyphen on STAR®, Stago; 2= Hemoclot® Hyphen on ACL-TOP®, Werfen; 3= Hemoclot® Hyphen on BCS®, Siemens; 4=Diluted TT® Werfen on ACL-TOP®, Werfen; 5=ECAT® Stago on STAR®, Stago; 6= ECAT® Stago on BCS® Siemens; 7= Ecarin "home" on ACL-TOP®, Werfen.

AC

CE

Figure 3: Rivaroxaban concentrations measured in rivaroxaban-spiked samples according to the reagents, for each drug level. A: 40 ng/mL; B: 101 ng/mL; C: 252 ng/mL; D: 499 ng/mL. 1= STA liquid anti-Xa® Stago; 2=Biophen heparin® Hyphen; 3=IL liquid anti-Xa® Werfen; 4=DiXa® Hyphen; 5=Anti-Xa Chromogenix® Werfen.

Figure 4: Apixaban concentrations measured in apixaban-spiked samples according to the reagents, for each drug level. A: 37 ng/mL; B: 141 ng/mL; C: 246 ng/mL; D: 492 ng/mL. 1= STA liquid anti-Xa® Stago; 2=Biophen heparin® Hyphen; 3=IL liquid anti-Xa® Werfen; 4=DiXa® Hyphen; 5=Anti-Xa Chromogenix® Werfen. 21

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

Figure 1

22

ACCEPTED MANUSCRIPT Figure 2

A

B

160

IP

80

T

p=0.7443

p=0.2194

70

40 30

120 100

US

20

CR

140

50

ng/mL

ng/mL

60

10 0

80 1

2

3

4

5

6

7

1

3

4

5

6

7

5

6

7

Reagent

C p=0.3762

PT

350

CE

250 200 150 1

2

3

AC

ng/mL

300

4

5

6

D p=0.8525 700 650 600

ng/mL

ED

M

AN

Reagent

2

550 500 450 400 350 300

7

1

Reagent

2

3

4

Reagent

23

ACCEPTED MANUSCRIPT Figure 3

B

T

A

80

IP

p=0.7443

p=0.2194 160 140

50

ng/mL

40 30

120

US

ng/mL

60

CR

70

20

100

10 0

80 2

3

4

5

6

7

1

2

3

AN

1

5

6

7

5

6

7

Reagent

ED

M

Reagent

4

C

p=0.8525

CE

300 250

150 1

2

3

AC

200

4

5

6

700 650 600

ng/mL

350

ng/mL

D

PT

p=0.3762

550 500 450 400 350 300

7

1

Reagent

2

3

4

Reagent

24

ACCEPTED MANUSCRIPT Figure 4

A

B

160

IP

80

T

p=0.7443

p=0.2194

70

40 30

120 100

US

20

CR

140

50

ng/mL

ng/mL

60

10 0

80 1

2

3

4

5

6

7

1

3

4

5

6

7

5

6

7

Reagent

C p=0.3762

PT

350

CE

250 200 150 1

2

3

AC

ng/mL

300

4

5

6

D p=0.8525 700 650 600

ng/mL

ED

M

AN

Reagent

2

550 500 450 400 350 300

7

1

Reagent

2

3

4

Reagent

25

ACCEPTED MANUSCRIPT

Highlights

CE

PT

ED

M

AN

US

CR

IP

T

Information on the performances of direct oral anticoagulants (DOAC) assays are required A wide range of DOAC concentrations were tested in a multicenter study (30 laboratories) Good accuracy and acceptable precision of DOAC commercial assays were found Improvements are necessary for measuring low DOAC concentrations PT or aPTT are unable to detect low concentrations of rivaroxaban and dabigatran and "on-therapy" concentrations of apixaban

AC

    

26