Interactions between rivaroxaban and antiphospholipid antibodies in thrombotic antiphospholipid syndrome

Transcription

1 Journal of Thrombosis and Haemostasis, 13: DOI: /jth ORIGINAL ARTICLE Interactions between rivaroxaban and antiphospholipid antibodies in thrombotic antiphospholipid syndrome D. R. J. ARACHCHILLAGE,* I. J. MACKIE,* M. EFTHYMIOU,* D. A. ISENBERG, S. J. M A C H I N * and H. COHEN* *Haemostasis Research Unit, Department of Haematology, University College London; Centre for Rheumatology, Division of Medicine, University College London; and Department of Haematology, University College London Hospitals NHS Foundation Trust, London, UK To cite this article: Arachchillage DRJ, Mackie IJ, Efthymiou M, Isenberg DA, Machin SJ, Cohen H. Interactions between rivaroxaban and antiphospholipid antibodies in thrombotic antiphospholipid syndrome. J Thromb Haemost 2015; 13: Summary. Introduction: Rivaroxaban can affect lupus anticoagulant (LA) testing and antiphospholipid antibodies (apl) may interfere with the anticoagulant action of rivaroxaban. Aims: To establish the influence of rivaroxaban on LA detection and of apl on the anticoagulant action of rivaroxaban. Methods: Rivaroxaban and 52 IgG preparations (20 LA+ve, 12 LA ve thrombotic antiphospholipid syndrome [APS] patients, and 20 normal controls [NC]) were spiked into pooled normal plasma (PNP) for relevant studies. LA detection was also studied in APS patients receiving rivaroxaban 20 mg once daily. Results: In vitro spiking of samples with rivaroxaban showed no false positive LA with Textarin time, Taipan venom time/ecarin clotting time (TVT/ECT), dilute prothrombin time (dpt) and in-house dilute Russell s viper venom time (DRVVT), but false positives in the majority of NC and LA negative IgG with two commercial DRVVT reagents at 2 but not rivaroxaban. Ex vivo studies: six LA+ve patients on rivaroxaban remained LA positive with TVT/ECT and DRVVT at peak ( ng/ml) and trough (30 85 ng/ml) rivaroxaban levels. Six LA-ve patients became (apparently) LA+ve with two DRVVT reagents (test/confirm ratio median [confidence interval], 1.6 [ ], 1.6 [ ]) but not with TVT/ECT at peak rivaroxaban levels, and remained LA-ve with both DRVVT reagents and TVT/ ECT at trough levels. apl positive IgG spiking of PNP had no effect on rivaroxaban s anticoagulant action on thrombin generation or rivaroxaban anti-xa levels. Conclusions: The TVT/ECT ratio and Textarin time were not Correspondence: Deepa R. J. Arachchillage, Haemostasis Research Unit, Department of Haematology, University College London, 1st Floor, 51 Chenies Mews, London WC1E 6HX, UK. Tel.: ; fax: deeparanjani@hotmail.com Received 19 January 2015 Manuscript handled by: S. Kitchen Final decision: F. R. Rosendaal, 24 March 2015 affected even at peak rivaroxaban levels, enabling detection of LA ex vivo. apl had no effects on rivaroxaban s anticoagulant action in vitro. Keywords: antiphospholipid syndrome; lupus anticoagulant; prothrombin activator (Taipan venom); rivaroxaban; Russell s viper venom time. Introduction Antiphospholipid syndrome (APS) is defined as the presence of thrombosis (venous and/or arterial or microvascular) and/or pregnancy loss or late obstetric morbidity in association with persistently positive antiphospholipid antibodies (apl) [1]. The syndrome may occur in isolation or in association with other conditions, notably systemic lupus erythematosus [1]. apl are heterogeneous and include primarily lupus anticoagulant (LA), and IgG and/ or IgM anticardiolipin antibodies (acl) and anti-b 2 glycoprotein-i (ab 2 GPI) antibodies. The international consensus (revised Sapporo [Sydney]) classification criteria define persistently positive apl as those being present on two or more occasions at least 12 weeks apart [1]. The clinical manifestations correlate best with the prolongation of phospholipid-dependent clotting assays associated with LA [2]. The current mainstay of the treatment for thrombotic APS is long-term anticoagulation with vitamin K antagonists (VKA), mainly warfarin [3]. Non-vitamin K antagonist oral anticoagulants (NOAC), dabigatran (a direct thrombin inhibitor), rivaroxaban and apixaban, both direct factor (F) Xa inhibitors, have shown promising results in the prevention of stroke and systemic embolism in patients with non-valvular atrial fibrillation in phase III clinical trials [4 6]. These agents have been licensed [7,8] and approved for this indication by the National Institute for Health and Care Excellence (NICE) in England [9,10] and the US Food and Drug Administration (FDA) [11]. They have also undergone major phase III trials for the treatment of

