Advertisement

Clinical usefulness of measuring prothrombin time and soluble fibrin levels in Japanese patients with atrial fibrillation receiving rivaroxaban

Open ArchivePublished:September 02, 2014DOI:https://doi.org/10.1016/j.jjcc.2014.07.021

      Abstract

      Background

      Rivaroxaban is currently used to prevent stroke in patients with atrial fibrillation. Measuring coagulation function may help clinicians to understand the effects of this drug and the associated risk of bleeding.

      Methods and results

      Rivaroxaban was given to 136 patients with non-valvular atrial fibrillation. Mean age was 74.5 ± 9.0 years (men: 63.2%) and mean CHADS2 score (±SD) was 1.8 ± 1.2. Prothrombin times (PTs) and plasma soluble fibrin (SF) levels were examined in 84 out of 136 patients at baseline and at least 2 weeks thereafter. In 48 patients we were able to collect blood at exact times, namely just before and 3 h after rivaroxaban administration, corresponding to the trough and peak concentrations. Mean peak PT in 48 patients was 17.1 ± 3.6 s and median peak SF level was 1.46 μg/mL. Multiple regression analysis showed that female sex, high brain natriuretic peptide, and high dose were independent factors prolonging the peak PT. Patients with peak PTs ≥20 s experienced significantly more bleeding events. Among 29 of 46 patients newly treated with rivaroxaban without any previous anticoagulant, we examined coagulation function at the exact trough and peak times. In 29 patients, peak PT was significantly more prolonged than the baseline or trough PT (p < 0.001 for both), whereas trough PT was comparable to the baseline PT. In contrast, both trough and peak SF levels in these newly treated patients were significantly reduced than at baseline (p = 0.003 and p < 0.001, respectively).

      Conclusions

      In Japanese patients with non-valvular atrial fibrillation receiving rivaroxaban, a prolonged peak PT (≥20 s) could indicate increased risk of bleeding, and both trough and peak SF levels were reduced relative to baseline. PT and SF are both valuable measures of coagulation status in patients receiving rivaroxaban, regardless of prior anticoagulant history.

