Characteristics of Intracerebral Hemorrhage During Rivaroxaban Treatment
Comparison With Those During Warfarin
Background and Purpose—Neuroradiological characteristics and functional outcomes of patients with intracerebral hemorrhage (ICH) during novel oral anticoagulant treatment were not well defined. We examined these in comparison with those during warfarin treatment.
Methods—The consecutive 585 patients with ICH admitted from April 2011 through October 2013 were retrospectively studied. Of all, 5 patients (1%) had ICH during rivaroxaban treatment, 56 (10%) during warfarin, and the other 524 (89%) during no anticoagulants. We focused on ICH during rivaroxaban and warfarin treatments and compared the clinical characteristics, neuroradiological findings, and functional outcomes.
Results—Patients in the rivaroxaban group were all at high risk for major bleeding with hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly, drugs/alcohol concomitantly (HAS-BLED) score of 3 and higher rate of past history of ICH. Moreover, multiple cerebral microbleeds (≥4) were detected more frequently in rivaroxaban group than in warfarin (80% versus 29%; P=0.04). Hematoma volume in rivaroxaban group was markedly smaller than that in warfarin (median: 4 versus 11 mL; P=0.03). No patient in the rivaroxaban group had expansion of hematoma and surgical treatment. Rivaroxaban group showed lower modified Rankin Scale at discharge relative to warfarin, and the difference between modified Rankin Scale before admission and at discharge was smaller in rivaroxaban than in warfarin (median: 1 versus 3; P=0.047). No patient in the rivaroxaban group died during hospitalization, whereas 10 (18%) warfarin patients died.
Conclusions—Rivaroxaban-associated ICH occurs in patients at high risk for major bleeding. However, they had a relatively small hematoma, no expansion of hematoma, and favorable functional and vital outcomes compared with warfarin-associated ICH.
Novel oral anticoagulants (NOACs) reduce incidence of stroke and intracerebral hemorrhage (ICH) in patients with nonvalvular atrial fibrillation.1 Several studies demonstrated hematoma expansion in patients with ICH occurring during warfarin therapy and poor clinical outcomes.2 However, information regarding hematoma size, its expansion, and functional and vital outcomes of patients with ICH occurring during NOAC treatment have been limited and remain largely unclear.
Cerebral microbleeds (CMBs) are said to be predictive of the occurrence of ICH or ischemic stroke3 and to increase the risk of warfarin-associated ICH.4 Furthermore, a considerable interest has been shown in association between CMBs and subsequent ICH in patients treated with NOACs.5 In the present study, we investigated clinical and neuroradiological characteristics of patients with ICH occurring during NOAC treatment.
From April 2011 through October 2013, 585 patients (342 men) with ICH were admitted to the Hirosaki Stroke and Rehabilitation Center for acute therapy <7 days after the onset (n=329) and for further rehabilitation therapy <60 days after the onset from other hospitals (n=256). Of all, 5 patients (1%) had ICH during NOAC treatment with nonvalvular atrial fibrillation, 56 (10%) during warfarin, and the other 524 (89%) during no anticoagulants. In the current study, we focused on patients treated with NOACs and warfarin and retrospectively compared their clinical characteristics, neuroradiological findings, and clinical outcome. This study was approved by the ethics committee of our institution.
ICH was diagnosed by physical examination and immediate computed tomography at admission. Follow-up computed tomographic scan was performed in all patients with ICH at day 2 after admission according to our prespecified clinical protocol or during 2 to 4 days after admission in other hospitals. Hematoma volume was determined by the ABC/2 method, and expansion of hematoma was defined as 33% increase in hematoma volume.6 There were no follow-up computed tomographic data in 8 patients with warfarin who died on the day of admission. For detection of CMBs, T2*-weighted MRI at 1.5 T was performed after acute therapy in all patients with rivaroxaban, and in 42 of 56 patients with warfarin, because 10 patients died in the acute phase and 4 patients had contraindications. Vitamin K, 10 to 40 mg with dependence on international normalized ratio of prothrombin time, was administered to all patients with warfarin for acute therapy. No anticoagulation reversal was performed in patients with rivaroxaban treatment. The congestive heart failure, hypertension, age, diabetes, prior stroke/transient ischemic attack (CHADS2) and hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly, drugs/alcohol concomitantly (HAS-BLED) scores before onset were determined in each patient, as previously described.7 The modified Rankin Scale (mRS) scores before onset and at discharge were evaluated.