2 Rivaroxaban and antiphospholipid antibodies 1265 acute deep vein thrombosis and pulmonary embolism and secondary prevention of venous thromboembolism (VTE) [12 15]; rivaroxaban has been licensed for these indications [7,8] and approved by NICE and the FDA [11,16]. Approximately 10% of patients with acute VTE have APS [17], and it is therefore likely that patients with APS were among those included in the phase III clinical trials of NOAC versus VKA in patients with VTE. However, apl status was not documented in these trials [12,13,15], and therefore the results of these trials may not be generalizable to APS patients with VTE. The Rivaraoxaban in Antiphospholipid Syndrome (RAPS) trial of warfarin versus rivaroxaban ( isrctn.org/isrctn ) is aiming to demonstrate in patients with APS and previous VTE that the intensity of anticoagulation achieved with rivaroxaban is not inferior to that of warfarin. Rates of recurrent thrombosis, bleeding and serious adverse events in both patient groups are also being assessed as secondary outcome measures. With the increasing use of rivaroxaban, it is clinically relevant to define appropriate methodology for ex vivo LA testing, both to enable APS diagnosis in patients treated with rivaroxaban and to inform national and international guidelines. However, in vitro and ex vivo studies on LA detection in thrombotic APS are lacking. It is generally not recommended to perform LA testing in patients on anticoagulation as they have prolonged basal clotting times and LA reagents can have different sensitivities to anticoagulants [3]. Most patients on anticoagulants can wait until their period of anticoagulation is complete before LA detection is attempted. However, the diagnosis of APS requires that two samples are tested at least 12 weeks apart and some patients may require long-term anticoagulation; the period of discontinuation of anticoagulation may be considered to put the patient at a higher risk of recurrent thrombosis. It is theoretically possible that apl could directly interfere with the anticoagulant effects of rivaroxaban, although this is unlikely due to its small molecular size, high specificity and affinity for its target. Rivaroxaban affects various parameters of thrombin generation (TG) ex vivo, measured using calibrated automated thrombography (CAT) [18]. The specific effects of apl on the anticoagulant action of rivaroxaban have not been defined. The aim of this study was to establish the influence of therapeutic-dose rivaroxaban on LA detection in vitro and ex vivo, and of apl on the anticoagulant action of rivaroxaban assessed using TG testing and anti-xa assays. Materials and methods Blood sample collection and processing The in vitro and ex vivo studies were approved by the Research Ethics Committee and the Research and Development Office at University College London Hospitals (UCLH) NHS Trust (Reference number: 13/EM/0150). Twenty-four thrombotic APS patients on anticoagulation with warfarin for VTE (target INR 2.5 [range ] for at least 3 months) were recruited from the RAPS trial. These patients had given their written informed consent for samples to be collected for translational research (Reference number: 12/SC/0566). Samples were collected in M citrate Vacutainers Ò (Becton Dickinson, Plymouth, UK) with minimal stasis and platelet-poor plasma (PPP) was prepared within 1 h of collection by double centrifugation at g for 15 min and stored at 80 C. These 24 patients comprised the first consecutive persistently LA-positive and LA-negative patients at the UCLH site randomized to the rivaroxaban arm or to remain on warfarin, respectively (12 patients in each group, comprising the first six persistently LA-positive and first six LA-negative patients; Fig. 1). The LA status of all patients had been confirmed previously during definition of their APS-defining apl, in accordance with the International Society of Thrombosis and Haemostasis (ISTH) and British Committee for Standards in Haematology (BCSH) guidelines on LA detection [3,19]. The 12 LA-negative patients were positive for either acl, ab 2 GPI or both (five patients had both acl and ab 2- GPI, three patients had acl alone and four patients had ab 2 GPI alone). Venous blood samples from the 12 patients on rivaroxaban, (six LA positive and six LA negative) were collected 2 4 h after the morning dose of rivaroxaban to represent peak levels, and also at least 18 h (range 18 24) following the last dose, which was anticipated to represent trough levels. All patient samples were collected after at least 30 days (range days) of stable anticoagulation with either rivaroxaban or warfarin. Samples were processed according to ISTH [19] and BCSH guidelines on LA testing [3]. Platelet-poor plasma (PPP) samples were stored in aliquots at 80 C and thawed for 10 min at 37 C immediately prior to analysis. INR values measured using Innovin (Siemens Healthcare Diagnostics) with the Sysmex CS-5100 analyser (Sysmex UK Ltd, Milton Keynes, UK) for the 24 RAPS patients on warfarin pre-randomization and those 12 patients who remained on warfarin were (mean [confidence interval, CI] 2.4 [ ] and 2.3 [ ], respectively. The peak plasma rivaroxaban levels in the 12 patients who had samples taken at 2 4 h following the last morning dose of rivaroxaban, measured using an amidolytic anti-xa assay (Biophen DiXaI; Hyphen Bio- Med, Neuville-Sur-Oise, France), were 240 [ ] ng/ ml. Rivaroxaban trough levels in samples taken from the same 12 patients more than 18 h (18 24) after the last morning dose of rivaroxaban were 50 [30 80] ng/ml. These rivaroxaban levels concur with those reported as being within peak and trough levels in population pharmacokinetics studies [20] of patients on rivaroxaban 20 mg daily.