      Keywords

      Introduction

      Anticoagulation therapy is mandatory for preventing stroke in patients with atrial fibrillation (AF) [
      • Wolf P.A.
      • Abbott R.D.
      • Kannel W.B.
      Atrial-fibrillation as an independent risk factor for stroke: the Framingham Study.
      ,
      • Gage B.F.
      • Waterman A.D.
      • Shannon W.
      • Boechler M.
      • Rich M.W.
      • Radford M.J.
      Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
      ,
      • Ogawa S.
      • Aizawa Y.
      • Atarashi H.
      • Inoue H.
      • Kamakura S.
      • Koretsune Y.
      • Kumagai K.
      • Mitamura H.
      • Okumura K.
      • Sugi K.
      • Yamashita T.
      • Yasaka M.
      • Hiejima K.
      • Kodama I.
      • Ohe T.
      • et al.
      Guidelines for pharmacotherapy of atrial fibrillation (JCS 2008): digest version.
      ]. The numbers of patients with AF are increasing, making anticoagulation therapy all the more necessary [
      • Go A.S.
      • Hylek E.M.
      • Phillips K.A.
      • Chang Y.C.
      • Henault L.E.
      • Selby J.V.
      • Singer D.E.
      Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study.
      ,
      • Ohsawa M.
      • Okayama A.
      • Sakata K.
      • Kato K.
      • Itai K.
      • Onoda T.
      • Ueshima H.
      Rapid increase in estimated number of persons with atrial fibrillation in Japan: an analysis from national surveys on cardiovascular diseases in 1980, 1990 and 2000.
      ,
      • Akao M.
      • Chun Y.H.
      • Wada H.
      • Esato M.
      • Hashimoto T.
      • Abe M.
      • Hasegawa K.
      • Tsuji H.
      • Furuke K.
      Current status of clinical background of patients with atrial fibrillation in a community-based survey: the Fushimi AF Registry.
      ]. Over the past 50 years, warfarin has been used in oral anticoagulation therapy, but there are several problems associated with its use. For example, administration of warfarin requires: (1) a fixed period to become effective; (2) adequate patient education for proper compliance; and (3) regular monitoring of the prothrombin time (PT) – international normalized ratio to maintain drug safety and efficacy [
      • Koudstaal P.J.
      • Vanlatum J.C.
      • Vangijn J.
      • Kappelle L.J.
      • Algra A.
      • Dehaene I.
      • Dhooghe M.
      • Marchau M.
      • Vanzandijcke M.
      • Delwaide C.
      • Depre A.
      • Laterre E.C.
      • Vanbuggenhout E.
      • Schurmans J.
      • Desmet E.
      • et al.
      Optimal oral anticoagulant-therapy in patients with nonrheumatic atrial fibrillation and recent cerebral ischemia.
      ,
      • Yasaka M.
      • Minematsu K.
      • Yamaguchi T.
      Optimal intensity of international normalized ratio in warfarin therapy for secondary prevention of stroke in patients with non-valvular atrial fibrillation.
      ,
      • Inoue H.
      • Okumura K.
      • Atarashi H.
      • Yamashita T.
      • Origasa H.
      • Kumagai N.
      • Sakurai M.
      • Kawamura Y.
      • Kubota I.
      • Matsumoto K.
      • Kaneko Y.
      • Ogawa S.
      • Aizawa Y.
      • Chinushi M.
      • Kodama I.
      • et al.
      Target international normalized ratio values for preventing thromboembolic and hemorrhagic events in Japanese patients with non-valvular atrial fibrillation: results of the J-RHYTHM Registry.
      ].
      Recently, novel oral anticoagulants (NOACs) have been proven to be non-inferior to warfarin in terms of preventing ischemic stroke, and they may be superior with respect to the risk of major bleeding [
      • Patel M.R.
      • Mahaffey K.W.
      • Garg J.
      • Pan G.
      • Singer D.E.
      • Hacke W.
      • Breithardt G.
      • Halperin J.L.
      • Hankey G.J.
      • Piccini J.P.
      • Becker R.C.
      • Nessel C.C.
      • Paolini J.F.
      • Berkowitz S.D.
      • Fox K.A.
      • et al.
      Rivaroxaban versus warfarin in nonvalvular atrial fibrillation.
      ,
      • Connolly S.J.
      • Ezekowitz M.D.
      • Yusuf S.
      • Eikelboom J.
      • Oldgren J.
      • Parekh A.
      • Pogue J.
      • Reilly P.A.
      • Themeles E.
      • Varrone J.
      • Wang S.
      • Alings M.
      • Xavier D.
      • Zhu J.
      • Diaz R.
      • et al.
      Dabigatran versus warfarin in patients with atrial fibrillation.
      ,
      • Granger C.B.
      • Alexander J.H.
      • McMurray J.J.V.
      • Lopes R.D.
      • Hylek E.M.
      • Hanna M.
      • Al-Khalidi H.R.
      • Ansell J.
      • Atar D.
      • Avezum A.
      • Cecilia Bahit M.
      • Diaz R.
      • Easton J.D.
      • Ezekowitz J.A.
      • Flaker G.
      • et al.
      Apixaban versus warfarin in patients with atrial fibrillation.
      ,
      • Heidbuchel H.
      • Verhamme P.
      • Alings M.
      • Antz M.
      • Hacke W.
      • Oldgren J.
      • Sinnaeve P.
      • Camm A.J.
      • Kirchhof P.
      European heart rhythm association practical guide on the use of new oral anticoagulants in patients with non-valvular atrial fibrillation.
      ]. Moreover, NOACs do not require frequent monitoring of their anticoagulation effects. However, these findings were obtained in large-scale clinical trials and are applicable only to the types of patients who met the inclusion criteria for each trial. In an actual clinical setting, patients with heart failure or poor medication adherence also need to be treated with anticoagulants. Furthermore, coagulation function must be measured in many other instances, such as in patients on anticoagulants undergoing emergency surgery or exhibiting hemorrhage.
      Rivaroxaban was the second NOAC to be approved worldwide after dabigatran [
      • Patel M.R.
      • Mahaffey K.W.
      • Garg J.
      • Pan G.
      • Singer D.E.
      • Hacke W.
      • Breithardt G.
      • Halperin J.L.
      • Hankey G.J.
      • Piccini J.P.
      • Becker R.C.
      • Nessel C.C.
      • Paolini J.F.
      • Berkowitz S.D.
      • Fox K.A.
      • et al.
      Rivaroxaban versus warfarin in nonvalvular atrial fibrillation.
      ,
      • Hori M.
      • Matsumoto M.
      • Tanahashi N.
      • Momomura S.
      • Uchiyama S.
      • Goto S.
      • Izumi T.
      • Koretsune Y.
      • Kajikawa M.
      • Kato M.
      • Ueda H.
      • Iwamoto K.
      • Tajiri M.
      Rivaroxaban vs. warfarin in Japanese patients with atrial fibrillation – the J-ROCKET AF study.
      ]. Data on coagulation function in patients treated with rivaroxaban are limited [
      • Hillarp A.
      • Baghaei F.
      • Blixter I.F.
      • Gustafsson K.M.
      • Stigendal L.
      • Sten-Linder M.
      • Strandberg K.
      • Lindahl T.L.
      Effects of the oral, direct factor Xa inhibitor rivaroxaban on commonly used coagulation assays.
      ,
      • Lindhoff-Last E.
      • Samama M.M.
      • Ortel T.L.
      • Weitz J.I.
      • Spiro T.E.
      Assays for measuring rivaroxaban: their suitability and limitations.
      ,
      • Samama M.M.
      • Martinoli J.L.
      • LeFlem L.
      • Guinet C.
      • Plu-Bureau G.
      • Depasse F.
      • Perzborn E.
      Assessment of laboratory assays to measure rivaroxaban – an oral, direct factor Xa inhibitor.
      ,
      • Molenaar P.J.
      • Dinkelaar J.
      • Leyte A.
      Measuring rivaroxaban in a clinical laboratory setting, using common coagulation assays, Xa inhibition and thrombin generation.
      ], and only Yamashita's group have reported the distribution of PT in Japanese AF patients treated with rivaroxaban [
      • Suzuki S.
      • Otsuka T.
      • Sagara K.
      • Kano H.
      • Matsuno S.
      • Kato Y.
      • Takai H.
      • Uejima T.
      • Oikawa Y.
      • Nagashima K.
      • Kirigaya H.
      • Yajima J.
      • Sawada H.
      • Aizawa T.
      • Yamashita T.
      Rivaroxaban in clinical practice for atrial fibrillation with special reference to prothrombin time.
      ]. Here, we measured PTs to determine the safety profile of rivaroxaban; we chose PT because it is the most widely used coagulation test. In addition, we measured plasma levels of soluble fibrin (SF), which is generated from thrombin–fibrinogen reactions and is a potentially useful marker of hypercoagulability [
      • Dempfle C.E.
      The use of soluble fibrin in evaluating the acute and chronic hypercoagulable state.
      ,
      • Sudo A.
      • Wada H.
      • Nobori T.
      • Yamada N.
      • Ito M.
      • Niimi R.
      • Hasegawa M.
      • Suzuki K.
      • Uchida A.
      Cut-off values of d-dimer and soluble fibrin for prediction of deep vein thrombosis after orthopaedic surgery.
      ,
      • Tsuji A.
      • Wada H.
      • Matsumoto T.
      • Abe Y.
      • Ota S.
      • Yamada N.
      • Sugiyama T.
      • Sudo A.
      • Onishi K.
      • Nakatani K.
      • Uchida A.
      • Ito M.
      • Suzuki K.
      • Nobori T.
      Elevated levels of soluble fibrin in patients with venous thromboembolism.
      ].

      Materials and methods

      Study design and participants

      This observational study of Japanese patients with non-valvular AF treated with rivaroxaban was conducted in routine clinical practice at Tosei General Hospital, Aichi, Japan. From 1 May 2012 to 31 May 2013, a total of 136 patients began treatment with rivaroxaban and were considered for study entry. Coagulation function was examined in 84 of these patients at baseline and at least 2 weeks later. In 48 of these 84 patients, we were able to collect blood accurately at times corresponding to the trough and peak concentrations of rivaroxaban. Patients gave written informed consent to participate, and the study was conducted in accordance with the ethical policy of Tosei General Hospital. We also performed outcome and safety evaluations in all 136 patients treated with rivaroxaban. The observation period lasted till 31 December 2013.