The Mann–Whitney U test or Fisher exact test was used to compare differences between 2 groups. Difference between mRS before admission and at discharge (ΔmRS) was compared between 2 groups by the Wilcoxon signed-rank test. Statistical analyses were performed using JMP 10 software and SPSS 16 (SAS, Cary, NC). A P value <0.05 was considered statistically significant.
All 5 patients on NOACs were on rivaroxaban treatment just before the occurrence of ICH. There were no patients with ICH taking other NOACs. The patient profiles are summarized in Table 1. No differences in CHADS2 and HAS-BLED scores, renal function, and antiplatelet use were found between the 2 groups, whereas the past history of ICH was found more in patients with rivaroxaban than in those with warfarin. International normalized ratio at admission was significantly higher in warfarin group than in rivaroxaban group.
Comparisons of Severity of ICH and Clinical Outcome
As shown in Table 2, hematoma volume was markedly smaller in rivaroxaban group than in warfarin group and was ≤10 mL in all patients on rivaroxaban. None of the patients on rivaroxaban showed expansion of hematoma after admission and underwent surgical treatment. In contrast, of the patients treated with warfarin, 10 (21%) showed expansion of hematoma, and 6 (11%) underwent surgical treatment for hematoma. Most importantly, multiple CMBs (≥4) were detected more frequently in patients on rivaroxaban than in those on warfarin. Representative computed tomographic images and T2*-weighted MRI in rivaroxaban group are shown in Figures I and II in the online-only Data Supplement. Regarding clinical outcome, the incidence of patients with mRS ≥4 at discharge did not differ between the 2 groups. However, when only patients with mRS of 0 and 1 before admission were analyzed, none of the patients in the rivaroxaban group showed mRS ≥4 at discharge, whereas a half of patients in the warfarin group showed mRS ≥4 at discharge. Furthermore, ΔmRS in the rivaroxaban group was significantly smaller than that in warfarin group. None of patients on rivaroxaban died during hospitalization, whereas 10 patients (18%) on warfarin died.
Detailed clinical characteristics and outcome of each of 5 patients treated with rivaroxaban are shown in Table in the online-only Data Supplement. Of these 5 patients, 4 had CHADS2 score ≥2, and all 5 had HAS-BLED score of 3 before admission.
In the present study, a significantly smaller size of hematoma was found in patients on rivaroxaban than in those on warfarin, and notably, expansion of hematoma was not observed in any of the patients on rivaroxaban, whereas it occurred in 21% patients on warfarin. Interestingly, rivaroxaban-associated ICH is apt to occur in patients having multiple CMBs or previous ICH.
Patients with HAS-BLED score ≥3 have been shown to be at high risk for major bleeding.7 In the present study, all 5 patients with ICH during rivaroxaban treatment had a HAS-BLED score of 3 and a high rate of past history of ICH, strongly suggesting that patients at high risk of bleeding are also at high risk of ICH during rivaroxaban treatment. Nonetheless, our findings that rivaroxaban-associated ICH showed a significantly small hematoma and no hematoma expansion compared with warfarin-associated ICH may imply a favorable limiting influence of rivaroxaban to hematoma enlargement even in patients at high risk of ICH.