3 1266 D. R. J. Arachchillage et al 6 LA positive 24 APS patients on warfarin for previous VTE, target INR 2.5 (12 LA positive, 12 LA negative) 12 patients randomised to remain on warfarin 12 patients randomised to rivaroxaban 6 LA negative 6 LA positive 6 LA negative Fig. 1. Patient selection for ex vivo lupus anticoagulant (LA) detection studies. Venous blood samples were collected for IgG purification from the above 24 patients (12 LA positive and 12 LA negative) and from eight additional patients with thrombotic APS [1] with persistent LA (i.e. total of 32 patients). APS classification categories (according to Miyakis et al. 2006) of the 20 LA positive were as follows: 14/20 (70%) category I (more than one laboratory criteria present [any combination]); 6/20 (30%) category IIa (LA present alone) [1]. IgG purification The polyclonal IgG fraction for in vitro studies was purified from the plasma of the 32 patients detailed above (20 LA positive and 12 LA negative) and 20 healthy normal controls (NC) using NAbTM protein G spin columns (Pierce; Fisher Scientific, Loughborough, UK). IgG was quantified by in-house IgG ELISA (5 10 mg/ml). Preparation of IgG and rivaroxaban for in vitro studies Rivaroxaban was diluted in dimethyl sulfoxide (DMSO; Sigma Aldrich, Poole, UK) and Owren s veronal buffer (OVB; Siemens Healthcare Diagnostics, Marburg, Germany) [21]. A total of 50 mg rivaroxaban was dissolved in 500 ml using DMSO and stirred thoroughly to make 100 lg/ml of rivaroxaban. One milliliter ml of 100 lg/ ml rivaroxaban preparation was diluted with 4 ml of OVB to get 20 lg/ml of rivaroxaban, which contains 20% of DMSO (V/V). Addition of the required volume of rivaroxaban preparation to PNP resulted in DMSO concentration of 1% (1% for LA testing and 0.75% for TG). Both our experiments and also work from others have shown that DMSO at this concentration has no effect on TG and other coagulation assays [22]. Rivaroxaban 50, 250 and 0 ng/ml (buffer control) and IgG (500 lg/ml) preparations in 20 mm HEPES buffer (20 mm HEPES, 140 mm NaCl, 0.02% Sodium Azide at ph 7.35) with 1% BSA from patients (20 LA positive and 12 LA negative) and NC were spiked into pooled normal plasma (PNP) to give a final plasma : IgG and rivaroxaban ratio of 80 : 20 v/v (80% : 20%) and incubated for 10 min at 37 C. PNP was prepared locally by mixing equal volumes of plasma from 20 NC. PPP was prepared within 1 h of collection, by double centrifugation at ambient temperature at g for 15 min of blood collected into M citrate Vacutainers Ò. All NC were tested for apl and were negative. Rivaroxaban concentrations were selected to represent expected peak ( lg/l) and trough (4 96 lg/l) rivaroxaban levels with 20 mg rivaroxaban daily based on population pharmacokinetic studies [20]. An IgG concentration of 500 lg/ml was selected based on our preliminary studies and also the experience of other workers [23]. Lupus anticoagulant A range of LA assays were performed in vitro, on rivaroxaban and IgG spiked into PNP, and ex vivo, on samples from 24 patients with APS from the RAPS trial. LA testing in the in vitro studies comprised: Textarin Ò time [24] (Pentapharm Ltd, Engelgasse, Switzerland); dilute prothrombin time (dpt; American Diagnostica, Sekisui Group, Stamford, CT, USA); dilute Russell s viper venom time (DRVVT) using three different reagents (two detailed below) including an in-house method (Diagen Bell & Alton phospholipid and RVV [Diagnostic Reagents Ltd, Thame, UK] and in-house platelet neutralizing reagent) [25]; and Taipan venom time (TVT)/Ecarin clotting time (ECT; Diagnostic Reagents Ltd). TVT [26,27] was performed using a 1 : 1 dilution in 25 mm CaCl 2, containing 1% w/v bovine serum albumin (A7030; Sigma Aldrich) and Diagen Bell & Alton phospholipid diluted 1 : 7 with imidazole buffer. All tests were completed within 1 h of reagent reconstitution. ECT was performed using ecarin venom diluted 1 : 4 with imidazole buffer. TVT/ECT normalized ratios were calculated with reference to PNP. The following DRVVT tests were performed in both in vitro and ex vivo samples: LA1 screening and LA2 confirmation reagents (Siemens Healthcare Diagnostics Products) and HemosIL drvvt screen and confirm reagents

4 Rivaroxaban and antiphospholipid antibodies 1267 (Instrumentation Laboratory, Warrington, UK). TVT and ECT were also performed in ex vivo studies. DRVVT on samples from patients on warfarin were tested using equal volume (50 : 50) mixtures of patient sample and PNP. In vitro LA studies, dpt, Textarin time and ECT, were performed using a KC4A coagulometer (Tcoag Ireland Ltd., Bray, Ireland). Siemens DRVVT, ECT and TVT were performed using a CS analyser (Sysmex UK Ltd) and HemosIL DRVVT using an ACL TOP 500 (Instrumentation Laboratory). The BCSH guidelines [3] state that two different LA tests should be used. Although one of these is usually the APTT (for convenience), it is not recommended in anticoagulated patients. The APTT shows poor specificity for LA and there is great variability between APTT reagents for sensitivity to rivaroxaban, and therefore the APTT is not included in this study. Effects of apl on rivaroxaban anticoagulant action The effects of apl on the anticoagulant effects of rivaroxaban were assessed using TG testing and anti-xa assays on PNP spiked with four different concentrations of rivaroxaban (0, 25, 50 and 100 ng/ml) and with two concentrations of IgG (500 and 250 lg/ml) from 32 (20 LA positive and 12 LA negative) APS