      Dose of rivaroxaban

      Patients received oral rivaroxaban at a dose of either 15 mg once daily (O.D.) or 10 mg O.D. Patients took rivaroxaban after breakfast, and their adherence to medication was monitored. Each dose was determined by the attending physician on the basis of the creatinine clearance (CCr), determined with the Cockcroft–Gault formula, i.e. men, [(140 − age)/(serum creatinine)] × (weight/72); women, 0.85 × [(140 − age)/(serum creatinine)] × (weight/72)], or of the presence of a previous history of bleeding complications or concurrent use of antiplatelet drugs, or both. Patients were allocated to three groups: patients who received 15 mg O.D. and had CCr ≥ 50 mL/min; those who received 10 mg O.D. and had CCr < 50 mL/min; and those who received 10 mg O.D. and had CCr ≥ 50 mL/min.

      Measuring coagulation function

      PTs and SF levels were examined at baseline (before the start of rivaroxaban treatment) and at trough and peak times at least 2 weeks after the start of treatment. We defined the trough time as immediately before the administration of rivaroxaban and the peak time as 3 h after drug administration. Therefore, patients brought the drug with them and then took it orally at the hospital after the first blood collection in the morning. Blood was collected again 3 h later. PT was determined with the recombinant human tissue factor-based thromboplastin reagent (HemosIL RecombiPlasTin, Instrumentation Laboratory, Lexington, MA, USA). SF was quantified by latex photometric immunoassay (IATRO SF II, LSI Medience Corporation, Tokyo, Japan).

      Outcomes

      Outcomes were defined in accordance with the ROCKET AF study and J-ROCKET AF study [
      • Patel M.R.
      • Mahaffey K.W.
      • Garg J.
      • Pan G.
      • Singer D.E.
      • Hacke W.
      • Breithardt G.
      • Halperin J.L.
      • Hankey G.J.
      • Piccini J.P.
      • Becker R.C.
      • Nessel C.C.
      • Paolini J.F.
      • Berkowitz S.D.
      • Fox K.A.
      • et al.
      Rivaroxaban versus warfarin in nonvalvular atrial fibrillation.
      ,
      • Hori M.
      • Matsumoto M.
      • Tanahashi N.
      • Momomura S.
      • Uchiyama S.
      • Goto S.
      • Izumi T.
      • Koretsune Y.
      • Kajikawa M.
      • Kato M.
      • Ueda H.
      • Iwamoto K.
      • Tajiri M.
      Rivaroxaban vs. warfarin in Japanese patients with atrial fibrillation – the J-ROCKET AF study.
      ]. Outcome events included stroke or any other form of embolism in a major organ. In the safety evaluation we counted major and non-major clinically relevant bleeding events. A major bleeding event was defined as clinically overt bleeding associated with a decrease in hemoglobin level of ≥20 g/L, transfusion of ≥2 units of packed red blood cells or whole blood, involvement of an intracranial site, or a fatal outcome. A non-major clinically relevant bleeding event was defined as clinically overt bleeding that did not meet the criteria for major bleeding but that required medical intervention or resulted in the impairment of daily activities.

      Statistical analysis

      Categorical variables are presented as numbers and percentages (%). Continuous variables are presented as means ± standard deviation or medians. Pearson's correlation was used to measure the strength of the relationship between two variables. To compare PTs between groups we used the unpaired t-test, paired t-test, chi-square test, or one-way ANOVA. To compare SF levels between groups we used the Wilcoxon signed-rank test or Kruskal–Wallis test. We performed stepwise multiple regression analysis to evaluate the factors affecting peak PT. A p-value < 0.05 was considered statistically significant. All statistical analyses were conducted with the SPSS statistical software program (SPSS version 18.0 for Windows, SPSS Inc., Chicago, IL, USA).

      Results

      Patient characteristics

      Patient characteristics are shown in Table 1. Mean age of all patients (n = 136) was 74.5 years, and 63.2% were men. Eighty-five patients suffered from paroxysmal AF and 51 suffered from persistent or long-standing persistent AF. CCr < 50 mL/min was observed in 30.2% of all patients. During the study, 15.4% of all patients were taking antiplatelet drugs concurrently.
      Table 1Characteristics of the patients.
      Overall (n = 136)Patients examined for coagulation function (n = 84)
      Age (years)74.5 ± 9.076.0 ± 8.5
       Age ≥75 years, n (%)70 (51.5)49 (58.3)
      Sex, n (%)
       Male86 (63.2)52 (61.9)
       Female50 (36.8)32 (38.1)
      Body mass index23.4 ± 4.123.1 ± 4.2
      Body weight (kg)59.4 ± 11.958.4 ± 12.3
      Daily dose (mg O.D.)12.1 ± 2.511.8 ± 2.4
       10 mg O.D., n (%)78 (57.4)53 (63.1)
       15 mg O.D., n (%)58 (42.6)31 (36.9)
      Previous anticoagulants
       None, n (%)81 (59.6)46 (54.8)
       Warfarin, n (%)32 (23.5)22 (26.2)
       Dabigatran, n (%)13 (9.6)9 (10.7)
       Unfractionated heparin, n (%)10 (7.4)7 (8.3)
      Concurrent antiplatelet drugs
       Aspirin, n (%)8 (5.9)5 (6.0)
       Clopidogrel, n (%)1 (0.7)1 (1.2)
       Cilostazol, n (%)2 (1.5)1 (1.2)
       Aspirin + clopidogrel, n (%)9 (6.6)7 (8.3)
       Aspirin + cilostazol, n (%)1 (0.7)0 (0)
      CHADS2 score
       Mean1.8 ± 1.21.9 ± 1.1
       0, n (%)19 (14.0)9 (10.7)
       1, n (%)38 (27.9)22 (26.2)
       2, n (%)41 (30.1)33 (39.3)
       ≥3, n (%)38 (27.9)20 (23.8)
      Congestive heart failure, n (%)36 (26.5)24 (28.6)
      Hypertension, n (%)93 (68.4)60 (71.4)
      Diabetes mellitus, n (%)22 (16.2)11 (13.1)
      Baseline stroke/systemic embolism, n (%)14 (10.3)6 (7.1)
      Baseline creatinine clearance
       Mean (mL/min)66.5 ± 27.963.8 ± 27.6
       15–29 mL/min, n (%)10 (7.4)7 (8.3)
       30–<50 mL/min, n (%)31 (22.8)19 (22.6)
       50–<80 mL/min, n (%)55 (40.4)39 (46.4)
       ≥80 mL/min, n (%)40 (29.4)19 (22.6)
      Dose and creatinine clearance
       15 mg O.D., CCr ≥50 mL/min58 (42.6)31 (36.9)
       10 mg O.D., CCr <50 mL/min39 (28.7)25 (29.8)
       10 mg O.D., CCr ≥50 mL/min39 (28.7)28 (33.3)
      O.D., once daily; CCr, creatinine clearance.