Warfarin inhibits vitamin K–dependent synthesis of coagulation factors II, VII, IX, and X and regulatory factor proteins C and S, whereas rivaroxaban targets only a single coagulation factor Xa. Furthermore, the half-life of rivaroxaban is shorter, 5 to 9 hours in healthy young subjects and 11 to 13 hours in elderly subjects, than that of warfarin, ≈36 to 42 hours.8 These differences in pharmacological action and kinetics may partly contribute to the less disadvantage of rivaroxaban compared with warfarin. Moreover, it is well known that tissue factor and coagulation factor VIIa formation is an important initiator of the coagulation cascade. Tissue factor is at high concentration in the brain and thereby plays an important role in the hemostatic system after ICH. Warfarin inhibits coagulation factor VII, which leads to lower formation of tissue factor and VIIa complex. On the contrary, rivaroxaban does not affect this formation, thereby not reducing the coagulation cascade in the brain.
Recent systematic analysis showed that an excess of CMBs is more frequent in patients with ICH during warfarin treatment compared with those not on warfarin.4 In this context, it is of considerable interest and importance to address the relationship between multiple CMBs and rivaroxaban-associated ICH. Similarly to warfarin, most likely explanation is that rivaroxaban may trigger ICH only in patients at particularly high ICH risk, such as those with multiple CMBs. An alternative possibility is that rivaroxaban use somehow may cause CMBs. The absence of longitudinal data makes it difficult to untangle these 2 possibilities.
There are several limitations in the present study. First, our study is a single-center retrospective analysis and, therefore, would limit generalization of results. Also, because our analyses including cut-point selection of CMB (≥4) and mRS change score were post hoc methods rather than testing of prespecified hypotheses, these may raise the possibility of false-positive results, and therefore our results are exploratory and need to be confirmed in independent data sets. Second, the small number of study patients, particularly rivaroxaban-associated ICH, is at risk of being statistically underpowered. Finally, precise mechanism by which rivaroxaban-associated ICH has relatively small volume and no expansion of hematoma remains largely uncertain in the present study.
In conclusion, ICH may occur during rivaroxaban treatment, especially in patients at a high risk of bleeding such as those with HAS-BLED score ≥3, past history of ICH, and multiple CMBs. Of particular note were the findings that even such patients had a relatively small hematoma, no expansion of hematoma, and favorable functional and vital outcomes. Thus, our study first provides clinical and neuroradiological characteristics of patients with ICH during rivaroxaban treatment.
We gratefully thank Hiroko Kogawa for her excellent technical support.
Dr Joji Hagii received research funding from Bayer Healthcare, and Speakers’ Bureau/Honorarium from Boehringer Ingelheim, Bayer Healthcare, Bristol-Myers Squibb, and Pfizer. Dr Norifumi Metoki received Speakers’ Bureau/Honorarium from Boehringer Ingelheim, Bayer Healthcare, Bristol-Myers Squibb, and Pfizer. Dr Minoru Yasujima received Speakers’ Bureau/Honorarium from Bristol-Myers Squibb and Pfizer. Dr. Ken Okumura received research funding from Boehringer Ingelheim, Bayer Healthcare, and Daiichi-Sankyo, and Speakers’ Bureau/Honorarium from Boehringer Ingelheim, Bayer Healthcare, Bristol-Myers Squibb, Pfizer, and Eisai. The other authors report no conflicts.
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.006661/-/DC1.
- Received July 3, 2014.
- Accepted July 14, 2014.
- © 2014 American Heart Association, Inc.
- Bokura H,
- Saika R,
- Yamaguchi T,
- Nagai A,
- Oguro H,
- Kobayashi S,
- et al
- Lovelock CE,
- Cordonnier C,
- Naka H,
- Al-Shahi Salman R,
- Sudlow CL,
- Sorimachi T,
- et al
- Toyoda K,
- Yasaka M,
- Nagata K,
- Nagao T,
- Gotoh J,
- Sakamoto T,
- et al
- Camm AJ,
- Kirchhof P,
- Lip GY,
- Schotten U,
- Savelieva I,
- Ernst S,
- et al