patients and 20 NC. These rivaroxaban concentrations were chosen following preliminary investigation with increasing concentrations of rivaroxaban spiked into PNP on TG to demonstrate measurable clear effects on each parameter of TG (Fig. 2). TG was performed using the CAT system (Thrombinoscope BV, Maastricht, the Netherlands) [28] in conjunction with the PPP reagents (5 pmol/l tissue factor [TF] and 4 lmol/l phospholipid; Diagnostica Stago, Asnieres, France). Rivaroxaban anti-xa levels were Thrombin (nm) Buffer control Rivaroxaban 50 ng ml 1 Rivaroxaban 100 ng ml 1 Rivaroxaban 150 ng ml 1 Rivaroxaban 200 ng ml 1 Rivaroxaban 250 ng ml 1 Rivaroxaban 300 ng ml 1 Rivaroxaban 350 ng ml 1 Rivaroxaban 400 ng ml Time (min) Fig. 2. Thrombin generation with increasing concentration of rivaroxaban in pooled normal plasma. measured using an amidolytic anti-xa assay (Biophen DiXaI, Hyphen BioMed), on a Sysmex CS-2000i analyser (Sysmex UK Ltd); all samples were tested in duplicate. Two quality control samples with known rivaroxaban levels (Hyphen BioMed, C1, rivaroxaban level ng/ ml and C2, rivaroxaban level ng/ml) were also assayed. Statistical analysis Data analysis was performed using Analyse-it software version 2.26 (Analyse-It Software Ltd, Leeds, UK). Results were reported as median or mean based on the distribution of results with 95% CI. For TG parameters, multiple group comparisons were performed using Kruskal Wallis ANOVA and paired comparisons using the Mann Whitney U-test after adjusting the significant P- values by Bonferroni correction. A P value of < 0.05 was considered significant. Results Effects of rivaroxaban on lupus anticoagulant In vitro studies Rivaroxaban caused a concentrationdependent increase in clotting time of PNP with the dpt and DRVVT, but not the TVT, ECT or Textarin time. The dpt screen ratio became abnormal at 2 rivaroxaban, but the dpt confirm ratio increased in parallel, with the screen/confirm ratio (SCR) unchanged. When IgG from LA-negative APS patients and NC was spiked into PNP and rivaroxaban was added, there were no false positives. IgG from 19 of the 20 LA-positive APS patients spiked into PNP remained LA positive by dpt, irrespective of the rivaroxaban concentration (SCR mean [CI] 1.4 [ ] and 1.4 [ ] for 50 and 2, respectively). The remaining patient was LA negative with the dpt method, without spiking with rivaroxaban, and this remained unchanged at both concentrations of rivaroxaban. The in-house DRVVT showed no false positive or negative LA results with either of the rivaroxaban concentrations (Table 1). With the commercial DRVVT reagents, the SCR became abnormal when 2 rivaroxaban was added to PNP spiked with IgG from NC or LA-negative APS patients, giving false positive results in 18/20 NC (90%) and 11/12 LA-negative APS patients (92%). Similar results were observed with both commercial reagents; those obtained with HemosIL reagents are shown in Table 2 as an example. This phenomenon was not observed with rivaroxaban. PNP spiked with LA-positive APS IgG remained LA positive at both levels of rivaroxaban (Table 2). There was no effect of rivaroxaban on Textarin time/ ECT ratio (Table 3) or TVT/ECT ratio and LA was detected at both rivaroxaban levels for all APS IgGspiked PNP samples (TVT/ECT ratio 1.4 [ ], 1.3

5 1268 D. R. J. Arachchillage et al Table 1 Dilute Russell s viper venom time (DRVVT) ratios with in-house method for IgG from the plasma of 20 lupus anticoagulant (LA)- positive and 12 LA-negative antiphospholipid syndrome (APS) patients spiked into pooled normal plasma with rivaroxaban 50 or 2 DRVVT ratio IgG 2 IgG 2 Screen ratio 1.5 ( ) 2.5 ( ) 3.6 ( ) 1.0 ( ) 1.5 ( ) 2.4 ( ) Confirm ratio 1.1 ( ) 1.6 ( ) 2.7 ( ) 1.0 ( ) 1.4 ( ) 2.3 ( ) Screen/ confirm ratio 1.3 ( ) 1.5 ( ) 1.4 ( ) 1.0 ( ) 1.0 ( ) 1.0 ( ) Screen ratio > 1.2 is suggestive of LA positivity and screen/confirm ratio > 1.19 confirms LA. There were no false positive or negative LA results at either rivaroxaban level. Results are given as median with 95% confidence interval. Table 2 Dilute Russell s viper venom time (DRVVT) ratios with a commercial reagent (HemosIL) for IgG from the plasma of 20 lupus anticoagulant (LA)-positive and 12 LA-negative antiphospholipid syndrome (APS) patients spiked into pooled normal plasma with rivaroxaban 50 or 2 DRVVT ratio IgG 2 IgG 2 Screen ratio 1.3 ( ) 1.8 ( ) 2.8 ( ) 1.0 ( ) 1.7 ( ) 2.6 ( ) Confirm ratio 1.1 ( ) 1.5 ( ) 2.1 ( ( ) 1.5 ( ) 2.1 ( ) Screen/ confirm ratio 1.3 ( ) 1.3 ( ) 1.3 ( ) 1.0 ( ) 1.0 ( ) 1.3 ( ) Screen ratio > 1.2 is suggestive of LA positivity and screen/confirm ratio > 1.19 confirms LA. There were false positive LA results at 250 ng/ ml rivaroxaban. 