      Prothrombin time

      The mean trough and peak PTs in the 48-patient group were 11.1 ± 1.0 s and 17.1 ± 3.6 s, respectively (Fig. 1a). Peak PTs were distributed normally and were more significantly prolonged than trough PTs (p < 0.001); the trough PTs were comparable to the normal value used in our hospital laboratory (10.9 ± 0.9 s). PTs were compared among the three groups, namely 15 mg O.D. with CCr ≥ 50 mL/min, 10 mg O.D. with CCr < 50 mL/min, and 10 mg O.D. with CCr ≥ 50 mL/min. Baseline PTs were widely distributed within all three groups (12.7 ± 4.2 s, 11.4 ± 1.9 s, and 13.0 ± 2.6 s, respectively) and depended on prior anticoagulant therapy. Trough PTs were comparable among the three groups (11.0 ± 1.0 s, 11.1 ± 0.8 s, and 11.2 ± 1.4 s, respectively; p = 0.850) (Fig. 1b) and were not significantly prolonged in each group compared with the normal values used in our hospital laboratory (data not shown). Peak PTs did not differ significantly among the three groups (17.7 ± 3.4 s, 16.7 ± 3.7 s, and 16.9 ± 3.8 s, respectively; p = 0.725) (Fig. 1c).
      Figure thumbnail gr1
      Fig. 1(a) Distributions of prothrombin time (PT) at trough and peak. Mean values at trough and peak PT were 11.1 ± 1.0 s and 17.1 ± 3.6 s, respectively. (b and c) Trough and peak PTs in each group according to CCr and dose. PT, prothrombin time; O.D., once daily; CCr, creatinine clearance. Horizontal lines in boxes represent medians. Tops and bottoms of boxes indicate 75th and 25th percentile, respectively. Tops and bottoms of bars indicate maximum and minimum non-outliers, respectively. Circles indicate outliers above or below 1.5 times the interquartile range from the 75th to the 25th percentile.
      In the 48-patient group, there was no correlation overall between peak PT and CCr (r = −0.012, p = 0.933; Fig. 2). Moreover, there was no correlation between peak PT and CCr in any of the three groups [15 mg O.D. with CCr ≥ 50 mL/min (r = −0.01, p = 0.972), 10 mg O.D. with CCr < 50 mL/min (r = −0.31, p = 0.196), 10 mg O.D. with CCr ≥ 50 mL/min (r = −0.173, p = 0.553)]. B-type natriuretic peptide (BNP) was negatively correlated with CCr (r = −0.416, p < 0.001). We also performed a stepwise multiple regression analysis to evaluate the factors affecting peak PT (Table 2). Female sex, high BNP, and high dose were significantly related to peak PT, whereas CCr was not. There were no differences in CCr between male and female patients (males: 63.9 mL/min; females: 63.8 mL/min, p = 0.987). Female patients tended to exhibit more prolonged peak PTs than males (males, 16.2 ± 3.4 s; females, 18.1 ± 3.5 s, p = 0.058), even though more males than females were receiving the higher dose (15 mg O.D.) (p = 0.007) and the CCr was comparable between the sexes.
      Figure thumbnail gr2
      Fig. 2Relationship between peak prothrombin time (PT) and creatinine clearance (CCr). There was no correlation between peak PT and CCr (r = −0.012, p = 0.933) in the overall group of patients examined for peak PT or in each of the three groups.
      Table 2Factors that prolonged peak prothrombin time.
      Variableβp-Value
      Female sex0.3700.011
      p<0.05.
      Log BNP0.3300.020
      p<0.05.
      Dose0.3280.026
      p<0.05.
      CCr−0.0680.681
      BNP, brain natriuretic peptide; CCr, creatinine clearance.
      * p < 0.05.

      Outcome assessment

      Outcomes were assessed in a total of 136 patients treated with rivaroxaban. One embolism event was observed. This patient was an 82-year-old male whose CCr was 53.6 mL/min and who was receiving 10 mg O.D. for secondary prevention of cerebral infarction. Unfortunately, coagulation function was not measured before the onset of the event; it was measured only on the day of the event when the patient was hospitalized (PT 11.4 s).
      Bleeding events were observed in 25 patients (Table 3), where 23 of whom were over 70 years. Major bleeding events were observed in four patients and consisted of two cases of intracranial hemorrhage, one of hypermenorrhea, and one of upper gastrointestinal tract bleeding. All four patients were being treated with 15 mg O.D. and all patients had CHADS2 scores ≤1. Minor bleeding events were observed in 21 patients and included gastrointestinal bleeding, gingival bleeding, nasal bleeding, macrohematuria, subcutaneous bleeding, subcutaneous hematoma formation following cardiac catheterization, hemosputum, and atypical genital bleeding.
      Table 3Characteristics of the patients who suffered bleeding events.
      Patients with major bleeding (n = 4)Patients with non-major bleeding (n = 21)Patients without bleeding events (n = 111)
      Median days of administration period (25th–75th percentile)208 (166–272)266 (27–496)230 (69–292)
      Median days to event (25th–75th percentile)198 (139–238)47 (24–233)
      Age (years)61.8 ± 11.781.9 ± 6.773.5 ± 8.4
      Sex, n
       Male11075
       Female31136
      Body mass index23.7 ± 2.223.3 ± 4.123.4 ± 4.2
      Body weight (kg)63.2 ± 9.956.2 ± 12.859.9 ± 11.8
      Concurrent antiplatelet drugs, n0615
      CHADS2 score0.8 ± 0.52.3 ± 0.91.8 ± 1.3
      CCr (mL/min)98.4 ± 22.748.5 ± 20.068.8 ± 27.5
      Dose and CCr, n
       15 mg O.D., CCr ≥50 mL/min4549
       10 mg O.D., CCr <50 mL/min0930
       10 mg O.D., CCr ≥50 mL/min0732
      CCr, creatinine clearance; O.D., once daily.

      Bleeding events and prothrombin time

      Accurate coagulation function tests of peak PT were performed in 15 of the 25 patients who suffered bleeding events. Among those patients, the mean trough PT was 11.4 ± 1.2 s and the mean peak PT was 18.4 ± 4.2 s. Receiver operating characteristic analysis showed that the peak PT with 80% sensitivity for a bleeding event was 20.8 s. We therefore allocated the patients to two groups, with peak PTs of ≥20 s or <20 s. Patients with peak PTs ≥20 s experienced significantly more bleeding events than those with peak PTs <20 s (62.5% vs. 22.7%, p = 0.022). The trough PTs were comparable between patients with and without bleeding events.

      Soluble fibrin

      In the 48-patient group, the median values for baseline, trough, and peak SF were 1.92, 1.72, and 1.46 μg/mL, respectively (Fig. 3a). The trough and peak SF levels were more significantly reduced than the baseline levels (p = 0.006 and p = 0.001, respectively), especially in patients who had extremely high SF levels at baseline. The ratios of SF reduction from baseline to peak (peak SF/baseline SF) did not differ significantly among patients treated with 15 mg O.D. (CCr ≥ 50 mL/min) or 10 mg O.D. (CCr < 50 mL/min or ≥ 50 mL/min) (Fig. 3b). There were no significant correlations between the peak PT and SF levels (data not shown).
      Figure thumbnail gr3
      Fig. 3(a) Distribution of baseline, trough, and peak plasma soluble fibrin (SF) levels. The trough and peak SF levels (1.72 μg/mL and 1.46 μg/mL, respectively) were significantly reduced compared to baseline values (1.92 μg/mL). (b) SF reduction ratio from baseline to peak in each group according to creatinine clearance (CCr) and dose. Horizontal lines in boxes represent medians. Tops and bottoms of boxes indicate 75th and 25th percentile, respectively. Tops and bottoms of bars indicate maximum and minimum non-outliers, respectively. Circles indicate outliers above or below 1.5 times the interquartile range from the 75th to the 25th percentile.

      Coagulation function in patients treated with rivaroxaban without previous anticoagulants

      From among the 46 patients receiving rivaroxaban who had not been treated with any previous anticoagulant, we were able to collect blood accurately at trough and peak from 29. PT was significantly prolonged at peak (16.9 ± 3.2 s) than at baseline (10.6 ± 1.0 s, p < 0.001) or trough (10.9 ± 0.9 s, p < 0.001) (Fig. 4a) in these 29 patients. PT was not significantly prolonged at trough than the baseline value and the normal value used in our hospital laboratory (10.9 ± 0.2 s).
      Figure thumbnail gr4
      Fig. 4Changes in coagulation function from baseline to trough and peak in 29 patients receiving rivaroxaban without any previous anticoagulant therapy. (a) Prothrombin time (PT) was significantly prolonged at peak (16.9 ± 3.2 s) compared to that observed at baseline (10.6 ± 1.0 s) or at trough (10.9 ± 0.9 s), whereas the trough value was comparable to the baseline. (b) The baseline median soluble fibrin (SF) level was 1.90 μg/mL. Following the initiation of treatment with rivaroxaban, the SF levels at both trough (1.47 μg/mL) and peak (1.45 μg/mL) were significantly reduced compared to those observed at baseline. Horizontal lines in boxes represent medians. Tops and bottoms of boxes indicate 75th and 25th percentile, respectively. Tops and bottoms of bars indicate maximum and minimum non-outliers, respectively. Circles indicate outliers above or below 1.5 times the interquartile range from the 75th to the 25th percentile.
      The SF level at baseline was distributed widely, with a median of 1.90 μg/mL (Fig. 4b) in these 29 patients. Following the initiation of treatment with rivaroxaban, the SF levels at both trough and peak were more significantly reduced than at baseline (trough, 1.47 μg/mL; p = 0.003; peak, 1.45 μg/mL; p < 0.001) in these 29 patients. In addition, SF levels at trough and peak were comparable (p = 0.314). Although a few patients displayed high SF levels at baseline, their trough and peak SF levels dropped dramatically within a short period after administration of the daily dose.