11/12 (92%) IgG from LA-negative APS patients demonstrated false positive LA results at 2. Similar results were observed with Siemens DRVVT reagents as well. Results are given as median with 95% confidence interval. Table 3 Textarin time (s), Ecarin clotting time (ECT) and Textarin time/ect ratios for IgG from the plasma of 20 lupus anticoagulant (LA)- positive and 12 LA-negative antiphospholipid syndrome (APS) patients spiked into pooled normal plasma with rivaroxaban 50 or 2 IgG 2 IgG 2 Textarin time (s) 35.7 ( ) 36.9 ( ) 36.2 ( ) 23.0 ( ) 23.1 ( ) 23.6 ( ) ECT (s) 25.0 ( ) 25.6 ( ) 25.4 ( ) 23.4 ( ) 23.0 ( ) 23.9 ( ) Textarin time/ ECT ratio 1.4 ( ) 1.4 ( ) 1.3 ( ) 0.9 ( ) 0.9 ( ) 1.0 ( ) Textarin time/ect ratio > 1.19 confirms LA. There were no false positive or negative LA results at either rivaroxaban level. Results are given as median with 95% confidence interval. Table 4 Taipan venom time (TVT), Ecarin clotting time (ECT) and TVT/ECT ratios for IgG from the plasma of 20 lupus anticoagulant (LA)- positive and 12 LA-negative antiphospholipid syndrome (APS) patients spiked into pooled normal plasma with rivaroxaban 50 or 2 IgG 2 IgG 2 TVT ratio 1.4 ( ) 1.4 ( ) 1.5 ( ) 1.0 ( ) 1.0 ( ) 1.0 ( ) ECT ratio 1.1 ( ) 1.2 ( ) 1.2 ( ) 1.0 ( ) 1.1 ( ) 1.1 ( ) TVT/ECT ratio 1.3 ( ) 1.2 ( ) 1.3 ( ) 1.0 ( ) 0.9 ( ) 0.9 ( ) Normal TVT and ECT ratios were and , respectively. A TVT/ECT ratio > 1.2 was considered to be positive for LA. There were no false positive or negative LA results. Results are given as median with 95% confidence interval. [ ] and 1.3 [ ] for 0, 50 and 2 rivaroxaban, respectively; normal TVT/ECT < 1.2). There were no false LA-positive results with TVT/ECT (Table 4). Ex vivo studies There were no false positive or negative LA results in the 24 RAPS patients when they were on warfarin, based on their APS-defining LA status determined prior to randomization in RAPS. Six confirmed

6 Rivaroxaban and antiphospholipid antibodies 1269 LA-positive patients receiving rivaroxaban remained LA positive with TVT/ECT and DRVVT reagents when samples were taken at 2 4 h ( ng/ml) following the last dose of rivaroxaban, except for one patient (178 ng/ ml) who appeared LA negative with one reagent. There was a greater prolongation of the screen DRVVT compared with the DRVVT confirm assay with both commercial reagents in patients treated with rivaroxaban at therapeutic levels, causing an increased normalized test/ confirm ratio. This led to six LA-negative patients becoming (apparently) LA positive with two DRVVT reagents (test/confirm ratio median [CI] 1.6 [ ], 1.6 [ ]) (Fig. 3A), but not with TVT/ECT (1.1 [ ]). When A Ratio B Ratio War Riva War Riva War Riva Screen ratio Confirm ratio Screen/confirm ratio War Riva War Riva War Riva Screen ratio Confirm ratio Screen/confirm ratio Fig. 3. Ex vivo dilute Russell s viper venom time (DRVVT) results with Siemens reagents for six patients who were negative for lupus anticoagulant (LA). (A) Paired results for DRVVT screen and confirm and screen/confirm ratios at baseline for patients on warfarin and at least 30 days on rivaroxaban (peak levels) 240 ng/ml (CI, ng/ml). (B) Paired results for DRVVT screen and confirm and screen/confirm ratios at baseline for patients on warfarin and at least 90 days on rivaroxaban (trough levels) 55 ng/ml (CI, ng/ml). HemosIL DRVVT reagents gave a similar pattern of results. Dotted line in each figure indicates the normal cut-off. War = warfarin; Riva = rivaroxaban. samples were taken at least 18 h (36 80 ng/ml) following the last dose of rivaroxaban, the six APS patients positive for LA remained positive with both commercial DRVVT reagents and also with TVT/ECT ratio. Six LA-negative patients receiving rivaroxaban remained LA negative with both DRVVT reagents (Fig. 3B) and with TVT/ECT for samples taken at least 18 h (18 24) after the last dose of rivaroxaban. Effects of apl on the anticoagulant action of rivaroxaban DMSO, OVB and PBS had no influence on TG. PNP spiked with IgG from patients with LA-positive APS (500 lg/ml) caused a prolonged lag time and time to peak thrombin generation compared with LA-negative IgG and NC IgG: mean lag time (min) [CI] 5.2 [ ], P = 0.02, for LA-positive IgG, compared with 3.2 [ ] for LA-negative IgG and 3.1 [ ] for NC; mean time to peak (min) 8.1 [ ], P = 0.01, compared with 6.4 [ ] for LA-negative IgG and 6.4 [ ] for NC (Fig. 4A). There were no differences in peak thrombin and ETP in PNP spiked with LA-positive IgG compared with PNP spiked with LA-negative IgG or NC IgG: mean peak thrombin (nm) [CI] 301 [ ], P = 0.09, for LA-positive IgG compared with 290 [ ] for LA-negative IgG and 287 [ ] for NC; ETP (nm.min) 1738 [ ], P = 0.08, compared with 1690 [ ] for LA-negative IgG and 1682 [ ] for NC (Fig. 4A). The differences in the lag time and time to peak between the three groups remained when PNP/IgG mixtures were also spiked with increasing concentrations of rivaroxaban (25, 50 and 100 ng/ml). However, there was no difference in peak thrombin and ETP between the three groups (Fig. 4B D). There was no difference in the lag time or time to peak in the three groups at 250 lg/ml IgG, demonstrating that the LA effect is dose dependent. When PNP containing 0, 25, 50 or 100 lg/ml rivaroxaban, was spiked with 250 or 500 lg/ml IgG from 20 LA-positive, 12 LA-negative and 20 NC, there was no change in the rivaroxaban concentration measured by anti-xa assay. Mean [CI] rivaroxaban anti-xa levels for 0, 25, 50 and 100 ng/ml PNP, respectively, spiked with IgG from the three groups were: 0, 20 [19 24], 46 [44 48] and 95 [94 98] ng/ml in LA-positive APS; 0, 22 [21 24], 45 [44 47] and 97 [96 98] ng/ml in LA-negative APS; and 0, 23 [21 24], 46 [44 47] and 97 [96 98] ng/ml in NC. Discussion This study provides new insights into interactions between rivaroxaban and apl. Initial studies in vitro (using IgG preparations from thrombotic APS patients) and ex vivo (using plasma from thrombotic APS patients treated with therapeutic-dose rivaroxaban), with peak and trough con-