      Discussion

      Our main findings were as follows: (1) the mean peak PT was 17.1 s during rivaroxaban treatment. (2) Female sex, high BNP, and high dose were the independent factors that affect prolonged peak PT. (3) Patients who had peak PTs ≥20 s experienced significantly more bleeding events. (4) Among the patients receiving rivaroxaban who had not received anticoagulants previously, the peak PT was more significantly prolonged than the trough and baseline PT values, whereas the trough PT was similar to the baseline value and to the normal value used in our hospital laboratory. (5) The SF levels in patients receiving rivaroxaban without any previous anticoagulant were significantly reduced at both trough and peak compared to baseline; however, there were no significant differences between the trough and peak SF levels.
      Our results for the distribution of PT among Japanese patients receiving rivaroxaban are similar to those previously reported [
      • Suzuki S.
      • Otsuka T.
      • Sagara K.
      • Kano H.
      • Matsuno S.
      • Kato Y.
      • Takai H.
      • Uejima T.
      • Oikawa Y.
      • Nagashima K.
      • Kirigaya H.
      • Yajima J.
      • Sawada H.
      • Aizawa T.
      • Yamashita T.
      Rivaroxaban in clinical practice for atrial fibrillation with special reference to prothrombin time.
      ]. As with other NOACs, rivaroxaban normally does not require routine coagulation function testing. In routine medical practice, the dose of rivaroxaban is determined according to the CCr and a fixed amount of the drug is administered [
      • Fox K.A.A.
      • Piccini J.P.
      • Wojdyla D.
      • Becker R.C.
      • Halperin J.L.
      • Nessel C.C.
      • Paolini J.F.
      • Hankey G.J.
      • Mahaffey K.W.
      • Patel M.R.
      • Singer D.E.
      • Califf R.M.
      Prevention of stroke and systemic embolism with rivaroxaban compared with warfarin in patients with non-valvular atrial fibrillation and moderate renal impairment.
      ,
      • Tanigawa T.
      • Kaneko M.
      • Hashizume K.
      • Kajikawa M.
      • Ueda H.
      • Tajiri M.
      • Paolini J.F.
      • Mueck W.
      Model-based dose selection for phase III rivaroxaban study in Japanese patients with non-valvular atrial fibrillation.
      ]. However, coagulation status must be evaluated in some patients on anticoagulants, such as those with complications of heart failure, poor medication adherence, elevated plasma concentrations of the drug, or hemorrhage, or undergoing emergency surgery.
      PT is used widely to evaluate coagulation function in the clinical setting. The results of anti-factor-Xa assay are correlated with the concentration of rivaroxaban, but this assay cannot be easily performed in all hospitals [
      • Hillarp A.
      • Baghaei F.
      • Blixter I.F.
      • Gustafsson K.M.
      • Stigendal L.
      • Sten-Linder M.
      • Strandberg K.
      • Lindahl T.L.
      Effects of the oral, direct factor Xa inhibitor rivaroxaban on commonly used coagulation assays.
      ,
      • Lindhoff-Last E.
      • Samama M.M.
      • Ortel T.L.
      • Weitz J.I.
      • Spiro T.E.
      Assays for measuring rivaroxaban: their suitability and limitations.
      ,
      • Samama M.M.
      • Martinoli J.L.
      • LeFlem L.
      • Guinet C.
      • Plu-Bureau G.
      • Depasse F.
      • Perzborn E.
      Assessment of laboratory assays to measure rivaroxaban – an oral, direct factor Xa inhibitor.
      ,
      • Molenaar P.J.
      • Dinkelaar J.
      • Leyte A.
      Measuring rivaroxaban in a clinical laboratory setting, using common coagulation assays, Xa inhibition and thrombin generation.
      ]. PT is a candidate measure for evaluating whether coagulation function remains in the safe range in each patient receiving rivaroxaban. In patients receiving rivaroxaban, peak PTs are naturally prolonged. In our female patients, the peak PT was marginally longer than in males. One of the reasons may have been the fact that body surface area was significantly smaller in female patients than in males (males: 1.7 m2; females: 1.4 m2, p < 0.001). The blood concentration of rivaroxaban could be higher in patients with smaller body surface areas, thus leading to a more prolonged PT value. One-third of rivaroxaban is excreted from the kidneys unaltered. However, no correlation was seen between CCr and peak PT; this may have been related to the fact that the CCr value was used by the attending physician to adjust the dose.
      We also did not find any statistical relationship between peak PT and age. This may have been due to CCr, which includes age as a variable, and was used in determining the dose. Female sex, high BNP, and high dose were independent factors affecting peak PT. However, these results should be interpreted with caution because of the small sample size.
      Among patients whose coagulation function was measured accurately, those with peak PTs ≥20 s experienced bleeding events more frequently than did those with lower values. Although the number of events was small, clinicians should be aware that rivaroxaban patients with peak PTs ≥20 s may be at increased risk of bleeding.
      Elevated levels of SF are indicative of a high risk of deep vein thrombosis [
      • Dempfle C.E.
      The use of soluble fibrin in evaluating the acute and chronic hypercoagulable state.
      ,
      • Sudo A.
      • Wada H.
      • Nobori T.
      • Yamada N.
      • Ito M.
      • Niimi R.
      • Hasegawa M.
      • Suzuki K.
      • Uchida A.
      Cut-off values of d-dimer and soluble fibrin for prediction of deep vein thrombosis after orthopaedic surgery.
      ,
      • Tsuji A.
      • Wada H.
      • Matsumoto T.
      • Abe Y.
      • Ota S.
      • Yamada N.
      • Sugiyama T.
      • Sudo A.
      • Onishi K.
      • Nakatani K.
      • Uchida A.
      • Ito M.
      • Suzuki K.
      • Nobori T.
      Elevated levels of soluble fibrin in patients with venous thromboembolism.
      ]; patients with SF levels of <3 to <6 μg/mL are less likely to develop deep vein thrombosis than those with higher levels [
      • Sudo A.
      • Wada H.
      • Nobori T.
      • Yamada N.
      • Ito M.
      • Niimi R.
      • Hasegawa M.
      • Suzuki K.
      • Uchida A.
      Cut-off values of d-dimer and soluble fibrin for prediction of deep vein thrombosis after orthopaedic surgery.
      ,
      • Tsuji A.
      • Wada H.
      • Matsumoto T.
      • Abe Y.
      • Ota S.
      • Yamada N.
      • Sugiyama T.
      • Sudo A.
      • Onishi K.
      • Nakatani K.
      • Uchida A.
      • Ito M.
      • Suzuki K.
      • Nobori T.
      Elevated levels of soluble fibrin in patients with venous thromboembolism.
      ]. However, there have as yet been no reports of SF levels in AF patients treated with NOACs. We used IATRO SF II™ to measure SF levels, by which the manufacturer guaranteed the linearity between 1.57 μg/mL and 7.9 μg/mL. Therefore, SF levels above approximately 1.5 μg/mL are reliable. Here, we measured SF levels at the initiation of therapy to evaluate the distribution of SF levels in patients receiving rivaroxaban. We found that trough and peak SF levels were substantially reduced after a few weeks of therapy, unlike the reported SF levels in patients with deep vein thrombosis [
      • Sudo A.
      • Wada H.
      • Nobori T.
      • Yamada N.
      • Ito M.
      • Niimi R.
      • Hasegawa M.
      • Suzuki K.
      • Uchida A.
      Cut-off values of d-dimer and soluble fibrin for prediction of deep vein thrombosis after orthopaedic surgery.
      ,
      • Tsuji A.
      • Wada H.
      • Matsumoto T.
      • Abe Y.
      • Ota S.
      • Yamada N.
      • Sugiyama T.
      • Sudo A.
      • Onishi K.
      • Nakatani K.
      • Uchida A.
      • Ito M.
      • Suzuki K.
      • Nobori T.
      Elevated levels of soluble fibrin in patients with venous thromboembolism.
      ]. This finding suggests that SF levels might be suitable for estimating the efficacy of rivaroxaban in patients at the beginning of therapy.
      Considerable numbers of patients with CCr ≥50 mL/min received 10 mg O.D. in consideration of the presence of a previous history of bleeding complications or concurrent use of antiplatelet drugs. In fact, there will always be some at-risk patients with CCr ≥50 mL/min in whom the 10 mg O.D. dose should be considered. In terms of SF levels and SF reduction ratio, we found no difference between the three groups allocated by dose and CCr. However, care should be taken to ensure that use of the lower dose is appropriate; notably, the patient who suffered the one embolic event in this study was on 10 mg O.D.
      As a subanalysis, we examined a group of patients who were receiving rivaroxaban and had not been treated previously with anticoagulants. In this group, the trough SF level was significantly reduced compared to baseline, whereas the trough PT was not prolonged and was similar to the baseline value. The Rocket AF study demonstrated that the preventive effects of treatment with rivaroxaban once daily with the evening meal on cerebral infarction were non-inferior to those of warfarin in patients with AF [
      • Patel M.R.
      • Mahaffey K.W.
      • Garg J.
      • Pan G.
      • Singer D.E.
      • Hacke W.
      • Breithardt G.
      • Halperin J.L.
      • Hankey G.J.
      • Piccini J.P.
      • Becker R.C.
      • Nessel C.C.
      • Paolini J.F.
      • Berkowitz S.D.
      • Fox K.A.
      • et al.
      Rivaroxaban versus warfarin in nonvalvular atrial fibrillation.
      ,
      • Hori M.
      • Matsumoto M.
      • Tanahashi N.
      • Momomura S.
      • Uchiyama S.
      • Goto S.
      • Izumi T.
      • Koretsune Y.
      • Kajikawa M.
      • Kato M.
      • Ueda H.
      • Iwamoto K.
      • Tajiri M.
      Rivaroxaban vs. warfarin in Japanese patients with atrial fibrillation – the J-ROCKET AF study.
      ]. Our results suggest that administration of rivaroxaban once daily in the morning could keep the SF level low, even at trough; thus rivaroxaban might inhibit thrombus formation sufficiently throughout the day.

      Limitations

      Our study was conducted at a single institution, and the observations were made over a relatively short period and in a limited number of subjects. Moreover, as PT measurements differ depending on the reagent, our results cannot be simply compared with those of other studies [
      • Samama M.M.
      • Martinoli J.L.
      • LeFlem L.
      • Guinet C.
      • Plu-Bureau G.
      • Depasse F.
      • Perzborn E.
      Assessment of laboratory assays to measure rivaroxaban – an oral, direct factor Xa inhibitor.
      ]. Although the peak time was defined as 3 h after drug administration, the time reflecting the actual peak concentration of the drug may differ according to the individual.

      Conclusion

      Patients with non-valvular AF exhibited prolonged peak PTs and reduced trough and peak SF levels after receiving rivaroxaban compared to baseline. A prolonged peak PT (≥20 s) places patients at increased risk of bleeding. Treating the patients with rivaroxaban at a fixed dose and without measuring their coagulation function may expose them to the risk of hemorrhage in the future. Measuring PT may help to screen for bleeding risk. PT and SF are both valuable measures of coagulation status in patients receiving rivaroxaban, regardless of prior anticoagulant history.

      Conflict of interest

      Hiroyuki Osanai, Masayoshi Ajioka, Yasuya Inden, and Toyoaki Murohara have received lecture fees from Bayer Yakuhin Ltd., Boehringer Ingelheim, Bristol-Myers Squibb, and Pfizer.

      Acknowledgments

      We gratefully acknowledge the help of the pharmacists and laboratory technicians at Tosei General Hospital, Aichi, Japan.