7 1270 D. R. J. Arachchillage et al A Thrombin (nm) NC lgg + Buffer LA negative lgg + Buffer LA positive lgg + Buffer B Thrombin (nm) Rivaroxaban 25 ng ml 1 + NC lgg Rivaroxaban 25 ng ml 1 + LA negative lgg Rivaroxaban 25 ng ml 1 + LA positive lgg Time (min) Time (min) 40 C 300 Rivaroxaban 50 ng ml 1 + NC lgg D 300 Rivaroxaban 100 ng ml 1 + NC lgg Thrombin (nm) Rivaroxaban 50 ng ml 1 + LA negative lgg Rivaroxaban 50 ng ml 1 + LA positive lgg Thrombin (nm) Rivaroxaban 100 ng ml 1 + LA negative lgg Rivaroxaban 100 ng ml 1 + LA positive lgg Time (min) Time (min) 40 Fig. 4. Example of thrombin generation in pooled normal plasma spiked with IgG from normal controls and lupus anticoagulant (LA)-negative and LA-positive antiphospholipid syndrome patients; IgG 500 lg/ml (A), with increasing concentration of rivaroxaban (25, 50 and 100 ng/ml; B D). centrations of rivaroxaban, established that a false positive DRVVT may occur with rivaroxaban, mainly at peak therapeutic plasma levels ( ng/ml). However, the TVT/ECT ratio and Textarin time were not affected, even at peak rivaroxaban levels, enabling reliable detection of LA. In vitro studies showed no false negative LA with all tests used for LA detection. In vitro studies also established, as hypothesized, that apl do not influence the anticoagulant action of rivaroxaban as assessed by TG testing and anti-xa assays. The present study on detection of LA in the presence of rivaroxaban provides a systematic and detailed approach to LA detection, both in vitro and ex vivo, using well characterized patients with thrombotic APS and defined LA status, and treated with rivaroxaban at both peak and trough levels of rivaroxaban. Limited previous studies provide no in vitro or ex vivo data in thrombotic APS patients, a group that can provide clinically relevant information on LA detection. In an in vitro study by van Os et al. [24], rivaroxaban was spiked into plasma from 13 SLE patients. Six of these patients, who were LA positive, demonstrated that their LA ratio was not influenced by rivaroxaban when tested with TVT/ECT, but the activated partial thromboplastin time (aptt) was slightly increased and was strongly influenced when tested with a commercial DRVVT reagent. In seven LA-negative patients, three became LA positive by aptt, but not by TVT/ECT (the DRVVT was not assessed in the presence of rivaroxaban) [29]. Merriman et al. [30] reported on 19 patients randomized into the EINSTEIN study [12] to receive rivaroxaban 20 mg daily who had no documented LA or APS; these patients exhibited apparent positive LA with a DRVVT screening test when samples were taken between 1 and 19.5 h following the previous dose of rivaroxaban. However, a DRVVT confirm step was not performed in this study, limiting the clinical applicability of the findings. Martinuzzo et al. [31] studied 26 patients for LA, four on therapeutic-dose rivaroxaban and the remainder on a prophylactic dose, who had no previously documented LA or history of APS. Based on

8 Rivaroxaban and antiphospholipid antibodies 1271 DRVVT screen and confirm tests with one commercial reagent, these authors reported that 75% of samples were apparently LA positive by DRVVT, and 100% had a prolonged colloidal silica clotting time screen (SCT) (which employs low phospholipid), with an LA-negative screen/confirm ratio (containing additional phospholipid). The TVT/ECT ratio and Textarin time are not affected even at peak rivaroxaban levels, enabling reliable detection of LA. False positive DRVVT may occur with rivaroxaban, mainly at peak therapeutic plasma levels ( ng/ml). There was a greater prolongation of the screen DRVVT compared with the DRVVT confirm assay with both commercial reagents in patients treated with rivaroxaban at therapeutic levels, causing an increased normalized test/confirm ratio. This led to six LA-negative patients becoming (apparently) LA positive with two DRVVT reagents. In vitro studies on LA detection using a dpt or a DRVVT with in-house reagents, did not produce false positive results as there were parallel increase in the LA screen and LA confirm DRVVT, producing a normal SCR. This is most likely to be due to differences in the phospholipid composition and concentration between the in-house and commercial DRVVT confirm reagents. However, at trough rivaroxaban levels (30 85 ng/ml), for samples taken at least 18 h (18 24) after the last dose of rivaroxaban, previously LA-negative patients (with LA status determined whilst on warfarin) remained LA negative with two commercial DRVVT reagents. These data support the use of the TVT/ECT to detect LA, even at peak plasma rivaroxaban levels, and suggest that the DRVVT may be acceptable at trough rivaroxaban levels. Textarin Ò is a serine proteinase, isolated from the venom of the Australian Brown Snake Pseudonaja textilis, which causes phospholipid, FV and calcium ion (Ca 2+ )-dependent prothrombin activation [24]. As the action of Textarin Ò is phospholipid dependent it may be affected by apl. When performed in parallel with the phospholipid-independent ECT, the ratio of the two clotting times can be used to detect LA. TVT employs a reagent isolated from the venom of the Taipan snake (Oxyuranus scutellatus), which directly activates prothrombin in the presence of phospholipid and calcium ions [26]. Therefore it is independent of the plasma concentration of FV, FX and VII. In contrast, Echis carinatus venom (Ecarin) is able to convert native and PIVKA prothrombin to meizothrombin in the absence of Ca 2+ and phospholipid [24]. These characteristics make Textarin Ò time and TVT, when performed with ECT, useful tests for the detection of LA in the presence of rivaroxaban. Patients with thrombotic APS are inherently different from other patients with VTE by virtue of their apl, particularly LA and anti-prothrombin antibodies [32], which are known to interfere with a number of hemostatic mechanisms, and which could therefore modulate the actions of anticoagulants [33]. It follows that information on the influence of apl on rivaroxaban is clinically relevant in thrombotic APS patients. Our in vitro studies demonstrated that LA led to prolonged lag times and time to peak in a TG system. These observations are in accordance with other reports [34,35]. In vitro studies also showed that apl did not affect the anticoagulant action of rivaroxaban at peak or trough levels, based on TG testing and anti-xa levels. This was predicted as rivaroxaban is a small molecule with high specificity and affinity for its target. These studies were undertaken because of the known effects of apl on hemostasis and therefore potentially on the anticoagulant effects of rivaroxaban. In conclusion, false positive DRVVT may occur in rivaroxaban-treated patients, mainly at peak therapeutic levels. The TVT/ECT ratio and Textarin time are not affected, irrespective of the rivaroxaban levels, enabling detection of LA in patients receiving rivaroxaban. In thrombotic APS patients treated with rivaroxaban, the TVT/ECT appears reliable even at peak therapeutic plasma levels of rivaroxaban. The DRVVT may be acceptable at trough rivaroxaban plasma levels, in samples taken at least 18 h following the previous dose of rivaroxaban. However, in most cases, the laboratory will not know whether trough rivaroxaban levels are present, unless an anti-xa assay is performed. Patients vary in compliance and time keeping, and the exact time of blood collection is not always recorded. The findings in our in vitro and ex vivo studies will guide reliable detection of LA in patients on rivaroxaban. This will enable diagnosis of APS in patients treated with rivaroxaban, as well as monitoring of LA in APS patients, and also inform the development of national and international guidelines. Finally, our in vitro studies indicate that apl do not influence the anticoagulant activity of rivaroxaban as measured by TG testing and rivaroxaban anti-xa levels. Addendum D. R. J. Arachchillage designed the study, performed the assays, acquired data, performed statistical analysis and prepared the first draft of the manuscript, with final approval of the version to be published. I. J. Mackie, M. Efthymiou, D. A. Isenberg, S. J. Machin and H. Cohen contributed to the study design, interpretation of data, critical revision of the intellectual content and final approval of the version to be published. Acknowledgements This study was supported by grants from LUPUS UK and Bayer Healthcare. Rivaroxaban for in vitro spiking was provided by Bayer Healthcare. The RAPS trial is funded by Arthritis Research UK (19807), with supporting funding from the University College London Comprehensive Clinical Trials Unit, LUPUS UK and Bayer plc, who also provided the rivaroxaban free of charge.

9 1272 D. R. J. Arachchillage et al Disclosure of Conflict of Interests The authors state that they have no conflict of interest. References 1 Miyakis S, Lockshin MD, Atsumi T, Branch DW, Brey RL, Cervera R, Derksen RH, de Groot PG, Koike T, Meroni PL, Reber G, Shoenfeld Y, Tincani A, Vlachoyiannopoulos PG, Krilis SA. International consensus statement on an update of the classification criteria for definite antiphospholipid syndrome (APS). J Thromb Haemost 2006; 4: Galli M, Luciani D, Bertolini G, Barbui T. Lupus anticoagulants are stronger risk factors for thrombosis than anticardiolipin antibodies in the antiphospholipid syndrome: a systematic review of the literature. Blood 2003; 101: Keeling D, Mackie I, Moore GW, Greer IA, Greaves M. Guidelines on the investigation and management of antiphospholipid syndrome. Br J Haematol 2012; 157: Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, Pogue J, Reilly PA, Themeles E, Varrone J, Wang S, Alings M, Xavier D, Zhu J, Diaz R, Lewis BS, Darius H, Diener HC, Joyner CD, Wallentin L, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009; 361: Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, Al-Khalidi HR, Ansell J, Atar D, Avezum A, Bahit MC, Diaz R, Easton JD, Ezekowitz JA, Flaker G, Garcia D, Geraldes M, Gersh BJ, Golitsyn S, Goto S, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2011; 365: Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, Breithardt G, Halperin JL, Hankey GJ, Piccini JP, Becker RC, Nessel CC, Paolini JF, Berkowitz SD, Fox KA, Califf RM; ROCKET AF Investigators. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011; 365: Pradaxa 150 mg hard capsules. Summary of product characteristics (SPC), EU. Boehringer Ingelheim International GmBH. 28 January Xarelto 20 mg film-coated tablets. Summary of Product Characteristics last updated on the emc: 6 January NICE technology appraisal guidance 249. Dabigatran etexilate for the prevention of stroke and systemic embolism in atrial fibrillation. March Accessed 17 December NICE technology appraisal guidance 256. Rivaroxaban for the prevention of stroke and systemic embolism in people with atrial fibrillation. May Accessed 18 December Approval Drugs by FDA. Accessed 18 December EINSTEIN Investigators, Bauersachs R, Berkowitz SD, Brenner B, Buller HR, Decousus H, Gallus AS, Lensing AW, Misselwitz F, Prins MH, Raskob GE, Segers A, Verhamme P, Wells P, Agnelli G, Bounameaux H, Cohen A, Davidson BL, Piovella F, Schellong