      References

        • Wolf P.A.
        • Abbott R.D.
        • Kannel W.B.
        Atrial-fibrillation as an independent risk factor for stroke: the Framingham Study.
        Stroke. 1991; 22: 983-988
        • Gage B.F.
        • Waterman A.D.
        • Shannon W.
        • Boechler M.
        • Rich M.W.
        • Radford M.J.
        Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
        JAMA. 2001; 285: 2864-2870
        • Ogawa S.
        • Aizawa Y.
        • Atarashi H.
        • Inoue H.
        • Kamakura S.
        • Koretsune Y.
        • Kumagai K.
        • Mitamura H.
        • Okumura K.
        • Sugi K.
        • Yamashita T.
        • Yasaka M.
        • Hiejima K.
        • Kodama I.
        • Ohe T.
        • et al.
        Guidelines for pharmacotherapy of atrial fibrillation (JCS 2008): digest version.
        Circ J. 2010; 74: 2479-2500
        • Go A.S.
        • Hylek E.M.
        • Phillips K.A.
        • Chang Y.C.
        • Henault L.E.
        • Selby J.V.
        • Singer D.E.
        Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study.
        JAMA. 2001; 285: 2370-2375
        • Ohsawa M.
        • Okayama A.
        • Sakata K.
        • Kato K.
        • Itai K.
        • Onoda T.
        • Ueshima H.
        Rapid increase in estimated number of persons with atrial fibrillation in Japan: an analysis from national surveys on cardiovascular diseases in 1980, 1990 and 2000.
        J Epidemiol. 2005; 15: 194-196
        • Akao M.
        • Chun Y.H.
        • Wada H.
        • Esato M.
        • Hashimoto T.
        • Abe M.
        • Hasegawa K.
        • Tsuji H.
        • Furuke K.
        Current status of clinical background of patients with atrial fibrillation in a community-based survey: the Fushimi AF Registry.
        J Cardiol. 2013; 61: 260-266
        • Koudstaal P.J.
        • Vanlatum J.C.
        • Vangijn J.
        • Kappelle L.J.
        • Algra A.
        • Dehaene I.
        • Dhooghe M.
        • Marchau M.
        • Vanzandijcke M.
        • Delwaide C.
        • Depre A.
        • Laterre E.C.
        • Vanbuggenhout E.
        • Schurmans J.
        • Desmet E.
        • et al.
        Optimal oral anticoagulant-therapy in patients with nonrheumatic atrial fibrillation and recent cerebral ischemia.
        N Engl J Med. 1995; 333: 5-10
        • Yasaka M.
        • Minematsu K.
        • Yamaguchi T.
        Optimal intensity of international normalized ratio in warfarin therapy for secondary prevention of stroke in patients with non-valvular atrial fibrillation.
        Intern Med. 2001; 40: 1183-1188
        • Inoue H.
        • Okumura K.
        • Atarashi H.
        • Yamashita T.
        • Origasa H.
        • Kumagai N.
        • Sakurai M.
        • Kawamura Y.
        • Kubota I.
        • Matsumoto K.
        • Kaneko Y.
        • Ogawa S.
        • Aizawa Y.
        • Chinushi M.
        • Kodama I.
        • et al.
        Target international normalized ratio values for preventing thromboembolic and hemorrhagic events in Japanese patients with non-valvular atrial fibrillation: results of the J-RHYTHM Registry.
        Circ J. 2013; 77: 2264-2270
        • Patel M.R.
        • Mahaffey K.W.
        • Garg J.
        • Pan G.
        • Singer D.E.
        • Hacke W.
        • Breithardt G.
        • Halperin J.L.
        • Hankey G.J.
        • Piccini J.P.
        • Becker R.C.
        • Nessel C.C.
        • Paolini J.F.
        • Berkowitz S.D.
        • Fox K.A.
        • et al.
        Rivaroxaban versus warfarin in nonvalvular atrial fibrillation.
        N Engl J Med. 2011; 365: 883-891
        • Connolly S.J.
        • Ezekowitz M.D.
        • Yusuf S.
        • Eikelboom J.
        • Oldgren J.
        • Parekh A.
        • Pogue J.
        • Reilly P.A.
        • Themeles E.
        • Varrone J.
        • Wang S.
        • Alings M.
        • Xavier D.
        • Zhu J.
        • Diaz R.
        • et al.
        Dabigatran versus warfarin in patients with atrial fibrillation.
        N Engl J Med. 2009; 361: 1139-1151
        • Granger C.B.
        • Alexander J.H.
        • McMurray J.J.V.
        • Lopes R.D.
        • Hylek E.M.
        • Hanna M.
        • Al-Khalidi H.R.
        • Ansell J.
        • Atar D.
        • Avezum A.
        • Cecilia Bahit M.
        • Diaz R.
        • Easton J.D.
        • Ezekowitz J.A.
        • Flaker G.
        • et al.
        Apixaban versus warfarin in patients with atrial fibrillation.
        N Engl J Med. 2011; 365: 981-992
        • Heidbuchel H.
        • Verhamme P.
        • Alings M.
        • Antz M.
        • Hacke W.
        • Oldgren J.
        • Sinnaeve P.
        • Camm A.J.
        • Kirchhof P.
        European heart rhythm association practical guide on the use of new oral anticoagulants in patients with non-valvular atrial fibrillation.
        Europace. 2013; 15: 625-651
        • Hori M.
        • Matsumoto M.
        • Tanahashi N.
        • Momomura S.
        • Uchiyama S.
        • Goto S.
        • Izumi T.
        • Koretsune Y.
        • Kajikawa M.
        • Kato M.
        • Ueda H.
        • Iwamoto K.
        • Tajiri M.
        Rivaroxaban vs. warfarin in Japanese patients with atrial fibrillation – the J-ROCKET AF study.
        Circ J. 2012; 76: 2104-2111
        • Hillarp A.
        • Baghaei F.
        • Blixter I.F.
        • Gustafsson K.M.
        • Stigendal L.
        • Sten-Linder M.
        • Strandberg K.
        • Lindahl T.L.
        Effects of the oral, direct factor Xa inhibitor rivaroxaban on commonly used coagulation assays.
        J Thromb Haemost. 2011; 9: 133-139
        • Lindhoff-Last E.
        • Samama M.M.
        • Ortel T.L.
        • Weitz J.I.
        • Spiro T.E.
        Assays for measuring rivaroxaban: their suitability and limitations.
        Ther Drug Monit. 2010; 32: 673-679
        • Samama M.M.
        • Martinoli J.L.
        • LeFlem L.
        • Guinet C.
        • Plu-Bureau G.
        • Depasse F.
        • Perzborn E.
        Assessment of laboratory assays to measure rivaroxaban – an oral, direct factor Xa inhibitor.
        Thromb Haemost. 2010; 103: 815-825
        • Molenaar P.J.
        • 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: 1799-1807
        • Suzuki S.
        • Otsuka T.
        • Sagara K.
        • Kano H.
        • Matsuno S.
        • Kato Y.
        • Takai H.
        • Uejima T.
        • Oikawa Y.
        • Nagashima K.
        • Kirigaya H.
        • Yajima J.
        • Sawada H.
        • Aizawa T.
        • Yamashita T.
        Rivaroxaban in clinical practice for atrial fibrillation with special reference to prothrombin time.
        Circ J. 2014; 78: 763-766
        • Dempfle C.E.
        The use of soluble fibrin in evaluating the acute and chronic hypercoagulable state.
        Thromb Haemost. 1999; 82: 673-683
        • Sudo A.
        • Wada H.
        • Nobori T.
        • Yamada N.
        • Ito M.
        • Niimi R.
        • Hasegawa M.
        • Suzuki K.
        • Uchida A.
        Cut-off values of d-dimer and soluble fibrin for prediction of deep vein thrombosis after orthopaedic surgery.
        Int J Hematol. 2009; 89: 572-576
        • Tsuji A.
        • Wada H.
        • Matsumoto T.
        • Abe Y.
        • Ota S.
        • Yamada N.
        • Sugiyama T.
        • Sudo A.
        • Onishi K.
        • Nakatani K.
        • Uchida A.
        • Ito M.
        • Suzuki K.
        • Nobori T.
        Elevated levels of soluble fibrin in patients with venous thromboembolism.
        Int J Hematol. 2008; 88: 448-453
        • Fox K.A.A.
        • Piccini J.P.
        • Wojdyla D.
        • Becker R.C.
        • Halperin J.L.
        • Nessel C.C.
        • Paolini J.F.
        • Hankey G.J.
        • Mahaffey K.W.
        • Patel M.R.
        • Singer D.E.
        • Califf R.M.
        Prevention of stroke and systemic embolism with rivaroxaban compared with warfarin in patients with non-valvular atrial fibrillation and moderate renal impairment.
        Eur Heart J. 2011; 32: 2387-2394
        • Tanigawa T.
        • Kaneko M.
        • Hashizume K.
        • Kajikawa M.
        • Ueda H.
        • Tajiri M.
        • Paolini J.F.
        • Mueck W.
        Model-based dose selection for phase III rivaroxaban study in Japanese patients with non-valvular atrial fibrillation.
        Drug Metab Pharmacokinet. 2013; 28: 59-70