S. Oral rivaroxaban for symptomatic venous thromboembolism. N Engl J Med 2010; 363: EINSTEIN PE Investigators, B uller HR, Prins MH, Lensin AW, Decousus H, Jacobson BF, Minar E, Chlumsky J, Verhamme P, Wells P, Agnelli G, Cohen A, Berkowitz SD, Bounameaux H, Davidson BL, Misselwitz F, Gallus AS, Raskob GE, Schellong S, Segers A. Oral rivaroxaban for the treatment of symptomatic pulmonary embolism. N Engl J Med 2012; 366: Agnelli G, Buller HR, Cohen A, Curto M, Gallus AS, Johnson M, Porcari A, Raskob GE, Weitz JI; PLIFY-EXT Investigators. Apixaban for extended treatment of venous thromboembolism. N Engl J Med 2013; 368: Schulman S, Kearon C, Kakkar AK, Mismetti P, Schellong S, Eriksson H, Baanstra D, Schnee J, Goldhaber SZ; RE-COVER Study Group. Dabigatran versus warfarin in the treatment of acute venous thromboembolism. N Engl J Med 2009; 361: NICE technology appraisal guidance 261. Rivaroxaban for the treatment of deep vein thrombosis and prevention of recurrent deep vein thrombosis and pulmonary embolism. g.uk/ta261. July Accessed 20 December Andreoli L, Chighizola CB, Banzato A, Pons-Estel GJ, de Ramire JG, Erkan D. Estimated frequency of antiphospholipid antibodies in patients with pregnancy morbidity, stroke, myocardial infarction, and deep vein thrombosis: a critical review of the literature. Arthritis Care Res (Hoboken) 2013; 65: Arachchillage DR, Efthymiou M, Mackie IJ, Lawrie AS, Machin SJ, Cohen H. Rivaroxaban and warfarin achieve effective anticoagulation, as assessed by inhibition of TG and in-vivo markers of coagulation activation, in patients with venous thromboembolism. Thromb Res 2015; 135: Pengo V, Tripodi A, Reber G, Rand JH, Ortel TL, Galli M, De Groot PG; Subcommittee on Lupus Anticoagulant/Antiphospholipid Antibody of the Scientific and Standardisation Committee of the International Society on Thrombosis and Haemostasis. Update of the guidelines for lupus anticoagulant detection. Subcommittee on Lupus Anticoagulant/Antiphospholipid Antibody of the Scientific and Standardisation Committee of the International Society on Thrombosis and Haemostasis. J Thromb Haemost 2009; 7: Mueck W, Lensing AW, Agnelli G, Decousus H, Prandoni P, Misselwitz F. Rivaroxaban: population pharmacokinetic analyses in patients treated for acute deep-vein thrombosis and exposure simulations in patients with atrial fibrillation treated for stroke prevention. Clin Pharmacokinet 2011; 50: Molenaar PJ, Dinkelaar J, Leyte A. Measuring Rivaroxaban in a clinical laboratory setting, using common coagulation assays, Xa inhibition and thrombin generation. Clin Chem Lab Med 2012; 50: Perzborn E, Heitmeier S, Laux V, Buchmuller A. Reversal of rivaroxaban-induced anticoagulation with prothrombin complex concentrate, activated prothrombin complex concentrate and recombinant activated factor VII in vitro. Thromb Res 2014; 133: Pericleous C, Clarke LA, Brogan PA, Latchman DS, Isenberg DA, Ioannou Y, Giles IP, Rahman A. Endothelial microparticle release is stimulated in vitro by purified IgG from patients with the antiphospholipid syndrome. Thromb Haemost 2013; 109: Triplett DA, Stocker KF, Unger GA, Barna LK. The Textarin/ Ecarin ratio: a confirmatory test for lupus anticoagulants. Thromb Haemost 1993; 70: Mackie IJ, Machin SJ. Acquired inhibitors of coagulation. In: Laboaratory Haematology, An Account of Laboaratory Techniquies. London: Churchill Livingstone, 1989: Moore GW, Smith MP, Savidge GF. The Ecarin time is an improved confirmatory test for the Taipan snake venom time in warfarinized patients with lupus anticoagulants. Blood Coagul Fibrinolysis 2003; 14: Rooney AM, McNally T, Mackie IJ, Machin SJ. The Taipan snake venom time: a new test for lupus anticoagulant. J Clin Pathol 1994; 47: Hemker HC, Giesen P, Al Dieri R, Regnault V, de Smedt E, Wagenvoord R, Lecompte T, Beguin S. Calibrated automated thrombin generation measurement in clotting plasma. Pathophysiol Haemost Thromb 2003; 33: 4 15.

10 Rivaroxaban and antiphospholipid antibodies van Os GM, de Laat B, Kamphuisen PW, Meijers JC, de Groot PG. Detection of lupus anticoagulant in the presence of rivaroxaban using Taipan snake venom time. J Thromb Haemost 2011; 9: Merriman E, Kaplan Z, Butler J, Malan E, Gan E, Tran H. Rivaroxaban and false positive lupus anticoagulant testing. Thromb Haemost 2011; 105: Martinuzzo ME, Barrera LH, D adamo MA, Otaso JC, Gimenez MI, Oyhamburu J. Frequent false-positive results of lupus anticoagulant tests in plasmas of patients receiving the new oral anticoagulants and enoxaparin. Int J Lab Hematol 2014; 36: Sciascia S, Sanna G, Murru V, Roccatello D, Khamashta MA, Bertolaccini ML. Anti-prothrombin (apt) and anti-phosphatidylserine/prothrombin (aps/pt) antibodies and the risk of thrombosis in the antiphospholipid syndrome. A systematic review. Thromb Haemost 2014; 111: Arachchillage DR, Efthymiou M, Mackie IJ, Lawrie AS, Machin SJ, Cohen H. Anti-protein C antibodies are associated with resistance to endogenous protein C activation and a severe thrombotic phenotype in antiphospholipid syndrome. J Thromb Haemost 2014; 12: Liestol S, Sandset PM, Mowinckel MC, Wisloff F. Activated protein C resistance determined with a thrombin generationbased test is associated with thrombotic events in patients with lupus anticoagulants. J Thromb Haemost 2007; 5: Regnault V, Beguin S, Wahl D, de Maistre, Coenraad HH, Lecompte T. Thrombinography shows acquired resistance to activated protein C in patients with lupus anticoagulants. Thromb Haemost 2003; 89: